Automotive Mechatronics PDF

Bosch Professional Automotive
Information
Konrad Reif Ed.
Automotive
Mechatronics
Automotive Networking · Driving
Stability Systems · Electronics
Bosch Professional Automotive Information
Bosch Professional Automotive Information is a definitive reference for
automotive engineers. The series is compiled by one of the world´s largest
automotive equipment suppliers. All topics are covered in a concise but
descriptive way backed up by diagrams, graphs, photographs and tables
enabling the reader to better comprehend the subject.
There is now greater detail on electronics and their application in the motor
vehicle, including electrical energy management (EEM) and discusses the
topic of intersystem networking within vehicle. The series will benefit
automotive engineers and design engineers, automotive technicians in
training and mechanics and technicians in garages.
Konrad Reif
Editor
Automotive Mechatronics
Automotive Networking, Driving Stability
Systems, Electronics
Editor
Prof. Dr.-Ing. Konrad Reif
Duale Hochschule Baden-Württemberg
Friedrichshafen, Germany
reif@dhbw-ravensburg.de
ISBN 978-3-658-03974-5 ISBN 978-3-658-03975-2(eBook)

DOI 10.1007/978-3-658-03975-2
Library of Congress Control Number: 2014946887
Springer Vieweg
© Springer Fachmedien Wiesbaden 2015
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication
or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965,
in its current version, and permission for use must always be obtained from Springer. Violations are liable
to prosecution under the German Copyright Law.
The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws
and regulations and therefore free for general use.
Printed on acid-free paper
Springer Vieweg is part of Springer Science+Business Media

www.springer-vieweg.de
Foreword V
▶ Foreword
As the complexity of automotive vehicles increases this book presents operational

and practical issues of automotive mechatronics. It is a comprehensive introduction
to controlled automotive systems and provides detailed information of sensors for
travel, angle, engine speed, vehicle speed, acceleration, pressure, temperature, flow,
gas concentration etc. The measurement principles of the different sensor groups are
explained and examples to show the measurement principles applied in different
types.
Complex technology of modern motor vehicles and increasing functions need a

reliable source of information to understand the components or systems. The rapid
and secure access to these informations in the field of Automotive Electrics and Elec-
tronics provides the book in the series “Bosch Professional Automotive Information”
which contains necessary fundamentals, data and explanations clearly, systemati-
cally, currently and application-oriented. The series is intended for automotive pro-
fessionals in practice and study which need to understand issues in their area of work.
It provides simultaneously the theoretical tools for understanding as well as the
applications.
VI Contents
▶ Contents
2 Basics of mechatronics 165 Overview of the physical effects for sensors

2 Mechatronic systems and components 167 Overview and selection of sensor
4 Development methods technologies
6 Outlook
168 Sensor measuring principles
8 Architecture 168 Position sensors
8 Overview 195 Speed and rpm sensors
11 Vehicle system architecture 207 Acceleration sensors
212 Pressure sensors
18 Electronic control unit 215 Force and torque sensors
18 Operating conditions 224 Flowmeters
18 Design 230 Gas sensors and concentration sensors
18 Data processing 234 Temperature sensors
22 Digital modules in the control unit 244 Imaging sensors (video)
26 Control unit software
30 Software Development 246 Sensor types
246 Engine-speed sensors
44 Basic principles of networking 248 Hall phase sensors
44 Network topology 249 Speed sensors for transmission control
48 Network organization 252 Wheel-speed sensors
50 OSI reference model 256 Micromechanical yaw-rate sensors
52 Control mechanisms 259 Piezoelectric “tuning-fork” yaw-rate sensor
260 Micromechanical pressure sensors
56 Automotive networking 262 High-pressure sensors
56 Cross-system functions 263 Temperature sensors
57 Requirements for bus systems 264 Accelerator-pedal sensors
59 Classification of bus systems 266 Steering-angle sensors
59 Applications in the vehicle 268 Position sensors for transmission control
61 Coupling of networks 271 Axle sensors
61 Examples of networked vehicles 272 Hot-film air-mass meters
275 Piezoelectric knock sensors
70 Bus systems 276 SMM acceleration sensors
70 CAN bus 278 Micromechanical bulk silicon acceleration
84 LIN bus sensors
90 Bluetooth 279 Piezoelectric acceleration sensors
100 MOST bus 280 iBolt™ force sensor
111 TTP/C 282 Torque sensor
124 FlexRay 283 Rain/light sensor
136 Diagnosis interfaces 284 Two-step Lambda oxygen sensors
288 LSU4 planar wide-band lambda oxygen
144 Automotive sensors sensor
144 Basics and overview
147 Automotive applications 290 Electric Actuators
150 Details of the sensor market 290 Electromechanical actuators
151 Features of vehicle sensors 295 Fluid-mechanical actuators
152 Sensor classification 296 Electrical machines
154 Error types and tolerance requirements
155 Reliability 302 Electrohydraulic Actuators
158 Main requirements, trends 302 Application and Function
Contents VII
302 Requirements 404 Hydraulic modulator

303 Design and Operating Concept 404 Development history
304 Actuator Types 405 Design
313 Simulations in Development 408 Pressure modulation
316 Electronic Transmission Control 412 Sensotronic brake control (SBC)

316 Drivetrain Management 412 Purpose and function
317 Market Trends 414 Design
318 Control of Automated Shift Transmission 414 Method of operation
AST
322 Control of Automatic Transmissions 416 Overview of common-rail systems
338 Control of Continuously Variable 416 Areas of application
Transmission 417 Design
340 ECUs for Electronic Transmission Control 418 Operating concept
347 Thermo-Management 422 Common-rail system for passenger cars
349 Processes and Tools Used in 427 Common-rail system for commercial
ECU Development vehicles
350 Modules for Transmission Control 430 High-pressure components of common-rail

350 Application system
351 Module Types 430 Overview
432 Injector
354 Antilock Braking System (ABS) 444 High-pressure pumps
354 System overview 450 Fuel rail (high-pressure accumulator)
356 Requirements placed on ABS 451 High-pressure sensors
357 Dynamics of a braked wheel 452 Pressure-control valve
358 ABS control loop 453 Pressure-relief valve
362 Typical control cycles
454 Electronic Diesel Control (EDC)
370 Traction Control System (TCS) 454 System overview
370 Tasks 456 Common-rail system for passenger cars
370 Function description 457 Common-rail system for commercial
372 Structure of traction control system (TCS) vehicles
373 Typical control situations 458 Data processing
374 Traction control system (TCS) for four 460 Fuel-injection control
wheel drive vehicles 468 Lambda closed-loop control for
passenger-car diesel engines
378 Electronic Stability Program (ESP) 473 Torque-controlled EDC systems
378 Requirements 476 Data exchange with other systems
379 Tasks and method of operation 477 Serial data transmission (CAN)
380 Maneuvers
388 Closed-loop control system and controlled 478 Active steering
variables 478 Purpose
478 Design
394 Automatic brake functions 480 Method of operation
394 Overview 481 Safety concept
396 Standard function 481 Benefits of active steering for the driver
398 Additional functions
VIII Contents
482 Drive and adjustment systems 496 Electromagnetic compatibility (EMC) and
482 Power windows interference suppression
483 Power sunroofs 496 EMC ranges
484 Seat and steering column adjustment 497 EMC between different systems in the
vehicle
485 Heating, ventilation and air conditioning 504 EMC between the vehicle and its
485 Electronic heater control surroundings
485 Electronically controlled air conditioning 508 Guarantee of immunity and interference
system suppression
488 Vehicle security systems 510 Fault diagnostics

488 Acoustic signaling devices 510 Monitoring during vehicle operation
489 Central locking system (on-board diagnosis)
490 Locking systems 513 On-board diagnosis system for passenger
494 Biometric systems cars and light-duty trucks
520 On-board diagnosis system for heavy-duty
trucks
Authors IX
 Authors
Basics of mechatronics Dipl.-Ing. Christian Gerhardt,
Dipl.-Ing. Hans-Martin Heinkel,  Dipl.-Ing. Klaus Miekley,
Dr.-Ing. Klaus-Georg Bürger. Dipl.-Ing. Roger Frehoff,

Dipl.-Ing. Martin Mast,
Architecture Dipl.-Ing. (FH) Bernhard Bauer,
Dr. phil. nat. Dieter Kraft,  Dr. Michael Harder,
Dipl.-Ing. Stefan Mischo. Dr.-Ing. Klaus Kasten,

Dipl.-Ing. Peter Brenner, ZF Lenksysteme GmbH,
Electronic control units Schwäbisch Gmünd,
Dipl.-Ing. Martin Kaiser, Dipl.-Ing. Frank Wolf,
Dr. rer. nat. Ulrich Schaefer, Dr.-Ing. Johann Riegel.
Dipl.-Ing. (FH) Gerhard Haaf.
Electric Actuators
Basic principles of networking
Dr.-Ing. Rudolf Heinz,
Automotive networking
Dr.-Ing. Robert Schenk.
Bus systems
Dipl.-Ing. Stefan Mischo,
Electrohydraulic Actuators
Dipl.-Ing. (FH) Stefan Powolny,
Electronic Transmission Control
Dipl.-Ing. Hanna Zündel,
Modules for Transmission Control
Dipl.-Ing. (FH) Norbert Löchel,
Dipl.-Ing. D. Fornoff,
Dipl.-Inform. Jörn Stuphorn,
D. Grauman,
Universität Bielefeld,
E. Hendriks,
Dr. Rainer Constapel, Daimler AG Sindelfingen,
Dipl.-Ing. T. Laux,
Dipl.-Ing. Peter Häussermann,
Dipl.-Ing. T. Müller,
Daimler AG Sindelfingen, 
Dipl.-Ing. A. Schreiber,
Dr. rer. nat. Alexander Leonhardi,
Dipl.-Ing. S. Schumacher,
Daimler AG Sindelfingen,
Dipl.-Ing. W. Stroh.
Dipl.-Inform. Heiko Holtkamp,
Universität Bielefeld.
Antilock Braking System (ABS)
Traction Control System (TCS)
Automotive sensors
Electronic Stability Program (ESP)
Sensor measuring principles
Automatic brake functions
Sensor types
Hydraulic modulator
Dr.-Ing. Erich Zabler,
Dipl.-Ing. Friedrich Kost
Dr. rer. nat. Stefan Finkbeiner,
(Basic Principles of Vehicle Dynamics),
Dr. rer. nat. Wolfgang Welsch,
Dipl.-Ing. Heinz-Jürgen Koch-Dücker
Dr. rer. nat. Hartmut Kittel,
(Antilock Braking Systems, ABS),
Dr. rer. nat. Christian Bauer,
Dr.-Ing. Frank Niewels and
Dipl.-Ing. Günter Noetzel,
Dipl.-Ing. Jürgen Schuh
Dr.-Ing. Harald Emmerich,
(Traction Control Systems, TCS),
Dipl.-Ing. (FH) Gerald Hopf,
Dipl.-Ing. Thomas Ehret
Dr.-Ing. Uwe Konzelmann,
(Electronic Stability Program, ESP),
Dr. rer. nat. Thomas Wahl,
Dipl.-Ing. (FH) Jochen Wagner
Dr.-Ing. Reinhard Neul,
(Automatic Brake Functions),
Dr.-Ing. Wolfgang-Michael Müller,
Dipl.-Ing. (FH) Ulrich Papert
Dr.-Ing. Claus Bischoff,
(Wheel-Speed Sensors),
Dr. Christian Pfahler,
Dr.-Ing. Frank Heinen and
Dipl.-Ing. Peter Weiberle,
Peter Eberspächer
Dipl.-Ing. (FH) Ulrich Papert,
X Authors
Sensotronic brake control (SBC) Active steering

Dipl.-Ing. Bernhard Kant. Dipl.-Ing. (FH) Wolfgang Rieger,
ZF Lenksysteme, Schwäbisch Gmünd.
Overview of common-rail systems
High-pressure components of common-rail Drive and adjustment systems
system Dipl.-Ing. Rainer Kurzmann,
Electronic Diesel Control (EDC) Dr.-Ing. Günter Hartz.
Dipl.-Ing. Felix Landhäußer,
Dr.-Ing. Ulrich Projahn, Heating, ventilation and air conditioning
Dipl.-Inform. Michael Heinzelmann, Dipl.-Ing. Gebhard Schweizer,
Dr.-Ing. Ralf Wirth Behr GmbH & Co., Stuttgart.
(Common-rail system),
Ing. grad. Peter Schelhas, Vehicle security systems
Dipl.-Ing. Klaus Ortner Dipl.-Ing. (FH) Jürgen Bowe,
(Fuel-supply pumps), Andreas Walther,
Dipl.-Betriebsw. Meike Keller Dr.-Ing. B. Kordowski,
(Fuel filters), Dr.-Ing. Jan Lichtermann.
Dipl.-Ing. Sandro Soccol,
Dipl.-Ing. Werner Brühmann Electromagnetic compatibility
(High-pressure pumps), Dr.-Ing. Wolfgang Pfaff.
Ing. Herbert Strahberger,
Ing. Helmut Sattmann Fault diagnostics
(Fuel rail and add-on components), Dr.-Ing. Matthias Knirsch,
Dipl.-Ing. Thilo Klam, Dipl.-Ing. Bernd Kesch,
Dipl.-Ing. (FH) Andreas Rettich, Dr.-Ing. Matthias Tappe,
Dr. techn. David Holzer, Dr,-Ing. Günter Driedger,
Dipl.-Ing. (FH) Andreas Koch Dr. rer. nat. Walter Lehle.
(Solenoid-valve injectors),
Dr.-Ing. Patrick Mattes and the editorial team in cooperation with the
(Piezo-inline injectors), responsible in-house specialist departments of
Dipl.-Ing. Thomas Kügler Robert Bosch GmbH.
(Injection nozzles),
Dipl.-Ing. (FH) Mikel Lorente Susaeta, Unless otherwise stated, the authors are all
Dipl.-Ing. Martin Grosser, employees of Robert Bosch GmbH.
Dr.-Ing. Andreas Michalske
(Electronic diesel control),
Dr.-Ing. Günter Driedger,
Dr. rer. nat. Walter Lehle,
Dipl.-Ing. Wolfgang Schauer,
Rainer Heinzmann
(Diagnostics).
Basics
2 Basics of mechatronics Mechatronic systems and components
#BTJDTPGNFDIBUSPOJDT
The term “mechatronics” came about as Mechatronic systems

a made-up word from mechanics and and components
electronics, where electronics means
“hardware” and “software”, and mechan- Applications
ics is the generic term for the disciplines Mechatronic systems and components are
of “mechanical engineering” and “hy- now present throughout almost the entire
draulics”. It is not a question of replacing vehicle: starting with engine-management
mechanical engineering by “electronifi- systems and injection systems for gasoline
cation”, but of a synergistic approach and diesel engines to transmission control
and design methodology. The aim is to systems, electrical and thermal energy
achieve a synergistic optimization of me- management systems, through to a wide
chanical engineering, electronic hard- variety of brake and driving dynamics sys-
ware and software in order to project tems. It even includes communication and
more functions at lower cost, less weight information systems, with many different
and installation space, and better quality. requirements when it comes to operability.
The successful use of mechatronics in a Besides systems and components, mecha-
problem solution is dependent upon an tronics are also playing an increasingly
overall examination of disciplines that vital role in the field of micromechanics.
were previously kept separate.
Examples at system level
A general trend is emerging in the further
development of systems for fully automatic
vehicle handling and steering: more and
more mechanical systems will be replaced
by “X-by-wire” systems in future.
1 Mechatronic system
Environment
Forces, travel, etc. Forces, travel, etc.

Basic system
(mostly mechanical)
Auxiliary
power
Actuator Sensor
engineering technology
Correcting Measured
variables variables
Reference
Feedback Processor variables
UAE1035E
K. Reif (Ed.), Automotive Mechatronics, Bosch Professional Automotive Information,

DOI 10.1007/978-3-658-03975-2_1, © Springer Fachmedien Wiesbaden 2015
Basics of mechatronics Mechatronic systems and components 3
A system that was implemented long ago is Examples at component level

the “Drive-by-wire” system, i.e. electronic Fuel injectors are crucial components in
throttle control. determining the future potential of Diesel-
engine technology. Common-rail injectors
“Brake-by-wire” replaces the hydrome- are an excellent example of the fact that an
chanical connection between the brake extremely high degree of functionality
pedal and the wheel brake. Sensors record and, ultimately, customer utility can only
the driver’s braking request and transmit be achieved by controlling all the physical
this information to an electronic control domains (electrodynamics, mechanical en-
unit. The unit then generates the required gineering, fluid dynamics) to which these
braking effect at the wheels by means of components are subjected.
actuators.
One implementation option for In-vehicle CD drives are exposed to partic-
“Brake-by-wire” is the electrohydraulic ularly tough conditions. Apart from wide
brake (SBC, Sensotronic Brake Control). temperature ranges, they must in particu-
When the brake is operated or in the event lar withstand vibrations that have a critical
of brake stabilization intervention by the impact on such precision-engineered sys-
electronic stability program (ESP), the SBC tems.
electronic control unit calculates the re- In order to keep vehicle vibration away
quired brake pressure setpoints at the in- from the actual player during mobile de-
dividual wheels. Since the unit calculates ployment, the drives normally have a
the required braking pressures separately spring damping system. Considerations
for each wheel and collects the actual val- about reducing the weight and installation
ues separately, it can also regulate the space of CD drives immediately raise ques-
brake pressure to each wheel via the tions concerning these spring-damper sys-
wheel-pressure modulators. The four tems. In CD drives without a damper sys-
pressure modulators each consist of an tem, the emphasis is on designing a me-
inlet and an outlet valve controlled by chanical system with zero clearances and
electronic output stages which together producing additional reinforcement for
produce a finely metered pressure reg- the focus and tracking controllers at high
ulation. frequencies.
Only by combining both measures
Pressure generation and injection are mechatronically is it possible to achieve
decoupled in the Common Rail System. good vibration resistance in driving
A high-pressure rail, i.e. the common rail, mode. As well as reducing the weight by
serves as a high-pressure accumulator, approx. 15 %, the overall height is also
constantly providing the fuel pressure re- reduced by approx. 20 %.
quired for each of the engine’s operating
states. A solenoid-controlled injector with The new mechatronic system for electri-
a built-in injection nozzle injects fuel di- cally operated refrigerant motors is based
rectly into the combustion chamber for on brushless, electronically commutated
each cylinder. The engine electronics re- DC motors (BLDC’s). Initially, they are
quest data on accelerator pedal position, more expensive (motor with electronics)
rotational speed, operating temperature, than previous DC motors equipped with
fresh-air intake flow, and rail pressure in brushes. However, the overall optimization
order to optimize the control of fuel me- approach brings benefits: BLDC motors
tering as a function of the operating condi- can be used as “wet rotors” with a much
tions. simpler design. The number of separate
parts is therefore reduced by approx. 60 %.
4 Basics of mechatronics Mechatronic systems and components
In terms of comparable cost, this more Development methods

robust design doubles the service life,
reduces the weight by almost half and Simulation
reduces the overall length by approx. 40 %. The special challenges that designers face
when developing mechatronic systems are
Examples in the field of micromechanics the ever shorter development times and
Another application for mechatronics is the increasing complexity of the systems.
the area of micromechanical sensor sys- At the same time, it is vital to ensure that
tems, with noteworthy examples such as the developments will result in useful
hot-film air-mass meters and yaw-rate products.
sensors.
Because the subsystems are so closely Complex mechatronic systems consist of
coupled, microsystems design also re- a large number of components from differ-
quires an interdisciplinary procedure that ent physical domains: hydraulic compo-
takes the individual disciplines of mechan- nents, mechanical components and elec-
ical components, electrostatics and possi- tronic components. The interaction be-
bly fluid dynamics and electronics into tween these domains is a decisive factor
consideration. governing the function and performance
of the overall system. Simulation models
are required to review key design deci-
sions, especially in the early development
stages when there is no prototype avail-
able.
2 Model library for a micromechanical yaw-rate sensor
Microsystem
Mechanical Electro-
components mechanical
components
Rigid Elastic Comb-like Detection

bodies bodies structures electrodes
From From Bending Segment Undivided Divided From From

UAE0942-1E
segments segments beams of a circle stator comb stator comb segments segments
of a circle of a rectangle of a circle of a rectangle
Basics of mechatronics Development methods 5
Basic issues can often be clarified by pro- However, an analysis of the typical compo-
ducing relatively simple models of the nents in mechatronic systems shows that
components. If more detail is required, they can be composed of a few simple ele-
more refined component models are ments specific to the domains. These stan-
needed. The detailed models focus mainly dard elements are, for example:
on a specific physical domain: • In the hydraulic system: throttle, valve
• This means that detailed hydraulic mod- or electric line
els of common rail injectors are avail- • In the electronic system: resistor, capac-
able, for example. These can be simu- itor or transistor
lated using special programs with nu- • In the mechanical system: ground with
meric calculation methods that are friction, transmission or clutch (or the
exactly tailored to hydraulic systems. equivalent for micromechanics)
Cavitation phenomena have to be taken
into consideration, among other things. The preferable solution is that these ele-
• Detailed models are also needed to de- ments should be stored in a central stan-
sign the power electronics that trigger dard model library that is also decentrally
the injector. Again, this involves the use accessible to product development. The
of simulation tools which must be devel- essence of the standard model library is
oped specifically to design electronic a documentation of all the standard ele-
circuits. ments. For each element, this comprises:
The development and simulation of the Description of physical behavior in
software that controls the high-pressure words
pump and the power electronics in the The physical equations, parameters
control unit with the aid of the sensor (e.g. conductivity or permeability),
signals also takes place using tools that state variables (e.g. current, voltage,
are specially designed for this area of magnetic flux, pressure) and
the overall system. The description of the associated inter-
faces
As the components in the overall system
interact with each other, it is not sufficient In addition, a major part of the environ-
to consider specific detailed models of the ment is a reference model written in a
components in isolation. The optimum so- modeling language that is independent
lution is also to take into account the mod- of the tool. Overall, the library includes
els of other system components. In most reference models from the mechanical,
cases, these components can be repre- hydraulic, electronic, electrodynamic
sented by simpler models. For example, and software areas.
the system simulation that is focussed on
the hydraulic components only requires
a simple model of the power electronics.
The application of various domain-specific

simulation tools during the development
of mechatronic systems is only efficient if
there is some sort of support for exchang-
ing models and parameters between the
simulation tools. The direct exchange of
models is highly problematic due to the
specific languages used for describing the
models of each of the tools.
6 Basics of mechatronics Development methods
V model depending on the technologies applied, for

The dependencies of the different product each of the associated domains (mechani-
development phases are illustrated in the cal engineering, hydraulics, fluid dynam-
“V model”: from requirement analysis to ics, electrics, electronics, and software).
development, implementation, testing and
system deployment. A project passes Recursions at each of the design levels
through three “top-down” levels during shorten the development stages signifi-
the development stage: cantly. Simulations, rapid prototyping, and
• Customer-specific functions simultaneous engineering are tools that al-
• Systems and low rapid verification, and they create the
• Components conditions for shortening product cycles.
A requirements specification (what) must Outlook

first be produced at each level in the form
of specifications. This results in the design The major driving force behind mecha-
specification, which is drawn up on the ba- tronics is continuous progress in the field
sis of design decisions (the actual creative of microelectronics. Mechatronics benefits
engineering work). The performance spec- from computer technology in the form of
ifications describe how a requirement can ever more powerful integrated computers
be met. The performance specs form the in standard applications. Accordingly,
basis for a model description which allows there is a huge potential for further in-
a review (i.e. validation) of the correctness creases in safety and convenience in
of each design stage together with previ- motor vehicles, accompanied by further
ously defined test cases. This procedure reductions in exhaust emissions and fuel
passes through each of three stages, and, consumption. On the other hand, new
3 Recursion method at one level

Deve
Requirement
specification (what) Tool-supported
test-case creation
lopm
Specifications Validation,
feasibility
ent p
Design decisions
(”creative engineering
work”)
roces
Model,
Test cases
prototype
s
(Virtual)
sample
Performance
specifications
Design
UAE0943-1E
specification (how)
Basics of mechatronics Outlook 7
challenges are emerging with regard to The design approaches of mechatronic

the technical mastery of these systems. systems should strive toward continuity
However, future “X-by-wire” systems in several aspects:
without the mechanical/hydraulic fall- • Vertical:
back level must also provide the pre- “Top-down” from system simulation,
scribed functionality in the event of a with the objective of overall optimiza-
problem. The condition for their imple- tion, through to finite element simula-
mentation is a high-reliability and high- tion to achieve a detailed understanding,
availability mechatronic architecture and “bottom-up” design engineering
which requires a “simple” proof of safety. from component testing through to
This affects both single components as system testing
well as energy and signal transmissions. • Horizontal:
“Simultaneous engineering” across

As well as “X-by-wire” systems, driver-as- several disciplines in order to deal with
sistance systems and the associated man/ all product-related aspects at the same
machine interfaces represent another area time
in which the consistent implementation of • Beyond company boundaries:
mechatronic systems could achieve signifi- Step by step, the idea a “virtual sample”
cant progress for both users and vehicle is nearing our grasp
manufacturers.
Another challenge is training in order to
further an interdisciplinary mindset and
develop suitable SE processes and forms
of organization and communication.
4 V-model general overview
Product
Develo
Customer
n
Functio
wishes
Validation Test
Requirement
analysis Acceptance test
pment
Model Test
cases
System
requirement
proces
specifications
Validation Test
System
System design System test

Model, Test
System target prototype cases
s
specifications
Component
requirement
specifications
Validation Test
nts
Component design, Component test

development Model, Test
ne
prototypes cases
Compo
Component
UAE0944-1E
target
specifications
Component manufacture
8 Architecture Overview
"SDIJUFDUVSF
Over the last three decades, tremendous Overview

progress has been made in automotive
engineering. Modern injection and ex- History
haust-gas treatment systems drastically The on-board electrical network of a car
reduced pollutants in the exhaust gas, around the year 1950 comprised approx.
while occupant-protection and vehicle 40 lines. Essentially, cables were only re-
stabilization systems improved safety quired for the battery, starter, ignition and
on the road. Much of this success is due the lighting and signaling systems.
to the introduction of electronically-con- With the first electronic injection and
trolled systems. The proportion of these ignition systems, cabling complexity began
systems used in cars increased continu- to increase fast. Sensors fitted in the en-
ously. The requirements of safety and gine compartment (e.g. speed sensor,
environmental compatibility, but also engine-temperature sensor) had to deliver
the demand for comfort and convenience signals to the engine control unit, while
functions, will increase yet further and the fuel injectors required their triggering
this will in no small part be achieved signals from the electronic control unit.
through the use of electronics. Up to A further increase in cabling complexity
around 90 % of innovations in the motor resulted from the introduction and rapid
vehicle will be realized by electronics widespread adoption of the antilock brake
and microprocessor-controlled systems. system (ABS). Meanwhile, comfort and
The networking of these electronics cre- convenience systems, e.g. electrical power-
ates the prerequisite for having this wide window units, would also form part of the
variety of electronic systems integrated standard equipment. All these systems re-
within the complete vehicle system to quire additional connecting lines for the
form a whole. However, this results in connection of sensors, control elements
a complexity that can only be overcome and actuators to the control unit.
at considerable expense.
1 Proportion of electrics/electronics in the motor vehicle
Automobile 2000 Automobile 2010

100%
Electronics Electronics Electronics Electronics Electronics Electronics
80%
Hydraulics Pneumatics
Hydraulics Electronics Hydraulics Elec- Elec-
60% Pneumatics Hydraulics tronics tronics
40%
Mechanics Mechanics Mechanics
Mechanics Mechanics Mechanics
20% Mecha- Mecha-
Mechanics
nics nics
0%
40% 30% 22% 8% 30% 30% 25% 7% 8%
Driving Safety Con- Info- Driving Safety Con- Info- Commun-
and braking venience tainment and braking venience tainment ication/
navigation
Fig. 1
SVA0032E
Source:
Proportion of electrics/electronics, Proportion of electronics,
approx. 22% approx. 35%
Mercer management
consulting

Architecture Overview 9
Technology of the present day shorten, the airbag and seat-belt preten-
In the 1990s the cabling work in a luxury sioners are set to emergency standby.
class vehicle amounted to around 3 km. The communication between the elec-
This figure clearly demonstrates how tronic control units cannot take more than
complex the vehicle system has become. fractions of a second. The more electronic
The growth of the proportion of electron- control units interact in the one complete
ics in the motor vehicle (Fig. 1) can mainly system, the more difficult it becomes for
be attributed to the growth in microelec- them to communicate undisturbed.
tronics and sensor technology. With the number of electronic control
At first, many of the new systems were units and the associated need for mutual
integrated into the vehicle by means of communication, the costs of developing
their own dedicated electronic control the systems rose as did the adaptation
unit. For the most part, the individual costs for making interfaces compatible.
electronic control units operated in mutual With the CAN bus (Controller Area Net-
independence. All the same, connecting work) developed by Bosch, a powerful and
lines became increasingly necessary be- widely used data bus system has become
tween electronic control units to enable commonplace in vehicles for the first time.
the exchange of data by means of PWM The data line of the CAN bus makes it pos-
signals, for example. Depending on the sible for the electronic control units to
vehicle class, there are between 20 and exchange specific and relevant items of
80 electronic control units fitted in today’s information with each other. At the start,
vehicles. They control such equipment as the network only comprised a few elec-
the engine, antilock brake system or the tronic control units, such as the engine-
airbags. The number of microcontrollers management system, the electronic stabil-
in the vehicle has therefore risen continuity program and the transmission control.
ously in recent years (Fig. 2). Gradually, further systems would expand
The components of the individual sys- this network, especially in the areas of
tems are optimally matched to each other. comfort and convenience and infotain-
The systems may originate from different ment. The CAN bus has gradually evolved
manufacturers that use previously agreed, into the standard for networking systems
albeit still their own, interfaces. The rain in the motor vehicle. Today it is the stan-
sensor, for example, “speaks” in a different dard for communication between elec-
way to the sensors for the engine manage-
ment. The following example demon- 2 Number of microcontrollers in the motor vehicle
strates just how networked the functions

150
in a modern vehicle are: the radar sensor 140
of the adaptive cruise control system 130
(ACC) measures the distance to the vehicle 120
Number of microcontrollers
110
traveling in front. If this distance is shorter
100
than a specified minimum distance, the 90
ACC electronic control unit sends this in- 80
formation to the engine management, the 70
60
ESP electronic control unit and the airbag
50
electronic control unit. The engine man- 40
agement reduces torque and thus driving 30
speed. If this is not sufficient, the elec- 20
SVA0033E
10
tronic stability program (ESP) must also
0
generate brake pressure to decelerate 1980 1985 1990 1995 2000 2005
the vehicle. If the distance continues to
10 Architecture Overview
tronic control units within different areas objectives simultaneously, development

of the electronics (drivetrain, suspension, partners are more frequently tapping into
body electronics and infotainment) and resources that are already available in sub-
forms a powerful backbone for networking systems. These can be sensors or actuators
these areas with each other. Additional bus as well as realized functions that are avail-
systems (e.g. LIN bus, MOST bus) are used able to different systems over the commu-
as subbuses or for transmitting at high nications network. For new systems and
data rates with comparatively low realtime functions, manufacturers strive to get by
requirements in the motor vehicle. on a minimum of additional resources.
In the meantime, engineers are faced with
Development trends a new challenge in the form of "networked"
The proportion of electrics and electronics thinking and subsystem integration, espe-
in the motor vehicle will continue to in- cially when the assemblies for the subsys-
crease. In the drivetrain, the number of tems originate from different development
components in the exhaust line (e.g. ex- partners (suppliers).
haust-gas sensors) is increasing due to Complaints in the field (i.e. with series-
stricter exhaust-emissions legislation. production vehicles) due to electrical or
While the demands for reductions in fuel electronic failures could be the conse-
consumption can, for example, be fulfilled quence of not having taken the interac-
by means of new valve-gear concepts, even tions of the subsystems into consideration.
this requires additional electronic compo- The causes – unmanageable behavior of
nents. A further increase in the proportion functionality spread among networked
of electronics results mainly from the systems, and their integration – are avoid-
growth of electronic systems in the areas able through the logical application of cer-
of safety, comfort and convenience, and tified development processes as early as
infotainment. in the specification phase. Furthermore,
modeling and tools for authoring a formal
Objectives description of architectures are gaining
Drivers demand a high level of reliability ever more in importance.
from a car. The vehicle manufacturer and
the supplier of assemblies, meanwhile, Broadened requirements for a complete
are constrained by other requirements motor vehicle system in the future are
such as minimization of manufacturing leading to increased networking of vehicle
costs, space restrictions and the weight components and subsystems. In this re-
of components. An opportunity to fulfill gard, new functions are being developed
these requirements in the face of the in- that go beyond the frontiers of traditional
creasing complexity of the “vehicle” sys- applications – and this is without addi-
tem is seen in the shift of the traditional tional expenditure on hardware wherever
implementation technologies of mechan- possible.
ics, hydraulics and electrics towards New development methods and technol-
microprocessor-controlled, electronic ogies are required to make this achievable.
systems. For this reason, the development With a top-down approach, new functions
of software will continue to gain in impor- are viewed from the perspective of the
tance in future. complete vehicle. This means that, in ac-
The current situation in the electrical cordance with the method of systems engi-
and electronic architecture of motor vehi- neering, functional requirements and non-
cles is characterized by an increase in functional requirements (e.g. quality ob-
functionality and an increasingly strained jectives, safety, costs, etc.) are set for the
costs situation. To achieve both of these vehicle as a whole and derived as specifi-
Architecture Overview 11
cations for the subordinate subsystems. Vehicle system architecture

These requirements are formulated as a
model and can thus be used as a specifica- Architecture
tion for the subsystems and the creation The architecture of a system represents
of test cases. This is what is known as an its “construction plan”. It describes the
“executable specification”, which makes structural and dynamic system character-
it possible to prove the completeness and istics as a whole. The architecture is usu-
the traceability of the requirements, for ally specified in a description language.
example, or to identify the requirements Special draft mechanisms are used for
for interaction and communication be- specific requirements. With architecture
tween subsystems. In this way, it is possi- being a construction plan for different
ble to form an optimized architecture for realization technologies and a means of
the complete vehicle and its subsystems proving that functional and nonfunctional
and components. The functional relation- requirements have been fulfilled in the
ships between the complete motor vehicle system draft, different views of the system
system and the subordinate subsystems architecture are required. Examples of this
can be surveyed in different levels of detail include:
and suitable interfaces can be defined for • Hardware architecture
the functions. This approach supports an • Software architecture
expanding networking of functions. Syner- • Network architecture in the area of real-
gies are exploited between vehicle areas ization technologies
(domains such as the drivetrain, interior, • Cost and resource consumption in the
infotainment) that were hitherto consid- area of economical analysis and
ered in isolation and resources are spared. • For the area of social requirements,
As an element of the development pro- aspects such as safety, availability and
cess that works in the opposite direction, legal conformity
the generation of new functions from avail-
able resources and existing systems (bot- The problems that arise in the integration
tom up) should also be taken into consid- of differently structured subsystems can
eration to minimize innovation risks. be reduced by means of an architecture.
This is how new functions are integrated
into existing systems, for example. Exam- Functional structure
ples of this approach are measures to avert The domain of vehicle motion has the task
the consequences of an accident by “pre- of ensuring the controlled movement of
paring” subsystems for an imminent crash the vehicle as well as its directional stabil-
(closing windows, closing the sliding sun- ity. This task can be subdivided into vari-
roof, activating the airbag, etc.) or the as- ous levels (Fig. 3).
sistance of the driver in emergency brak- The navigation level is home to the plan-
ing situations in ESP in future. In this way, ning tools for the driving route. These are
it is possible to reduce the number of elec- merely informational in nature and have
tronic control units and counteract rising no interventional influence on vehicle
system costs. motion.
The development process described At vehicle guidance level, the decisions
characterizes the CARTRONIC® concept of the driver are implemented by means of
that Bosch developed in the 1980s. the steering wheel and accelerator pedal
The results of this concept are being but also various assistance systems for ve-
incorporated into the Autosar Initiative hicle handling (e.g. ACC, course stability
(see Autosar Initiative)
12 Architecture Vehicle system architecture
systems). At this level, the driver is able functional component represents the tasks
to overrule the assistance systems at any of the navigation level, which are to inform
time. the driver of the driving route determined
At the stability level, there are the sub- by means of a mapping system (Fig. 3).
systems that are able to correct the deci- Vehicle guidance represents the guidance
sions taken at handling level if these hap- level, and stability intervention the tasks
pen to be outside the range of safe refer- of the stabilization level. The vehicle mo-
ence variables (e.g. ABS, ESP). This may tion coordinator determines the correcting
be the case when cornering or on wet road variables for the actuators, e.g. of the
surfaces, for example. drive and electronic stability program
At stabilization level, correcting vari- (ESP), from the information input by
ables for implementation by the vehicle’s vehicle guidance and stability interven-
actuators are determined. Information tion.
about the environment (e.g. road condi- Figure 4 shows how the functional com-
tion, air temperature, rain sensor signal) ponents of guidance level, stabilization
is still required at the various levels for level and vehicle actuators are related in
the implementation of the relevant tasks. a hierarchical structure within vehicle
These tasks can be assigned to func- motion. Communication relationships
tional components, which are the architec- between the components and interactions
tural elements of the functional architec- with other domains, e.g. body and interior,
ture. In this way, the driver information are also featured in the model.
3 Levels in the vehicle motion domain
Environment Driving and assistance

systems
Navigation level
Road network
Driver information
Calculated distance
Road layout Guidance level

Traffic regulations
Vehicle guidance
Vehicle guidance data Vehicle

Road conditions
Stability level Actuators
Vehicle condition
Vehicle motion $ive
coordinator
$e
$
ing
View Stability
Weather conditions intervention $
SVA0034E
Vehicle motion
13
Architecture Vehicle system architecture
In the same way as Vehicle motion is Systematic creation of EE system

refined, these functional components architectures
require further detailing until the refined The increasing amount of networking in
components represent manageable, traditional vehicle domains for the realiza-
clearly delimited tasks that make flexible, tion of new functions can be illustrated
modular implementation possible through using the ACC (Adaptive Cruise Control)
different realization technologies. Defined driver-assistance system as an example.
interfaces between the components enable Adaptive, same-lane driving is made possi-
communication and the exchange of data. ble by the networking of a combined
For example, the transmission control cruise and distance control system with
issues a request through the engine-man- the engine-management system, brake sys-
agement system for a specific reduction tem, transmission and cockpit. Here, sub-
in torque during a gearshift. This value is systems from the drivetrain, chassis and
exchanged as a physical variable via the infotainment (interaction with the driver)
interface. domains are used to realize the new func-
With its integration into a suitable pro- tion with minimal cost.
cedural model, the functional structure is The decision as to whether a function
the starting point for subsequent stages in (e.g. ACC) is realized in a dedicated logic
the development process. close to the sensor or in one of the exist-
ing, subscriber electronic control units
has no bearing on the function itself.
4 Example of a functional structure for the vehicle motion domain
Vehicle motion
Vehicle guidance
Acceleration
requirement
Steering angle
Vehicle motion coordinator
Brake torque
Drive torque Steering angle
Steering angle
Drive Brake Steering Chassis Stability intervention
Stop lamp
SVA0035E
Body, interior
Rather, the decision is affected by non- ization stages. This required a decoupled
functional requirements such as safety, development process and the exploitation
availability, costs or resource availability. of synergies between subsystems. The de-
In addition to the functional requirements, velopment frameworks took into consider-
these requirements mainly determine how ation the dependencies and interface con-
the function is realized. The “how” is de- tents within the individual domains and
scribed by the architecture of the system. with the rest of the vehicle, as is the case
Different requirements result in different with a networked system such as ACC,
system architectures. for example.
CARTRONIC® concept Bosch introduced this concept to the Auto-

With the CARTRONIC® architecture con- sar Initiative (Workpackage 10.x).
cept, all closed and open-loop control
tasks in the vehicle have been structured Software architecture
in accordance with logical, functional The independence of the functional struc-
viewpoints and modeled in the form of ture, or architecture, from the later real-
a functional architecture. Delimited func- ization stage results in a decoupling of
tions (and their dependencies) that imple- functionality and technology and thus
ment specific functional requirements forms the first stage of a model-based
have been represented by defined archi- development process. The functional
tectural elements. The functional struc- structure can be used on several occasions
ture, i.e. the structural description, repre- and expanded as the foundation for draft-
sented a hierarchical decomposition of ing system architectures. This architecture
the subsystems down to manageable size. is characterized by architecture drivers
Interactions between elements of the func- (specific criteria of the architecture) that
tional structure have been described by are essentially the product of nonfunc-
communication relationships. Since the tional requirements (e.g. costs, quality,
use of the architecture concept could have reusability, relocatability).
led to different functional structures, Further precision of the development
it was essential to reach agreement on frameworks devised from the functional
the tasks and interfaces. It was necessary structure, and of their interfaces in partic-
to choose interfaces that were based on ular, is required if it is to be possible to
physical variables and thus supported evaluate an electronic control unit for the
aspects such as reusability and inter- relocatability of functions and the integra-
changeability. tion of software, which is a contribution
of various participants in the project.
The motor vehicle system with all its open While retaining the realization-indepen-
and closed-loop tasks was dismantled into dent information from the functional
subsystems that implement clearly defined structure – such as an agreed torque inter-
tasks. These subsystems include the en- face – the frameworks are supplemented
gine management, brake system, transmis- by realization-specific information such
sion control, ACC, lighting management, as data type, quantization, runtime prop-
etc. Different levels of functional structure erties or resource requirements.
detail can be assigned to the system and
subsystem levels (Fig. 5). It was therefore In the same way as hardware and disci-
possible to create development frame- plines such as mechanics or hydraulics,
works for selected functional components software can also be classified as a realiza-
and component groups on which to base tion technology. Product-line or platform
implementation in the form of partial real- approaches have long been a familiar
15
feature of mechanical development or pro- Network architecture

duction optimization and their application With the spreading of open standards
is virtually universal. The trend towards such as the CAN bus, the integration of
software-based system solutions and im- functions into application-specific elec-
provements, in conjunction with the vastly tronic control units, and satellites linked
expanding software scopes of electronic by subnetworks, network architecture has
systems, has given rise to the demand for become the synonym for the complexity
this strategy to also be transferred to soft- management of distributed systems.
ware-intensive systems. Extensions and “attachment solutions”
The decisive challenge faced today is are easily integrated until the limits of
less to do with technical feasibility but network capacity are reached. If these
more about how to develop methods fur- possibilities were to be exploited without
ther and apply already developed methods checking the system draft, this would
and processes in product developments result in unmanageable increases in com-
in a systematic and disciplined way, and plexities and integration conflicts. Biologi-
to anchor them within the organization. cal systems solve these unmanageable
The product-line approach in software de- complexities through specialization and
velopment was transferred to the motor the creation of subnetworks with new
vehicle domain with the participation of forms of organization. Their objectives
Bosch with methodological support from are stability and the ability to survive.
the software engineering institute (SEI). This model has, to a certain extent,
The method will be used systematically evolved on its own in motor vehicle sys-
in future Bosch product generations. tems through assignment to traditional
5 Functional architecture
System level
Subsystem
level
Hardware Software
architecture architecture
Network architecture
SVA0036E
fields of application or domains and the reliability are fulfilled by the multiple use
comparatively slow growth of networking of proven standards. Autosar concerns
within these domains. itself with all vehicle domains.
Based on the uniform electronics plat-
Bus systems for the individual domains are form, which primarily consists of standard
becoming more specialized due to their software modules, each vehicle manufac-
plainly different requirements. With the turer is then free to build its own specific
CAN in the drivetrain as the point of ori- content. They enable integration into the
gin, new bus systems such as the LIN sub- electronics network. These software func-
bus have begun to infiltrate the area of tions permit differentiation between the
body electronics or FlexRay in the case of competition.
safety-relevant x-by-wire systems. In the Not only does software have to conform
multimedia field, where demands for high to the Autosar standard. The electronic
data rates but low safety requirements control units must be built in such a way
prevail, bus systems such as Bluetooth that the Autosar software is able to run on
have started to make an appearance. them. The Autosar members are hoping
Breaking through these traditional do- that the new development methods yield
mains with ever more applications leads such benefits as shorter development
to known consequences, e.g. dramatic in- times and lower development costs.
crease in complexity, high start-up costs,
increasing integration times and costs, and Until now, it was often the case that dedi-
more demanding work in customer service cated electronic control units would be
as a consequence of diagnostics no longer developed and fitted for new functions
being manageable. A solution for these (e.g. electronic transmission control,
multidimensional optimization tasks has in antilock brake system, air conditioning).
the past been sought in the software field. The number of electronic control units
In the case of technical systems in particu- fitted in the vehicle grew continuously;
lar, the paradigm is still king, especially in today’s generation of vehicles are
software realizations, because the absence equipped with between 20 and 80 elec-
of physical boundaries supports unlimited tronic control units. In future vehicle gen-
growth. erations, it is intended that all functions
be covered by a network of 10 to 20 elec-
Autosar Initiative tronic control units. Some of these will
The Autosar Initiative (AUTomotive Open function a little like main computers that
Systems ARchitecture) was founded in will bundle the important function groups
July 2003 by several vehicle manufacturers together. These include the drivetrain,
and suppliers – Bosch among them. Their suspension management system, body
global objective is the joint development and interior and the multimedia/telematics
of an open system architecture for future domain. On data buses, sensors with inte-
automotive applications. The aims of the grated electronics output processed and
partnership include the standardization verified signals, while the buses carry the
of fundamental system functions (basic relevant control commands to actuators
software) and function interfaces; they with integrated triggering electronics.
will replace the company-specific, individ- In future, new functions will often be
ual solutions used to date. Model-based able to use the existing computer architec-
concepts and methods ought to reduce ture up to its performance limit and will
complexity in spite of an expanding range be widely realized in the form of a soft-
of functions. The demands for quality and ware add-on. This would therefore render
17
unnecessary the additional electronic con- Outlook

trol unit that would have been required Increasingly greater demands for safety,
today. The system only needs to be supple- reliability and availability are being placed
mented by the sensors and actuators re- on the network architecture of modern
quired. vehicles. This is where energy network
architecture will play a key role. In the
Software will no longer be an inevitable face of vehicle functions increasingly be-
component of hardware, but will increasing realized electronically, this architec-
ingly become a stand-alone product. ture makes for a reliable supply of power
The first examples of business and col- to systems and thus forms the basis for
laborative models between supplier and the reliability and safety of future systems.
manufacturer have already been put into One possible future technology is the
practice at Bosch, e.g. in drivetrain man- transmission of power and information
agement. on the supply line. The following benefits
arise from the powerline communication
Examples (PLC) concept used in the public grid:
Individually-controlled drive components • Weight and cost reductions as well as
at each of the wheels with different wheel space savings from the discontinuation
positions and wheel loads permit optimum of data lines
use of tire force potential. This results in • Easier retrofitting for retrofit systems
increased driving dynamics and safety (spare parts trade)
while at the same time reducing consump- • Reduction in complexity of the wiring
tion, wear and emissions. For this to be harness in respect of manufacture and
possible, all active elements in the drive- installation
train, suspension and steering must be • Increase in system safety, especially in
networked. mechanically stressed zones (e.g. door,
One example of superordinate functions mirror) that are characterized by pre-
realized by networking is the ASIS (Active mature aging of lines and increased risk
Shift Strategy) drive strategy for automatic of failure
transmissions in passenger cars. Based on • Powerline as a redundancy path for sys-
the evaluation of various control elements tems relevant to safety
(e.g. accelerator pedal) and conclusions • Simultaneous, or parallel, implementa-
drawn from the information of other sys- tion of several bus systems or services,
tems (e.g. cornering detection from the e.g. diagnostics
wheel speeds), this strategy is able to con-
trol the gearshift in such a way as to meet
the driver’s real-time demand for agility
and power through selection of the appro-
priate gear. In future, the telematics will
be able to support additional, improved
driving strategies, e.g. through the use
of GPS signals for transmission control
in terms of predictive driving.
18 Electronic control unit Operating conditions, design, data processing
&MFDUSPOJDDPOUSPMVOJU
Digital technology furnishes an extensive Design

array of options for open and closed-loop
control of automotive electronic systems. The printed circuit board with the electri-
A large number of parameters can be in- cal components (Fig. 1) is installed in a
cluded in the process to support optimal housing of plastic or metal. A multiple
operation of various systems. The control plug connects the control unit to the sen-
unit receives the electrical signals from sors, actuators and electrical power sup-
the sensors, evaluates them, and then ply. The high-performance driver circuits
calculates the triggering signals for the that provide direct control of the actuators
actuators. The control program, the are specially integrated within the housing
“software”, is stored in a special memory to ensure effective heat transfer to the
and implemented by a microcontroller. housing and the surrounding air.
The control unit and its components are The majority of the electrical compo-
referred to as hardware. The Motronic nents are of the surface-mounted device
control unit contains all of the algo- technology type. This concept provides
rithms for open and closed-loop control extremely efficient use of space in low-
needed to govern the engine-manage- weight packages. Only a few power com-
ment processes (ignition, induction and ponents and the connectors use push-
mixture formation, etc.). through assembly technology.
Operating conditions Hybrid versions combining compact

dimensions with extreme resistance to
The control unit is subjected to very thermal attack are available for mounting
high demands with respect to directly on the engine.
• Extreme ambient temperatures
(in normal vehicle operation from Data processing
–40 to +60...+125 °C)
• Extreme temperature changes Input signals
• Indirect materials and supplies In their role as peripheral components,
(oil, fuel etc.) the actuators and the sensors represent
• The effects of moisture and the interface between the vehicle and the
• Mechanical stress such as vibration control unit in its role as the processing
from the engine unit. The electrical signals of the sensors
are routed to the control unit via a wiring
The control unit must operate reliably harness and the connector plug. These sig-
when the vehicle is started with a weak nals can be of the following type:
battery (e.g. cold start) and with high
charge voltages (vehicle electrical system Analog input signals
fluctuations). Within a given range, analog input signals
can assume practically any voltage value.
Other requirements arise from the need Examples of physical quantities which
for EMC (ElectroMagnetic Compatibility). are available as analog measured values
The requirements regarding immunity to are intake-air mass, battery voltage, in-
electromagnetic interference and limita- take-manifold and boost pressure, coolant
tion of high-frequency interference signal and intake-air temperature. They are con-
emission are extremely stringent. verted into digital values by an analog-
digital converter in the microcontroller

Electronic control unit Data processing 19
of the control unit and used for calcula- Signal conditioning

tions by the microcontroller CPU. The Protective circuits limit the voltages of in-
maximum resolution of these analog sig- put signals to levels suitable for process-
nals is 5 mV. This translates into roughly ing. Filters separate the useful signal from
1,000 incremental graduations based on most interference signals. When neces-
an overall measuring range of 0 to 5 V. sary, the signals are then amplified to the
input voltage required by the microcon-
Digital input signals troller (0 to 5 V).
Digital input signals only have two states. Signal conditioning can take place com-
They are either “high” or “low” (logical 1 pletely or partially in the sensor depend-
and logical 0 respectively). Examples of ing upon the sensor’s level of integration.
digital input signals are on/off switching
signals, or digital sensor signals such as Signal processing
the rotational-speed pulses from a Hall The control unit is the switching center
generator or a magnetoresistive sensor. governing all of the functions and se-
Such signals are processed directly by quences regulated by the engine-manage-
the microcontroller. ment system. The closed and open-loop
control functions are executed in the mi-
Pulse-type input signals crocontroller. The input signals that are
The pulse-shaped input signals from provided by the sensors and the interfaces
inductive-type sensors containing infor- to other systems (such as the CAN bus) are
mation on rotational speed and reference used as input variables. and are subjected
mark are conditioned in their own control to a further plausibility check in the com-
unit stage. Here, spurious pulses are sup- puter. The control unit program supports
pressed and the pulse-shaped signals generation of the output signals used to
converted into digital rectangular signals. control the actuators.
1 Design of a control unit using the example of an ME Motronic (sectional view through housing cover)
UAE0992Y
20 Electronic control unit Data processing
Output signals Control unit-internal communication

The microcontroller uses the output sig- In order to be able to support the micro-
nals to control output stages that usually controller in its work, the peripheral
provide enough power for connecting the components must communicate with it.
actuators directly. It is also possible to ac- This takes place using an address/data
tuate certain output stage relays for con- bus which, The microcontroller outputs
sumers that use up a great deal of power the RAM address whose contents are to
(e.g. motor fans). be read (for example) via the address bus.
The output stages are proof against The data bus is then used to transmit the
short circuits to ground or battery voltage, relevant data. For former automotive ap-
as well as against destruction due to elec- plications, an 8-bit bus topology sufficed.
trical or thermal overload. Such malfunc- This meant that the data bus comprised
tions, together with open-circuit lines or 8 lines which together could transmit
sensor faults are identified by the output- 256 values simultaneously. 65,536 ad-
stage IC as an error and reported to the dresses can be accessed using the 16-bit
microcontroller. address bus in this system. Presently,
more complex systems demand 16 bits,
Switching signals or even 32 bits, for the data bus. In order
Actuators can be switched on and off using to save on pins at the components, the data
the switching signals (e.g. motor fans). and address buses can be combined in a
multiplex system, i.e. addresses and data
PWM signals are dispatched through the same lines
Digital output signals can be in the form but offset from each other with respect
of PWM (Pulse-Width Modulated) signals. to time.
These are constant-frequency rectangular Serial interfaces with just a single data
signals with variable on-times (Fig. 2), line are used for data that does not have
Various actuators can be moved to various to transmitted extremely quickly (e.g. fault
operating positions using these signals memory data).
(e.g. exhaust-gas recirculation valve,
boost-pressure actuator). EOL programming
The extensive variety of vehicle variants
with differing control programs and data
records, makes it imperative to have a sys-
tem which reduces the number of control
2 PWM signals unit types needed by a given manufac-
turer. To this end, the Flash-EPROM’s com-
a plete memory area can be programmed
at the end of production with the program
b
and the variant-specific data record
(this is the so-called End-of-Line, or EoL,
programming).
Signal voltage
Another way of reducing the type

a diversity is to store several data variants
(e.g. transmission variants) in the memory,
b
which are then selected using coding at the
end of the production line. This coding is
UAE0738E
Fig. 2
a Period duration stored in an EEPROM.
(fixed or variable)
Time
b Variable on-time
| Performance of electronic control units 21
Control unit
▶ Performance of electronic control units
The performance of electronic control units I/O facilities for timer-controlled signals and
goes hand-in-hand with advances achieved in an integrated analog-digital converter at the
the field of microelectronics. The first gasoline end of the 1980’s. It was then possible to cre-
injection systems were still analog – with lim- ate relatively powerful systems. Figure 3 shows
ited flexibility in the implementation of control a comparison between the performance of a
functions. These functions were constrained fuel-injection system (LH3.2) and an ignition
by the hardware. system (EZ129K) – equipped with 80C515
Progress advanced in quantum leaps with controllers – and that of the succeeding
the arrival of digital technology and the micro- Motronic systems. The ME7 has approximately
controller. The entire engine management sys- 40 times the performance capability of the
tem was taken over by the universally applica- LH/EZ combination with a clock frequency of
ble semiconductor microchip. The actual con- 40 MHz. With the benefit of a new generation
trol logic in microcontroller-controlled systems of microcontrollers and a further increase in
is in a programmable semiconductor memory. clock frequency on the ME9, this figure will
From systems that initially simply con- increase to a factor of well over 50.
trolled fuel injection, complex engine-manage- In the foreseeable future microcontrollers
ment systems were then developed. They con- will process more than just digital control se-
trolled not only fuel injection but also the quences. Signal processors are integrated that
ignition system including knock control, ex- can also directly process the signals provided
haust-gas recirculation and a whole variety by knock sensors, for example.
of other systems. This continuous process of
development is bound to continue in a similar Advances in the development of semiconduc-
vein over the next decade as well. The integrator memory chips are also worthy of note.
tion of functions and, above all, their complex- Complex control programs require an enor-
ity are constantly increasing. This pattern of mous amount of memory space. The capacity
development is only possible because the of memory chips at the start of the 1980s
microcontrollers used are also undergoing was still only 8 kilobytes. The ME7 now uses
a similar process of improvement. 1-megabyte chips and soon memory capacities
Microcontrollers in the Intel 8051 family of 2 megabytes will be required. Figure 3
were used quite some time until they were re- shows this pattern of development and likely
placed with 80515 derivatives with additional future trends.
Fig. 3
Chart illustrating
▶ Performance
3 Development of electronic control units capability of
engine-management
2,500
MPC555 56MHz systems

154
160
kB
kB
50
Flash: 2.5MB ME9.0

▶ Number of control
C167 40MHz
Connector pins (control unit)
136
unit connector pins

kB
ME7.0.1 Flash: 1MB

▶
1,024 kB
C167 24MHz Program memory

132
120
kB
ME7.0 Flash: 1MB capacity

20
Performance
C167 24MHz
ME7.0 ▶ Data memory
10
Flash: 512 kB
ROM
RAM
80C517A 16MHz
capacity (RAM)
M4.4.1
128 kB
4
Flash: 128 kB
86
80C517 15.8MHz
8 kB
M4.3 Flash: 64 kB
By way of comparison:
The performance
80535 12MHz
M1.8
55
capability of a state-
32 kB
EPROM: 32 kB
SMK1930E
0.5 kB
of-the-art engine-
1
80535 12MHz
LH3.2 + EZ129K
60
EPROM: 32 kB
management system
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 far exceeds that of
Apollo 13.
22 Electronic control unit Digital modules in the control unit
Digital modules in the In a microcomputer, however, the above-

control unit mentioned functions are integrated on a
silicon wafer (system-on-a-chip). This is
Microcontroller not functional on its own (standalone) and
Structure is therefore referred to as a single chip
A microcontroller consists of the following microcomputer.
interacting components (Fig. 1): The microcontroller is used to control
• Central processing unit (CPU): this con- self-regulating systems such as an engine-
tains the control unit and the arithmetic management system. Depending on the
and logic unit. The control unit executes application, they may also have expansion
the instructions from the program mem- modules connected to them (e.g. addi-
ory, whereas the arithmetic and logic tional memory for data and program
unit performs arithmetical and logical code).
operations. The user program is fixed and is not re-
• Input and output devices (I/O, Input/Out- placed for different applications. This is
put), which handle the exchange of data the difference between a microcontroller
with peripheral devices. Peripheral de- system, for example, and a PC.
vices include input and output devices
and external data storage media. Programming
• Program memory, in which the operating The only command form capable of direct
program (user program) is permanently interpretation by a microprocessor is a
stored (ROM, PROM, EPROM or flash bit pattern, i.e. the binary representation
EPROM). of a number. Since, however, this form of
• Data memory, which is accessed for instruction is not easy to work with for a
reading and writing (RAM). This con- programmer, and is therefore susceptible
tains the data that is currently being to errors, easily memorable abbreviations
processed. Non-volatile memory (mnemonics) are used. These are automat-
(EEPROM) is used for data that must ically translated by an assembler program
not be deleted when the supply voltage into bit patterns (machine code) that can
is switched off. be understood by the microprocessor.
• The bus system connects the individual For more complex systems and pro-
elements of the microcontroller. grams, high-level programming languages
• A clock generator (oscillator) ensures such as C are needed, as otherwise it would
that all operations in the microcon- be impossible to keep extensive programs
troller take place within a defined timing manageable and free of errors. Such lan-
pattern. guages require sophisticated translation
• Logic circuits are modules with special- programs (compilers) which convert the
ized tasks such as program interrupts. text of the high-level language into a form
They are integrated in individual that can be processed by the microcon-
I/O units. troller.
The machine code is stored in the
The chief components of a microcomputer program memory, where it remains
are generally separate modules connected permanently. The CPU accesses these
to one another on a printed-circuit board. components via the bus system, reads
The microprocessor within such as system the numerically coded commands and
– the CPU – is not functional on its own: executes then.
it is always part of a microcomputer.
1
Microcomputer
Microprocessor Data memory Program memory Data memory

Central processing unit (user memory) Read-only memory Non-volatile read/
Microcontroller
(CPU) Volatile read/write (ROM, EPROM, write memory
memory (RAM) flash EPROM) (EEPROM)
for variable data For programs and
Arithmetic and logic unit permanent data records
(ALU) Memory capacity
4-, 8-, 16-, 32-bit Memory capacity Memory capacity 32 bytes to
64 bytes to 32 kbytes 2 kbytes to 512 kbytes 1 kbyte
Clock Bus 4-, 8-, 16-, 32-bit data circuit

generator
(oscillator)
I/O I/O I/O I/O I/O I/O I/O

Interrupt Event Signal acquisition Analog/ Digital Serial Bus
controller counter and output with digital inputs/ interface controller
time reference (A/D) outputs
converter (I/O) (UART,
…
(Timer, time SPI,
processing unit, CAN)
Electronic control unit
input capture,
output-compare
register)
Communication
Monitoring Resolution Resolution with external
circuit 50 ns 8 to 10 bit Data rate chips via
(watchdog) Counter Time range 4 to 32 8 to 32 200 bit/s to address/
8 to 64 bit 50 ns to 1s channels channels 1 Mbit/s data bus
… … … … … …
UAE0454-1E
Digital modules in the control unit
23
24 Electronic control unit Digital modules in the control unit
Semiconductor memories Memory modules are organized on a bit or

Applications word basis, depending on the application.
Memories are used to store large volumes A “word” is a group of bits that can be pro-
of cessed as a single unit. The word length is
• Digital signals representing data equal to the number of bits processed as
(I/O data, statuses, intermediate results a single unit. Eight bits are referred to as
involving frequent and rapid reading a byte.
and writing) Memories can be organized on the basis
• Program code (usually permanently of different word lengths. An 8 M x 8-RAM,
stored) and for example, has a memory capacity of
• Constants (permanently stored) 8 million times 8 bits (64 Mbits). The data
is organized into bytes (8 bits), making the
Storage involves memory capacity 8 Mbytes.
• Recording (writing)
• Permanent retention (actual storage) Word lengths of 4, 8, 16 and 32 bits are
and common in microcontroller systems.
• Location and retrieval (reading) of The word length is one of the factors that
information determines the performance capability of
the system. The word length that is used
Memories utilize physical effects that depends on the performance capability
clearly and easily create and show two requirements of the system.
different states (e.g. conducting/non-con-
ducting or loaded/not loaded). The infor- The most important terms are explained
mation that is to be stored must therefore below according to their standardized def-
exist in binary form, i.e. encoded as a initions, where applicable, or their most
series of “yes or no” statuses (logical “1” common usage (see Figure 2 for overview).
or logical “0”). Such a “yes or no” unit of
information is called a bit (binary digit).
2 Overview of semiconductor memories
Semiconductor memories
Non-volatile memories Volatile memories
Factory programmed User programmed
Programmable
on a program- Programmable Static Dynamic
in the circuit memories memories
ming device
UV Electrically
Read-only erasable erasable
UAE0465-1E
ROM PROM EPROM Flash EEPROM SRAM DRAM

Read- Program- Erasable EEPROM Electrically Static Dynamic
only mable read- PROM erasable RAM RAM
memory only memory PROM
Electronic control unit Digital modules in the control unit 25
Random-access memory (RAM) Erasable ROM

Random-access memory or RAM is a There are also ROMs whose contents can
short-term memory that allows direct ac- be erased and reprogrammed as outlined
cess to any storage location. Information below.
can be read/written from/to the memory
any number of times. EPROM (Erasable PROM)
This type of erasable read-only memory
Static RAM (SRAM) can have its contents completely wiped
Static RAMs use bistable switching ele- by irradiation with UV light and can then
ments as the data storage cells. Their func- be reprogrammed using a programming
tionality is similar to that of a flip-flop, device.
a simple circuit with two transistors, of
which either the one (logical “1”) or the EEPROM (Electrical EPROM)
other (logical “0”) conducts at any one The EEPROM (also known as E2PROM) can
time. In SRAM, the information remains be electrically erased and reprogrammed.
stored until the storage cell concerned Every storage cell of an EEPROM be indi-
is addressed and overwritten, or the vidually overwritten. For that reason,
operating voltage is switched off. this type of memory module can also
SRAM is therefore volatile memory. be used as nonvolatile data memory
(e.g. for learned information in engine
Dynamic RAM (DRAM) management systems).
Unlike SRAM, the information is stored
as an electrical charge in the gate capacity Flash EEPROM
of a CMOS transistor in dynamic RAM A more sophisticated variant of the
(DRAM). As such capacitors are suscepti- EPROM and EEPROM is flash EEPROM.
ble to leakage, the charge is gradually lost. In this case, electrical flash pulses are used
In order to retain the information, the to erase specific storage areas or the entire
charge has to be refreshed at regular contents of the memory. The erased areas
intervals (every few ms). can subsequently be reprogrammed.
The flash memory can be reprogrammed
Read-only memory on a programming station. However, the
Read-only memory (ROM) is permanent- advantage of flash EEPROM is that is can
storage memory that allows any memory also be reprogrammed while still inside
location to be accessed directly but – as the the sealed control unit.
name indicates – allows the information Flash EEPROM is used in cases where
only to be read and not modified. relatively large quantities of data need
A ROM is nonvolatile memory, i.e. the to be stored, but must also be modifiable
information it contains is retained even (e.g. program memory in vehicle control
when the operating voltage is switched off. units).
It is usually used to store program code
(control programs) and fixed data (func-
tion tables, encoding rules, engine charac-
teristic data maps) that need to be retriev-
able at any time. The information may
be indelibly entered in the memory by
the manufacturer or the user by means
of appropriate programming of specially
prepared memories (PROMs or program-
mable ROMs).
26 Electronic control unit Control unit software
Control unit software execute the command depends on the

microcontroller that is used and the clock
Real-time capability frequency. The microcontrollers that are
One of the requirements on electronic sys- currently used in vehicles can execute up
tems is real-time capability. This means to 1 million commands per second.
that control procedures must react to in- Because of the limited speed at which
put signals within an extremely short time. the program can be executed, a software
For example, a wheel that has a tendency structure with which time-critical func-
to lock must be detected so quickly that tions can be processed with high priority
the ABS control algorithm in the control is required.
unit can reduce the brake pressure via The engine-management program has
the hydraulic modulator quickly enough. to react extremely quickly to signals from
Brake slip is therefore reduced before the speed sensor, which records the en-
the wheel can lock. Engine management gine speed and the crankshaft position.
systems make considerable real-time ca- These signals arrive at short intervals that
pability demands so that crankshaft angles can be a matter of milliseconds, depending
can be adhered to with extreme accuracy on the engine speed. The control unit pro-
at fast engine speeds for injection and igni- gram has to evaluate these signals with
tion timing purposes. high priority. Other functions such as
The complexity of an electronic system reading in the engine temperature are not
therefore makes extremely high demands as urgent, since the physical variable only
of the software that is developed. The soft- changes extremely slowly in this case.
ware structure is explained below on the
basis of an example. Interrupt control
As soon as an event occurs that requires
Software structure an extremely rapid response (e.g. speed
The microcontroller in the control unit ex- sensor pulse), the program that is cur-
ecutes commands sequentially. The com- rently running must be interrupted. This
mand code is obtained from the program can be done using the microcontroller’s
memory. The time taken to read in and interrupt control facility. Events can trig-
ger a program execution interrupt, where-
upon the program jumps and executes
1 CPU power distribution principle the “interrupt routine”. When this routine
has been executed, the program resumes
Tooth at the point at which it was interrupted
interrupt
(Fig. 1).
An interrupt can be triggered by an ex-
Synchro
interrupt ternal signal, for example. Other interrupt
sources are timers integrated in the micro-
Ignition controller, with which timed output signals
interrupt can be generated (e.g. ignition signal:
microcontroller ignition output is switched
at a point in time that is calculated before-
Time frame
hand). However, the timer can also gener-
ate internal time frames.
Fig. 1 Background
Depiction of several program
SAE1003E
program levels on the Time t

example of the software
from a Motronic system
Electronic control unit Control unit software 27
The control unit program reacts to several to the speed sensor signal. For this pur-
of these interrupts. An interrupt source pose, the engine-speed signal is connected
can therefore request an interrupt while to a microcontroller interrupt input. Every
another interrupt routine is currently falling signal edge at this input interrupts
being executed. Every interrupt source the current calculations that are in prog-
therefore has a fixed priority assigned to ress and forces a branch to the interrupt
it. The priority controller decides which routine. After executing the commands in
interrupt is allowed to interrupt another the interrupt routine, the program contin-
interrupt. ues execution at its point of origin.
In order to perform certain operations
Tooth interrupt the control unit program requires the time
The crankshaft is equipped with a pulse taken for the crankshaft to travel between
wheel (Fig. 2a) that has a certain number one tooth and the next. This calculation is
of teeth on its circumference. The teeth performed by an internal timer. This is a
are scanned by the speed sensor. This freewheeling 16-bit counter (Fig. 3) that
allows the crankshaft position to be re- increments at a certain rate, depending on
corded. The typical distance between the microcontroller oscillator clock cycle.
a pair of teeth on the crankshaft sensor This time frame amounts to about 0.5 µs.
wheel is 6°. In order to determine the When the falling tooth flank occurs,
crankshaft position, the control unit pro- the current counter status is recorded.
gram must execute certain routines as The difference (and therefore the tooth
each tooth is detected. At 6,000 rpm the interval) is calculated using the stored
detection time between two teeth is ap- counter from the previous tooth.
proximately 300 µs. Every command in
these routines must be executed within Example: crankshaft position calculation
this time. This requires a rapid response The engine-management system (Motronic
for gasoline engines, EDC for diesel en-
gines) must know the crankshaft position
2 Crankshaft sensor ring with speed sensor at any given point in time. This is a pre-
requisite for injecting into the right cylin-
a
der at the right time and ensuring that ig-
nition takes place at the calculated ignition
angle (Motronic systems). In order to
3 Time calculations via internal timer

Counter status
Tn+1
Tn
Time t
Speed sensor
b
signal
Digital
signal
SAE1004E
SAE1005E
Tn: Counter status at tooth n Fig. 2

Tn+1: Counter status at tooth n+1 a Design
b Speed sensor signal
28 Electronic control unit Control unit software
detect the engine position and the engine Since the synchronization program runs
speed, the control unit evaluates the speed over several teeth at fast engine speeds,
sensor signal (Fig. 2b). it has to be interrupted by the tooth inter-
There is gap in the crankshaft sensor rupt. The tooth interrupt is given higher
wheel in which two teeth are missing. priority than the synchronization pro-
The tooth space has a defined position gram.
in relation to the top dead center (TDC) of
cylinder no. 1. The control unit program Ignition interrupt
has to synchronize itself with this tooth The ignition output takes place within a
space. This is done by measuring the times certain crankshaft range, depending on
between two consecutive falling tooth the value from the ignition map. Since the
flanks. The time for the tooth space is specified ignition angle has to be adhered
considerably greater than the time before to exactly, the ignition output is controlled
and after the gap. Following a “short – by an interrupt. Like the synchronization
long – short” sequence the last thing to program, the ignition interrupt is also
be scanned was the falling flank of the called up once per combustion cycle.
second tooth after the space. The control unit program is aware of the
The crankshaft has rotated by 6° for crankshaft position in the 6° framework.
each falling tooth flank that has been de- However, this framework is not accurate
tected by the control unit program. This is enough for ignition angle output. For this
how the control unit program knows the reason, accurate ignition output between
crankshaft position within this time frame. two teeth must take place as well as this
Since cylinder no. 1 is in the tooth space approximate counting for the last 0 to
position in the vicinity of top dead center 6 crankshaft degrees. This is done using
(TDC) or bottom dead center (BDC), an ad- a timer (Fig. 5). Ignition angle output that
ditional signal is required to determine the was purely timer-controller would lead
position. The camshaft sensor provides to an ignition angle output error at high
a different voltage level in both cases. engine speed dynamics.
The control unit is therefore able to Firstly, the ignition coil must be enabled
uniquely assign the crankshaft and for a defined time (the so-called dwell pe-
camshaft positions. riod). In order to do this, the program cal-
culates the switch-on time by calculating
Combustion-synchronous interrupt
Some calculations have to be performed 4 Triggering the interrupt synchronously with
for every combustion cycle. For example, combustion
the ignition angle and the injection have
to be recalculated synchronously with
combustion for each cylinder. The pro-
gram does this by branching to the “syn-
Start synchronization
chronization program” after certain teeth
program
(Fig. 4). This interrupt takes place after
every 30 teeth (ignition interval) for a four-
Tooth counter elapsed
cylinder engine, and after every 20 teeth Synchronization program triggered
for a six-cylinder engine.
The synchronization program is fixed to Second tooth after the gap: reference mark
a certain tooth position and has to be exe- Preset tooth counter for triggering the
SAE1006E
cuted with high priority. For this reason it synchronization program

is activated via an interrupt (triggered by a
command in the tooth interrupt routine).
Electronic control unit Control unit software 29
backwards from the ignition angle at which 5 Dwell and ignition time output
the ignition coil has to be switched off.
This makes it possible to calculate the
tooth after which the ignition coil has to
be switched off (approximate counting Ignition output
in 6° time frame). The remaining angle
(detailed counting 0 to 6°) is converted Switch on the ignition coil
Enter the time for ignition
into an output time using the current en-
gine speed. As soon as the specified tooth Enter time for detailed counting
position has been reached using approxi-
mate counting, a time is loaded with the Synchronization program:
SAE1007E
Calculate the dwell period and ignition angle
output value from the detailed counting.
Preset the tooth counter for the dwell period
When this time period expires, the timer time (approximate counting)
triggers an interrupt. The commands that
switch on the ignition coil are programmed
in this interrupt routine. Then the timer Background program
is preset to the dwell period value, which All other activities that do not run in
causes an interrupt to be triggered when an interrupt routine or a time frame are
the timer elapses, switching off the igni- processed in the background program.
tion coil and therefore initiating ignition. At fast engine speeds, the synchronization
program and the tooth interrupt are called
Time frame frequently, leaving little CPU time for the
Many control algorithms have to run background program. The time taken for a
within a certain time frame. Lambda con- complete run-through of the background
trol, for example, has to be processed program therefore increases rapidly with
within a fixed time frame (e.g. 10 ms) so the increasing engine speed. The back-
that the correcting variables are calculated ground program must therefore only
quickly enough. contain low-priority functions.
30 Electronic Control Units (ECUs) Software Development
Software Development QA2F

When: After function implementation
A study of the current series projects together What: Review of each individual function
with the use of development capacities and checking of following docu-
demonstrate that approximately 60% of the ments:
time spent on ECU development has to be 쐌 specification,
devoted to creating the necessary software. 쐌 function description,
For this reason, it is absolutely essential that 쐌 source code,
modern tools and processes be used. 쐌 data definitions, and
쐌 test documentation.
Development Process
Definition of Development Process QA2
A depiction of the development steps in the When: Prior to software delivery
form of a V-model (Figure 1) serves as the ba- What: Review of all QA2F documents
sis for all software-development activities.
This model is used to detail the process steps QA3
which facilitate implementation within a When: Prior to start of series production
product-development department. What: Series-production review for hard-
ware and software
Quality Assessment
Quality assessments are scheduled at defined An essential part of the development process
points of the development process (Figure 2) is also the distinction between specification
for the purpose of process monitoring: and implementation. This separation allows
programming by contract, whereby project
QA1 teams use the software knowledge of ReUse
When: At the start of the project teams who implement the functions (e.g.
What: Resource check (capacity, develop- Keyword 2000 protocol) for various cus-
ment environment, responsibility) tomers. To this end, the project teams write
out function contracts which establish the
QA1F boundary conditions of implementation. The
When: Prior to function implementation scopes and levels of testing of the individual
What: Function specification check functions are determined in the project-spe-
cific quality plan (PQSP) with the customer
requirements in mind.
1 Simplified V-model 2 Development process in detail
System System
Specification Test initialization delivery
Function analysis QA2
specification
System
QA1 integration/test
Function Function
initialization delivery
Function analysis QA2F
specification
æ UTS0325E
æ UTS0326E
Function
QA1F integration/test
Function
Implementation development
Electronic Control Units (ECUs) Software Development 31
This also includes determining the QA x F Programming Guidelines

scopes. In any heterogeneous system of development
that is spread over countries and continents, a
The PQSP is a central element of project im- standardized procedure for creating software
plementation and should be fully discussed is a vital and integral part of the time to mar-
between vehicle manufacturer and supplier. It ket process. These guidelines address the fol-
lays down, among others things, the responsi- lowing points and are binding for all pro-
bilities, the customer relationships, the devel- grammers:
opment tools, the scopes of testing and docu-
mentation, etc. 쐌 general guidelines (terminology, vocabu-
lary, variation handling),
The process in the ReUse team is set out as 쐌 guidelines for software developments in C
follows: (templates, structure),
쐌 definitions and declarations (include,
defines, typedefs),
쐌 The person in charge of the project
쐌 check instructions (if, for, while, break,
formulates the task (if necessary,
return ...),
adoption of customer request).
쐌 coding specifications and instructions
쐌 The function contract is drawn up
(typecast, arithmetic, pointer),
(with details of task, project, desired
쐌 particular features when using variables
date, scope, and reference documents)
(alignment, address),
and given to the ReUse teams.
쐌 instructions on data consistency
(preemptive), and
쐌 instructions on resource relief.
쐌 The persons in charge of ReUse and These guidelines also serve as a source of
the project discuss the application and knowledge for effective code configuration in
establish the scopes and deadlines order to counteract the limitations in relation
together. to memory capacity and run time in the pro-
gramming of microcontrollers.
쐌 The person in charge of ReUse decides

on the variant and version handling
and implements the task.
쐌 All documents are handed over to the
person in charge of the project on
completion.
In order to ensure that the created software

can be reused to the greatest possible extent,
there are C-programming guidelines binding
an all programmers which are called up in the
relevant reviews (e.g. QA2F).
Tools for Creating Software As the wide variety of tools demonstrates, the
As well as the formal aspects such as process process involved in creating the software for
and programming guidelines, it is crucially an ECU of the latest generation is highly
important to ensure that the tools are subject complex. Figure 4 provides a simplified
to constant support in the interests of prod- overview of the interplay between the indi-
uct quality. Figure 3 provides an overview of vidual tools from the specification through to
the tools currently used for the various devel- the finished ECU program.
opment phases. Significant features of this
tool chain are: By way of example, two component parts of
쐌 constant support throughout the entire de- the tool chain will now be explained in closer
velopment process and detail:
쐌 product-specific, optimized solutions with 쐌 Design with ASCET-SD and
tools partly developed in-house. 쐌 Vehicle simulation with TCM-Simutec.
3 Tools in the development process 4 Simplified process sequence
Organization: MS Project Design

ASCET-SD Function model
Proto- Implemen- Object model
Design Test
typing tation Prototyping
ASCET- ASCET- ESPRIT INCA-PC ASCET-SD Source files
SD SD Innovator TCM-
StP Codewright LabCar Code generation
DAMOS++ ASCET-
ASCET-SD LabCar ESPRIT ETC program
æ UTS0327E
Test/application
Documentation: MS Word
INCA/PC
Key:
ASCET: advanced simulation and control
æ UTS0328E
engineering tool
ASCET-SD: ASCET software developer
StP: software through pictures
(Aonix) for OO modeling
ESPRIT: engineering software-production
5 Function design with ASCET-SD
user interface for tools
Innovator: Software-development
environment (MID)
Codewright: Software-development
environment (Premia)
DAMOS: database for microcontroller-
oriented systems
INCA-PC: integrated car application system
TCM-Simutec: Vehicle simulator
æ UTS0329Y
ASCET-LabCar: Vehicle simulator for HiL simulation

ClearCase: Configuration-management tool
Design with ASCET-SD Vehicle simulation with TCM-Simutec

ASCET-SD (Figure 5) offers the following To test the functions of a transmission-
functions for designing software: control system in the laboratory, there is a
쐌 interactive creation of function simulator for the vehicle and transmission
descriptions and function models, environment which provides the input signals
쐌 a graphical user interface, for ETC. One such simulator is shown in Fig-
쐌 support of object-oriented design, data- ure 7.
flow-oriented design and state machines. The front panel of the simulator is
equipped with assorted rotary potentio-
The ERCOSEK operating system is an integral meters, switches, and pushbuttons which en-
part of the development environment, and fa- able input variables such as output speed, se-
cilitates real-time simulation of the function lector-lever position, transmission
model. temperature, etc., to be specified.
The top of the simulator accommodates a
ASCET-SD offers the following support for breakbox which permits access to every ECU
the rapid prototyping of functions: pin. Measuring instruments can be easily
쐌 ASCET-SD operates as a bypass computer connected to these sockets to enable, for
for the series ECU, i.e., individual ECU example, a PWM signal of a pressure-regula-
functions run on the PC, while the other tor output to be viewed on an oscilloscope.
functions continue to be executed by ETC The laboratory car also contains computer
(Figure 6). cards which simulate the other ECUs in the
쐌 The connection is established via the CAN vehicle network (e.g. ECUs for engine man-
or the INCA probe. agement, for ABS, etc.) and also their signals.
The next step is the automatic C-code creation Process and Maturity Model
and the creation of the corresponding data A clear definition of the development process
files for the application from the models. and the corresponding implementation in the
projects are made possible by a software de-
For further information, log on to velopment which can be evaluated with a ma-
turity model such as CMM (capture maturity
http://www.etas.de model).
6 Test setup for ASCET-SD in bypass 7 TCM-Simutec (laboratory car)
2 4
1 Fig. 6
1 ASCET-SD and
3 INCA-PC
2 ASCET hardware
æ UTS0330Y
æ UTS0311Y
(ETAS ES 1000.2)
3 ETK
4 ETC-Simutec
(laboratory car)
8 Software layer model The operating system with its services and the
hardware-compatible software are imple-
Transmission software from mented on this hardware:
vehicle manufacturer or Bosch
ERCOSEK EEPROM Hardware KWP 2000
(OS) driver Input/Output driver
Program
library Device Device Device
driver driver driver
Component
driver The interface layer and program library for
the application software contain:
Operating system Diagnosis Diagnosis Security
handling monitoring software
æ UTS0331E
functions (SSK)
Hardware EEPROM KWP2000 Shift by wire
handling application functions
The application software

Software Structure (customized software) comprises:
The software structure described in closer de- e.g. ASIS (RB/ZF) Transmission software
tail in the following is implemented within AGS (BMW)
the transmission-control system. This layer Software sharing (interface)
model (Figure 8) comprises
쐌 the application software (transmission soft-
ware) with program library provided by Operating System
the vehicle manufacturer or supplier (in It is absolutely essential to use an OSEK-con-
this case by Bosch), forming operating system to fulfill the cur-
쐌 the operating system, rent real-time demands on an ECU. The ER-
쐌 the component driver, and COSEK operating system from ETAS is used in
쐌 the hardware. Bosch transmission control units (available
for all kinds of microcontroller).
The separation of hardware and application
software ensures that the software can be eas- An operating system is subdivided into
ily ported to new hardware platforms. Only processes and tasks (Figure 9):
the second layer, consisting of the operating A process is a function which has no call or
system and component driver (BIOS), has to return parameters.
be adapted. A task consists of different processes and is
The contents of the individual layers will characterized by
now be broken down in the following: 쐌 the sequential execution of processes,
쐌 the allocation of processes 씮 task,
The ECU hardware, the first layer of the soft- 쐌 each task being assigned a priority,
ware layer model, consists of the microcon- 쐌 tasks being assigned to a time base.
troller (here, by way of example, the
MPC555), the memory, the interfaces (SPI, For task changing, there is either cooperative
CAN and UART), and the peripheral chips scheduling or preemptive scheduling (task
(ASICs): management):
Cooperative Scheduling
CPU core SPI TPU MIOS CAN UART In the case of cooperative scheduling,
timer and (2) (2)
a task can only be interrupted between two
Memory, hardware, driver, etc.
processes by a higher-priority task (Figure 10).
The advantages of this procedure are low and a response time that is not dependent on
memory requirement (register banks, stack), process implementation. The disadvantages
simple management, and data consistency. are increased memory requirement (stack,
The disadvantages are the limited response register banks) and data-consistency prob-
time (dependent on the process run time) lems.
and the jitter over the task period.
Mixed Scheduling
Preemptive Scheduling ERCOSEK offers the option of mixing both
Owing to the drawbacks of cooperative types of scheduling in one application.
scheduling, preemptive scheduling is used in A combination of hardware and software
operating systems which operate as real-time scheduling serves this purpose. Figure 12
systems. shows the distribution between cooperative
With this form of scheduling, a higher-pri- and preemptive using the priorities assigned
ority task can interrupt a lower-priority task to the tasks.
at any time (Figure 11). The advantages of A software call starts the operating system.
this procedure are the very short response It can support different application modes
times, the minimal jitter over the task period, (e.g. different task sets for initialization, oper-
9 Processes and task 11 Preemptive task change
Task Activation
and start
Task B
Task B
Process 1
Task
p1B p2B p3B p4B

Process 2
Process 3
…
p1A p3A p4A

æ STS0332E
æ STS0334E
Prozess n Task A
Time t
10 Cooperative task change 12 Priority distribution
Activation Task B
Hardware-based
Start Task B
scheduling
Task B preemptive
Priority
Task
p1B p2B p3B p4B

Software-based
scheduling
p1A p2A p3A p4A

æ STS0333E
æ STS0335E
cooperative
Task A
Time t Distribution
13 Application-mode change ation and ECU run-on, Figure 13). Each ap-
plication mode consists of an initialization
phase and an execution phase. Interrupts are
Application
prohibited during initialization of an applica-
Mode n Mode n+1 tion mode.
Further documents on the subject of
ERCOSEK / OSEK can be found on the
Mode
Internet at:
Init Execution Init Execution http://www.etas.de
http://www.osek-vdx.org
æ STS0336E
Acquisition of Input and
Output Variables
Zeit t
Access to the hardware is obtained within the
framework of the software layer model in ac-
14 Hardware access in the layer model cordance with three layers (Figure 14):
쐌 user layer,
쐌 configuration layer, and
User layer 쐌 hardware layer.
Access to hardware capsule: The first implementation example to be

featured is access (A) via global RAM cells.
via global as direct access Table 1 describes the name of the RAM cell,
RAM cells (function call)
the signal direction (input or output), the sig-
nal type (analog, digital, frequency, or PWM),
the description, and the physical conversion.
The second example featured is access via
(B) function interfaces. Table 2 is structured
Configuration layer along the same lines as Table 1, but here the
global RAM cell is replaced by a function call.
Configuration: Filtering:
Access to Elimination of The objective in the configuration layer is
hardware malfunctions to obtain independence from platform and
channels
project in the conversion of the hardware ac-
Coherence: Scaling: cesses into real software. This is achieved on
Data timing Conversion
into physical the one hand by using tools which automati-
variables cally create the C-code for access to the hard-
ware, and on the other hand by using C-
macros which are then resolved on a proces-
Hardware layer
sor-specific basis.
ADC DIO PWM SPI Serial Freq

æ STS0337E
"Low-level"-channels
1 Hardware access via global RAM cells
RAM cell I/O Type Content Scaling

ugt_Batt In ANA 16 bit Battery voltage 0...25 000 mV
CGT In ANA 8 bit Oil temperature –40...+215°C
ccu_Chip In ANA 8 bit Substrate temperature –40...+215°C
fgt_Fet Out DIG 8 bit Status HSD-Fet 0.. 1 (On/Off)
fpo_L1 In DIG 8 bit Selector lever pos 1 0..1 (On/Off)
fpr_PinM In DIG 8 bit M button 0..1 (On/Off)
NAB In FREQ 16 bit Output speed 0...20 000 rpm
NAB32 In FREQ 8 bit Output speed/32 0...255 rpm/32
NTU In FREQ 16 bit Turbine speed 0...20 000 rpm
NTU32 In FREQ 8 bit Turbine speed/32 0...255 rpm/32
hmv1 Out PWM 16 bit Solenoid-valve output 0...1000 per mil
idr1s Out ANA 16 bit Nominal current 0..12 000 mA
pressure regulator
Table 1
2 Hardware access via functions
Software function I/O Return value Content Scaling

GetHWIO_U_IgnRunCrnk() In ANA Battery voltage 0...32 V
16 bit
GetHWIO_T_TransOil() In ANA Oil temperature –40...+215°C
16 bit
GetHWIO_b_HSD() In ANA Status HSD-Fet 0...1 (on/off)
8 bit
GetHWIO_e_TapUpDwnReq() In ENUM Tip (+/–) function 0 x 00...0 x 40
8 bit
TsHWIO_PRNDL In DIG Transmission control panel 0...1
GetHWIO_s_PRNDL(void) 8 bit
TsHWIO_FreqParams In FREQ Turbine speed Time stamp +
GetHWIO_s_NTU(void) struct counter value
TsHWIO_NAB_DualEdgeParams In FREQ Output speed, Time stamp +
GetHWIO_s_NAB_DualEdge() struct edge can be changed over counter value +
operating edge
SetHWIO_e_NAB_DualEdgeCptr Out – nout Rising, falling, both
Mode( BYTE ) Edge changeover Table 2
A further signal-acquisition component 15 ECU network
involves the exchange of signals via the a

communication interface. The CAN bus
(controller area network) has gained
acceptance in this field in the last few years. ECU 1 ECU 2 ECU 3 ECU 4
CAN replaces the conventional wiring

harness or the previously standard network of
ECUs (Figure 15). The bus system must b
satisfy the following requirements here:
Real-time updating for
safety functions: 10 ms ECU 1 ECU 2 ECU 3 ECU 4
Convenience functions: 10...100 ms
Maximum cable length: 40 m
æ STS0338E
CAN - Bus
Fig. 15
a Conventional
b With CAN
3 Bit rates as a function of cable (bus) length The system must also be resistant to tempera-
Maximum bit rate Bus length
ture and moisture. The CAN bus has also
kbit/s m gained acceptance in the field of automation
1000 40 technology. Table 3 lists the maximum possi-
500 100 ble data rates for different cable lengths.
250 250 Figure 16 shows the circuit-engineering
125 500
implementation of the CAN interface in an
40 1000
ECU.
Table 3
16 CAN interface In the microcontroller itself, message

handling is conducted via a dual-port RAM
Microcontroller (Figure 17). Since this RAM chip, as the name
suggests, can be described from two sides
CAN controller (CAN transceiver and microcontroller), the
CPU workload is substantially relieved for
TX0 TX1 RX0 RX1
signal transfer.
TxD RxD Ref Rs +6V
VCC Complete arbitration (message organization,
CAN transceiver 100nF who sends what when) on the CAN bus is
Gnd
CAN_L CAN_H
performed automatically by the CAN trans-
ceiver. It does not require any computing
CAN_H
power in the microcontroller (Figure 18).
A distinction is made within CAN messages
æ UTS0339E
RT CAN-BUS RT between standard and extended data frames

BUS CAN_L BUS (Figures 19 and 20).
termination termination
17 Use of dual-port RAM with the CAN bus 19 CAN standard data frame
Standard Data Frame Inter Frame Space
recessive
Message 1 low high
1 11 1 1 1 4 0…64 15 1 1 1 7 3
Message 2 dominant
Bus idle
Start of Frame
RTR Bit (D)

IDE Bit(D)
(reserved (D))
Data Length Code
Identifier Field
Data Field
CRC Sequence
CRC Delimiter
ACK Slot
ACK Delimiter
End of Frame
Intermission
CAN-
Bus CPU
Message n
æ UTS0340E
æ UTS0342E
workload
Acceptance Message Host CPU
filter management
Arbitration Control CRC Acknowledge
Field Field Field Field
18 CAN-bus arbitration 20 CAN extended data frame
Bus idle recessive

Extended Data Frame Inter Frame Space
NODE A recessive
dominant 1 11 1 1 18 1 2 4 0…64 15 1 1 1 7 3
dominant
Bus idle
Start of Frame
Identifier
SRR Bit (R)
IDE Bit(R)
Extended Identifier
RTR Bit(D)
(2 reserved (D))
Data Field
CRC Sequence
CRC Delimiter
ACK Slot
ACK Delimiter
End of Frame
Intermission
Bus idle
Data Length Code
recessive
NODE B
æ UTS0343E
æ UTS0341E
dominant
NODE B loses the arbitration
switches to receive Arbitration Control CRC Acknowledge
mode Field Field Field Field
The standard data frame is characterized by Object-Oriented Approach

the following data (Figure 19): Vehicle Control
Data capacity: 0...8 bytes As an introduction, Figure 21 shows the 1-2
Identifier length: 11 bits US and 2-1 DS shift curves for a driving pro-
Message length: max. 130 bits gram.
The shift curves shown in Figure 22 extend
In contrast, the extended data frame this system for different driving programs
(Figure 20) has the following from super economy (XE) to super sport
characteristic data: (XS). They clearly show that the upshift point
Data capacity: 0...8 bytes in the sporty driving program moves towards
Identifier length: 29 bits higher vehicle speed or higher engine speed
Message length: max. 150 bits and thereby achieves optimum utilization of
engine performance.
Gear Selection and Adaptive Functions
Transmission control has undergone various
phases or expansion stages within the frame-
work of development. 21 1-2 US shift curve
The basic functions and the adaptive

%
programs for shifting points and pressure
100
control are now standard in the field of elec-
Accelerator-pedal position
tronic transmission control (ETC). Where

2-1 RS
S
1-2 H
the various marques and vehicles differ is in

the different strategies employed in the auto-
matic adaptation of the transmission to the 50 1
driving style and the traffic situation. This
also represents an area of software which is
being increasingly taken up by the vehicle
manufacturer directly and which is no longer
in the hands of a supplier. The adaptive func-
æ STS0344E
0
tions for shifting-point control and pressure 0 30 km/h
control have already been discussed in the Vehicle speed υF
chapter sections entitled Shifting-Sequence
Control and Adaptive Pressure Control.
The following text will now deal with 22 1-2 curve with several driving programs
Bosch-specific implementations of automatic

adaptations (learn functions). The adaptive %
shifting strategy determines a gear cyclically 100
from the driver command, the vehicle status,

2-1 RS XS
2-1 RS XE
S XS
S XE
and the driving situation. It is adaptive in re-

1-2 H
1-2 H
lation to the driver type (sportiness) and also

takes into account automatic or manual gear 50 1
preselections (tip/nudge operation, as is fa-
miliar, for example, from Porsche Tiptronic©
or BMW Steptronic©). The complete software
package has been modeled with an object-
oriented approach for optimum reuse.
æ STS0345E
0
0 50 km/h
Vehicle speed υF Fig. 21 and 22
1 Upshift
To select a shift curve using the driver type On the other hand, there are the variables
and the total running resistance, it is necessary which are crucial to shifting-point selection:
for them first to be recorded and evaluated 쐌 driver type,
once. The overall structure of vehicle control 쐌 driving situation, and
shown in Figure 23 serves this purpose. 쐌 driving program.
On the one hand, there are the variables This is made clearer with the newly
which determine the vehicle and its status: arranged graphic overall structure of gear
쐌 transmission control panel (TCP), selection in Figure 24. Each of these variables
쐌 the transmission itself, is then further divided into different sub-
쐌 the engine, evaluations.
쐌 the accelerator-pedal position, and
쐌 the vehicle variables (e.g. vehicle speed, The driver-type evaluation specifies whether
wheel speed, etc.). the current driving style is economical or
sporty. The driver-type determination as
23 Overall structure of vehicle control featured in Figure 25 can be shown for this
purpose.
Determine Cyclical ()
The result of the driver-type determination is
a driver-type counter (Figure 26) with an allo-
Determine status () Determine driver type () cated driving program (XE to XXS).
Transmission Driver type
control panel
Driver-type evaluation is followed by hill
Engine Determine drive situation () recognition (based on the vehicle running
Transmission
Driving situation resistance), which distinguishes between
different types of uphill and downhill driving
Vehicle variables Set transmission () with the following allocation (Figure 27):
Drive program B0 Downhill 2
Accelerator pedal
B1 Downhill 1
æ STS0346E
B2 Level
Vehicle control
B3 Uphill 1
B4 Uphill 2
24 Overall structure of gear selection 25 Driver-type determination
Determine driver type ()
Vehicle
variable Determine driver type () Give driver type ()
Accelerator Administrator Determiner
pedal
Transmission Driving
control panel situation Startup evaluation
Gear selection
SESP
Engine Driver type
Gradient evaluation
Electronic Kickdown evaluation
stability program Linear-acceleration Counter
evaluation
Display Special evaluation
æ STS0348E
æ STS0347E
Lateral-acceleration
Transmission
evaluation
Driver-type recognition
This is followed by determination of the Summary

driving situation (Figure 28). The following features can be summarized for
the object-oriented adaptive driving strategy:
Ratio Criteria (RC) 쐌 cyclical determination of the gear,
Driver type and driving situation form one of 쐌 consideration of driver command,
three ratio criteria (RC), all of which demon- vehicle status, driving situation,
strate different functions: 쐌 adaptive in relation to driver type
(sportiness),
RC Driver Type 쐌 division into static and dynamic driving
RC driver type makes a gear suggestion using programs,
a shift curve, depending on the relevant dri- 쐌 automatic and manual gear preselection
ver type. The driver thus serves as the (tip/nudge operation),
ratio supplier (RS). 쐌 new, flexible prioritization procedure, and
쐌 object-oriented structure.
RC FastOff
RC FastOff prevents upshifts if FastOff has
been detected, i.e., a
shift prevention (USP, DSP) takes place. 27 Hill recognition
RC TCP
RC TCP changes the order of priority of the
ratio supplier in accordance with operation of B2
the transmission control panel (TCP). An RC B3
B1
selection adapted to the B4
shift situation (SS) B0
takes place.
B1
B2
æ STS0350Y
26 Driver-type counter 28 Driving situation
Determine drive situation ()

Startup evaluation
Gradient evaluation
Lateral-acceleration evaluation Coordinator
"Kick Fast" evaluation
Special evaluation
Determine status ()
Driver-type counter Hot mode Curve

0.....255
Urban driving Winter
Variable Hill
shifting sequence
æ STS0349E
æ STS0351E
Fast Off Warm-up

XE E S XS XXS
Drive program (characteristics) Driving situation
Diagnostic Functions 쐌 contains filters, time stamps, and assorted

As well as the actual transmission-control flags,
functions, monitoring functions are taking up 쐌 controls the internal program run via the
an ever-increasing percentage of the software. flags.
At present, approximately 30% of the soft-
ware stored in the ECU is assigned to Backup Fault Memory (BFM)
diagnostics. The backup fault memory (BFM) exhibits the
Full diagnostics serves to enhance driving following features:
safety and also with the aid of substitute 쐌 optional use,
functions to increase system availability. 쐌 ring memory, saved in the non-volatile
Diagnostic management must make the memory,
following functions available: 쐌 typically 5 to 10 memory locations,
쐌 management of the fault memory, 쐌 contains fault entries which have been
쐌 CARB fault management, applicable to deleted from the PFM.
each individual fault code (only specific
faults alter the vehicle’s exhaust characteris- Snapshot Memory
tics and are therefore relevant to CARB), The snapshot memory is optional and con-
쐌 filtering (time- and event-controlled), tains further environmental conditions for
쐌 requirement of measures (substitute the first PFM entry.
functions, limp-home),
쐌 provision of data for the application of The entries in this fault memory can be read
diagnostic management for each individual out in the workshop with the diagnosis tester,
fault type, and provide important information relating
쐌 monitoring functions call up diagnostic to vehicle repair.
management; this therefore only has a few
calls in the task lists. Monitoring Functions
The following text deals with the most
Fault Memory important monitoring functions in a
The fault memory is divided into different transmission-control system:
subsections, which are handled differently in
relation to the memory location and accord- Solenoid-Valve Monitoring
ing to the input of the “ignition off ” signal. The following conditions apply to monitor-
ing a solenoid valve in the transmission:
Primary Fault Memory (PFM) 쐌 activation in PWM mode,
The primary fault memory (PFM) exhibits 쐌 alternatively on/off function
the following features: (100%/0% PWM),
쐌 saved in the non-volatile memory 쐌 PWM with fundamental frequency 1 kHz,
(EEPROM), 쐌 pulse-to-no-current ratio dependent on
쐌 typically 10 memory locations, the supply voltage,
쐌 contains fault code/type, environmental 쐌 jump from approx. 95% to 100% in PWM
conditions, CARB/“warm-up” counter, and mode,
flags. 쐌 analog feedback for diagnosis, separate for
“on” and “off ” states,
Secondary Fault Memory (SFM) 쐌 general test condition:
The secondary fault memory (SFM) exhibits battery voltage ≥ threshold (7 V),
the following features: 쐌 fault detection as per Table 4.
쐌 one memory location for each fault code
(only in the volatile RAM),
4 Solenoid-valve monitoring a PRC section must be followed by a “com-

PWM (%) SCGO SCG O SCP
plete check” which generates the relevant
0 – HW HW –
PRC part answer with the aid of correction
0...5 – SW SW – values. The part answers are XORed and form
5...95 – SW SW SW the complete part answer of the PRC to the
95...100 – – – SW question/answer communication. The cur-
100 HW – – HW rent question serves as the input variable into
Key: the PRC (or, more precisely, into each of the
PWM Pulse width modulation
SCGO Short circuit to ground or
PRC parts), i.e., the program run is moni-
open circuit tored on a question-dependent basis. A
SCG Short circuit to ground dummy must simulate the PRC-monitored
O Open circuit
SCP Short circuit to positive points during the initialization phase because
HW Diagnosis by hardware these points are not run through during ini-
SW Diagnosis by software
tialization. Table 4
A further method of solenoid-valve monitor- If faults occur in the PRC, i.e. are detected, an
ing is ISIG evaluation (inductive signature). incorrect complete part answer of the PRC is
This method serves to monitor the voltage formed, which for its part causes an incorrect
characteristic at the solenoid valve and evalu- complete answer to be sent to the monitoring
ates the drop (UISIG) which occurs when the module (watchdog in the external ASIC, see
spool is moved (Figure 29). The objective also section entitled “ASIC”). This increases
here is to monitor the function of the on/off its fault counter by 1 (a fault-free complete
valve. Since the voltage drop is only very answer results in a decrease to the minimum
small and very short (tISIG), a special evalua- fault-counter content 0). The monitoring
tor circuit must be used here. module shuts down the driver stages when
the fault-counter content 5 is reached and
Pressure-Regulator Monitoring initiates a reset at fault-counter content 7.
The pressure regulator must be permanently
monitored because its function is crucial to
the function of the transmission.
Program-Run Check (PRC)

The program-run check (PRC) ensures that
the following incidents are detected:
쐌 mix-up of components,
쐌 double execution of code parts, and 29 ISIG signal characteristic
쐌 skipping of code parts.
Each module or each safety-relevant code

part must have a check point at the beginning
UISIG ≈ 10 mV
and the end in order to ensure a correct run

with maximum high probability. The num-
Voltage U
bers represent the relevant module (function

or process) and range from 0 to 9. The check-
sum algorithm (MISR procedure) ensures the tISIG ≈ 10 ms
sequence relationship.
æ STS0352E
Fig. 29
Each reference (every 10, 20, and 30 ms ref- UISIG Voltage drop
erence) has its own enumerator and delivers tISIG Time interval for
Time t
its own part answer. The last checked point in voltage drop
44 Basic principles of networking | Network topology
#BTJDQSJODJQMFTPGOFUXPSLJOH
With the tremendous speed at which Network topology

computer technology is advancing, the
number of electronic systems in use is A network is understood to be a system
increasing more and more. This growth in which a group of elements can exchange
is also continuing in automotive engi- information via a transportation medium.
neering. However, this also means that If the elements are visualized as nodes and
the complexity of an overall system (the the communication relationships as lines,
vehicle in this case) is on the increase. a picture is created of a network where
Individual systems such as engine man- many nodes are related to several other
agement have been improved over the nodes. The nodes in a communication net-
last few years. However, innovations are work are also often referred to as network
mainly achieved by means of interaction subscribers or stations.
between several individual systems. In motor vehicles, complex control units
The individual components need to be such as those for the engine management
networked so that the multitude of infor- system (Motronic or electronic diesel con-
mation that is managed by the individual trol, EDC), the electronic stability program
systems can also be used elsewhere (ESP), the transmission control system
throughout the system. Different com- or the door modules can be network sub-
munication systems are used depending scribers (Fig. 1). However, a sensor with a
on requirements (e.g. transmission reli- conditioning circuit that merely prepares
ability, fault tolerances, costs). and digitizes a measured value can also
act as a network subscriber and make the
measured signals available to other net-
1 Vehicle system networking
UVA0014-1Y

Basic principles of networking | Network topology 45
work subscribers. The transport medium Bus topology

via which the communication takes place This network topology is also referred
is referred to as a bus or a data bus. to as a linear bus. The core element of a
bus topology is a single cable to which all
A network topology is understood to be nodes are connected via short connecting
the structure consisting of network nodes cables (Fig. 2). This topology makes it ex-
and connections. This merely shows which tremely easy to add other subscribers to
nodes are interconnected, but does not de- the network. Information is transmitted by
pict underlying details such as the length the individual bus subscribers in the form
of the connection. Every network sub- of so-called messages and distributed over
scriber must have at least one connection the entire bus.
to another network subscriber in order Nodes transmit and receive messages.
to participate in network communication. If a node fails, the data that is expected
Different requirements are made of the from this node is no longer available to
network topology for a variety of commu- the other nodes on the network. However,
nication network applications, while the the remaining nodes can continue to ex-
topology determines some of the charac- change information. However, a network
teristics of the overall network. All net- with a bus topology fails completely if the
work topologies are based on the following main line is defective (due to a cable break,
four basic topologies for example).
Bus topology
Star topology Star topology
Ring topology and The star topology consists of a main node
Mesh topology (repeater, hub) to which all other nodes
are coupled via a single connection (Fig. 3).
Other structures (hybrid topologies) can A network with this topology is therefore
be created by combining these basic topol- easy to extend if free capacity is available
ogies. (connections, cables).
2 Linear bus topology 3 Star topology

UMA0036-3D
SVA0003Y
SVA0002Y
46 Basic principles of networking | Network topology
In star topologies, data is exchanged be- Ring topology

tween the individual node connections In the ring topology, each node is con-
and the main node, whereby a distinction nected to its two neighbors. This creates
is made between active and passive star a closed ring (Fig. 4). A distinction can
topologies. In active star topologies, the be made between single rings and double
main node contains a computer that pro- rings.
cesses and relays information. The perfor- In a single ring, data transfers are uni-
mance capability of a network is essen- directional from one station to the next.
tially determined by the performance The data is checked when it is received.
capability of this computer. However, If the data is not intended for this station
the main node does not have to have spe- it is repeated (repeater function), boosted
cial control intelligence. In passive star and relayed to the next station. The data
systems, it merely connects the bus lines that is being transferred is therefore re-
of the network subscribers together. layed from one station to the next in the
The following applies to active and pas- ring until it has reached its destination or
sive stars: if a network subscriber fails or arrives back at its point of origin, where
a connecting line to the main node is de- it will then be discarded.
fective, the rest of the network continues As a soon as a station in a single ring
to operate. However, if the main node fails fails, the data transfer is interrupted and
the entire network is disabled. the network breaks down completely.
In the automotive area, star structures
are under discussion for safety and secu- Rings can also be set up in the form of a
rity systems such as brakes and steering. double ring (e.g. FTTI), in which the trans-
In this case, the risk of a complete network fer of data is bidirectional. In this topology,
failure is prevented by designing the main the failure of a station or a connection
node to be physically redundant. This between two stations can be overcome,
means that several main nodes are used since all data is still transferred to all
to which the nodes whose information is operational stations in the ring.
needed for safe operation of the vehicle However, if several stations or connec-
can be connected in parallel. tions fail, the possibility of a malfunction
cannot be ruled out.
4 Ring topology 5 Mesh topology

SVA0004Y
SVA0005Y
Basic principles of networking | Network topology 47
Mesh topology Hybrid topologies

In a mesh topology, each node is con- Hybrid topologies are a combination of
nected to one or more other nodes (Fig. 5). different network topologies. Examples of
In a fully meshed network, each node is such combination are:
connected to every other node. Star bus topology: the hubs of several
If a mode or connection fails it is possi- star networks are interconnected as a
ble for the data to be rerouted. This type linear bus (Fig. 6).
of network therefore has a high degree of Star ring topology: the hubs of several
system stability. However, the cost of net- star networks are connected to the main
working and transporting the message is hub (Fig. 7). The hubs of the star net-
high. work are connected in the form of a ring
Radio networks form a type of mesh to- in this main hub.
pology, since the transmissions from each
station are received by every other station
that is within range.
A mesh topology is bus-like as far as
exchanging messages is concerned, and
star-like regarding data transfers, since
every station receives all transmissions
from every other station, but connection
failures can be overcome.
6 Star bus topology 7 Star ring topology

SVA0007Y
SVA0006Y
48 Basic principles of networking | Network organization
Network organization Message-oriented method

In this method it is not the receiver node
Addressing that is addressed, but the message itself
In order to make it possible to transmit (Fig. 8b). Depending on the content of
messages via a network and evaluate the the message, it is identified by a message
contents thereof, the useful data (payload) identifier that has been predefined for
that is transmitted is also accompanied by this message type. In this method, the
data transfer information. This can be ex- transmitter does not need to know any-
plicitly contained within the transmission thing about the destination of the message,
or implicitly defined using preset values. since each individual receiver node de-
Addressing represents important informa- cides whether or not to process the mes-
tion for data transfer information. It is sage. Of course, the message can be re-
needed in order for a message to be sent to ceived and evaluated by several nodes.
the correct recipient. There are different
ways of doing this. Transmission-oriented method
Transmission characteristics can also be
Subscriber-oriented method used to identify a message. If a message is
The data is exchanged on the basis of node always transmitted within a defined time
addresses. The message sent by the trans- window, it can be identified on the basis
mitter contains the data to be transmitted of this position. By way of a safeguard,
and also the destination node address this addressing is often combined with
(Fig. 8a). All receivers compare the trans- message or subscriber-oriented address-
mitter receiver address to their own ing.
address, and only the receiver with the
correct address evaluates the message. Bus access method
The majority of conventional communi- A node must access the bus in order to
cation systems (such as Ethernet) operate transmit a message. In the bus access
using the subscriber addressing principle. method a distinction is made between
Predictable methods in which the bus
access is determined by certain time-
dependent network characteristics,
whereby only one node can transmit
at a time and
8 Addressing modes Random methods whereby any node
can attempt to transmit data if the bus
a
appears to be free
Adr 1 Adr 2 Adr 3 Adr 4 In the predictable method the bus access
right is determined before bus access.
Adr 3 It can thereby be ensured that only one
subscriber is using the bus at a time.
Access collisions because of simultaneous
b
bus usage will be prevented if all subscrib-
Id 1 Id 2 Id 2 ers use this method.
Id 4 Id 3
Id 3 Id 3 Id 5
Id 6 Id 5 Id 6
Fig. 8
Id 7
UVA0012-1Y
a Subscriber-oriented
method
Id 3
b Message-oriented
method
Basic principles of networking | Network organization 49
In the random method, the nodes can Multimaster

simultaneously attempt to use the bus as In a Multimaster network, several nodes
soon as it appears to be free. The timing can access the transport medium indepen-
of the bus access is therefore random. dently without the assistance of another
There is a risk of transmission collisions node. Bus access is uncontrolled. Every
using this method, which will require at- node can access the bus and transmit a
tention. This can be dealt with by repeat- message if the bus appears to be free.
ing transmissions after a collision has been This means that each node is its own mas-
detected (e.g. Ethernet), by giving the ter, and that any node can start a message
transmissions different codings (CDMA), transfer with equal status. However, this
controlling communication via a master or also means that collision detection and
prioritizing message types or transmitters. handling methods have to be in place.
For example, this may be in the form of a
Time division multiple access (TDMA) decision-making phase with prioritization
TDMA is a deterministic (predictive) ac- or delayed transmission repeats. The use
cess method. In this case each node is as- of priority control prevents a bus conflict
signed a time window in which it is allowed if several nodes wish to use the bus at the
to transmit (a priori). A fixed schedule same time, since the network node that
is therefore required for the network. has high priority or wishes to transmit
There is not usually a main communication a message with high priority forces its
subscriber controlling the communication way through in the event of a conflict and
procedure. However, concepts exist in transmits its message first. Normal mes-
which in which it is possible to switch be- sage transmission resumes when the line
tween different schedules if necessary. is free again.
The internal clocks of the different stations The Multimaster architecture has a posi-
must run extremely synchronously with tive effect on the availability of the system,
TDMA, since the transmit windows have since no individual node is in control of
to be adhered to with extreme precision. communication whose failure would lead
to total communication breakdown.
Master-slave
In the master-slave system, one node
on the network operates as the master.
This node determines the communication
frequency by interrogating its subordinate 9 Master-slave method
nodes (slaves). A slave only replies if it is

spoken to by the master (Fig. 9). However,
some master-slave protocols allow a slave Master
to contact a master in order to transmit a
message (e.g. transmit information about
the position of the power-window unit to
the door module).
A B A B A B
UVA0013-1E
Slave Slave Slave

1 2 3
50 Basic principles of networking | OSI reference model
OSI reference model ranged functional areas (layers). Not all of

the layers in the OSI model are needed in a
Network protocols are usually defined simple communication system. Layers can
in layers, which combine properties and also be combined for many applications.
tasks. The properties of the deeper-lying Network protocols in the automotive area
layers are assumed in the next level up. are often divided up into
This has the advantage that individual Physical layer
layers are exchangeable, provided that Communication layer and
the interfaces that are provided between Application layer (user layer)
the layers remain unchanged.
The ISO OSI reference model (Open Physical layer
Systems Interconnection) provides a basis The electrical and procedural parameters
for describing and comparing many com- of the physical connection between the
munication protocols. This was developed network subscribers is defined in the
by the ISO (International Standardization physical layer.
Organization) and led to the adoption of
international standards by ISO and IEEE Signal level
(Institute of Electrical and Electronic In digital technology, data is represented
Engineers). by sequences of the two binary statuses,
In the OSI model, data communication 0 and 1. In order to transmit the data on
systems are depicted in different layers a bus, these statuses must be represented
(Fig. 10). The complex task of data commu- on the transmission agent. It is particularly
nication is distributed among clearly ar- important to avoid short-circuits on the
10 OSI reference model 11 Creating dominant and recessive levels
5V
Data Data
Bus
User layer line
Application layer E1 E2
T1 T2
Communication layer
Presentation layer
Control Status T Bus level
E
Session layer
0 conducts 0V
Transport layer
1 blocks 5V
Network layer
E1 E2 Bus level
Data link layer 0 0 0V dominant
0 1 0V dominant
Physical layer 1 0 0V dominant
1 1 5V recessive
Physical layer
UVA0027E
SVA0008Y
Physical connection
Basic principles of networking | OSI reference model 51
bus when one node is transmitting a status If no data is being exchanged, the bus level
of 1 and another is transmitting a status is 5 V (microcontroller operating voltage,
of 0. Fig. 11). When the start bit is transmitted
The binary statuses can be depicted in (dominant level), the other station con-
many different ways. The serial interface nected to the bus (receiver) is notified
of the PC, for example, uses +12 V and that a data transfer is starting (Fig. 12).
-12 V, and CAN-B uses voltages of 0 V and The length of the start bit determines a
5 V. The voltages of the serial interface are bit time that represents the basis for the
unsuitable for a bus, since short-circuits entire data transfer. Every subsequent
can occur if several subscribers wish to data bit has the same length. The recipro-
transmit conflicting binary statuses simul- cal of this time corresponds to the data
taneously. transfer rate, i.e. the number of bits that
If the coding allows one level to over- can be transmitted in one second in a
write another, the overwriting level is re- continuous data stream. All participat-
ferred to as dominant, and the subordinate ing stations must be set to the same data
level as recessive. transfer rate.
It is also possible to depict dominant After the start bit has been received,
and recessive levels using visual media. the transmission of an 8-bit data word
A status of 1 (recessive) then corresponds commences (1 byte) with the lowest signif-
to e.g. dark, and a status of 0 (dominant) icant bit (LSB, Low Significant Bit). The re-
corresponds to light. In an optical fiber, ceiver that has synchronized itself to the
an individual node can override all of the start bit scans the data bus between each
others by feeding light into the conductor. data bit and therefore assembles the trans-
ferred data byte.
Bit stream The eighth data bit is followed by the
The application information cannot usu- parity bit. This bit indicates whether the
ally be transmitted directly. In order to number of transmitted ones is odd or even.
make transmission possible, the informa- It therefore allows the receiver to perform
tion is first incorporated as a payload in a simple check for possible transmission
the frame of a message that contains infor- errors. The sequence is completed with
mation to be transmitted. Since all proto- the stop bit, which is placed onto the bus
cols have been developed in accordance
with different requirements, the frame
format differs from protocol to protocol. 12 UART interface transmission frame
The frame needs to converted into a

bit stream to actually transmit the infor-
mation, i.e. a sequence of bits that can be
transmitted via the transport medium as
physical states.
Example of the UART interface

Start bit
The microcontrollers that are used in
8 data bits (1 byte)
control units have a simple interface
(UART, Universal Asynchronous Receiver/ Parity
Transmitter) on the chip, via which they Stop bit
can communicate with the outside world
SVA0009Y
(e.g. with a PC). The essential features of

a data transmission are read out via this
interface.
52 Basic principles of networking | Control mechanisms
with a dominant level. The next data trans- Control mechanisms

fer can then take place.
Event control
Communication layer In an event-driven bus system, messages
Control units can only interconnect and are transmitted as soon as an event that
exchange data if they speak the same triggers the transmission of a message
“language”. This language determines has occurred (Fig. 13a). Examples of such
the rules that are used to exchange data events are:
between the individual network sub- Pressing a button on the air conditioning
scribers. system control panel
The communication layer accepts data Operating the hazard warning flasher
from the application layer, prepares it for switch
transmission and forwards it to the physi- Incoming message that requires a reac-
cal layer. tion (e.g. information from rpm sensor
The essential features of this protocol to speedometer needle motor)
layer are: Expiration of a fixed time period (time
Message frame format frame, e.g. 100 ms), after which mes-
Bus access control sages are transmitted cyclically
Message addressing
Detection and handling of collisions Since the stations are not synchronized
Network node synchronization with each other, situations where several
Checksum calculation stations wish to access the bus simultane-
ously are unavoidable. In order to allow
Application layer
The application layer consists of the appli-
cation that processes and provides the in- 13 Event control
formation. The application layer is the only

protocol layer to be affected by user or
a Occurrence of events
sensor input. high
1
Priority
2
low
Level
1
1 2 3
0
Time
b Occurrence of events
high
1
Priority
4
3 7
2 6
5
low
Level
1
1 3 4 2 5 6 7
SVA0010Y
Time
Basic principles of networking | Control mechanisms 53
a message to be transmitted without falsifi- Timer control

cation, only one station at a time can trans- In the most recent developments in dy-
mit data on the bus. Collision avoidance namic driving systems such as brakes and
mechanisms are available for preventing steering, an increasing number of mechan-
or solving bus conflicts. ical and hydraulic components are being
If a node wishes to transmit a message replaced with electronic systems (x-by-
whilst the bus is occupied, the transmis- wire). Mechanical connections such as
sion is delayed (Fig. 13b). A station that is the steering column are becoming super-
ready to transmit must then wait until the fluous, and the functionality thereof is be-
transmission that is currently in progress ing taken over by sensors and actuators.
has been completed. The reliability, safety and failure tolerance
Since bus access is subsequently rene- requirements of these systems are ex-
gotiated, the transmission may be delayed tremely high. This means:
yet again. These delays become problem- Messages must be received on time
atic if the bus becomes overloaded by a The latency time of critical messages
large number of network subscribers that must be extremely small
wish to transmit messages. In this case The system must have a redundant
messages may be lost if the transmitter design
abandons the transmission due to exces- The failure of a node must affect the rest
sive delays. of the system as little as possible and
Event-driven bus systems are suitable It must be possible to achieve a safe op-
for reacting to asynchronous (unforeseen) erating status from any fault situation
events as quickly as possible. In an ideal
case, they reduce the delay between the X-by-wire systems require close network-
occurrence of the event and the message ing by the various components. The exter-
transmission (latency time) compared to nal increase in complexity places new de-
time-driven systems. However, the latency mands on the safety, failure tolerance and
time can vary considerably depending on availability of the communication system.
the network loading. The demands that are made of the elec-
tronic and network architecture therefore
Advantages also increase. A reliable, fault-tolerant net-
High level of flexibility and capability of work architecture is required so that data
retrofitting new nodes in the network is transmitted with guaranteed transmis-
Good response time to asynchronous sion characteristics, and electronic system
external events malfunctions are handled in the most effi-
Bus usage depending on event fre- cient way.
quency in line with requirements
No network loading by unused events, System architectures for real-time applica-
since only events that have actually tions meet these requirements because
occurred trigger a transmission their behavior is predictable and verifiable
because of the way in which they are con-
Disadvantages structed. In these protocols, time windows
Static bus occupancy, non-deterministic within which a node is permitted to trans-
(i.e. not possible to prove that a message mit are assigned to the control units in the
was transmitted at the right time) communication network (nodes) during
network planning (Fig. 14). In order to com-
ply with the time window, the nodes must
be synchronized as precisely as possible.
54 Basic principles of networking | Control mechanisms
All transmissions are processed sequen- Advantages

tially in accordance with the network plan- Deterministic system
ning (without collisions). Once each node Punctual data transmission
has transmitted its message, the cycle re- Reliable detection and isolation of
starts with the first transmitter. This makes defective network nodes
it possible to determine how chronologi-
cally up-to-date the data is at any time. Disadvantages
Since missing messages are detected Overall system must be planned for
immediately, time-triggered concepts are distributed developments
more reliable than event-driven systems. Capacity for expanding the communica-
If a fast data rate is required in a time- tion system must be planned in
triggered system, the time delay between Good response time to asynchronous
the occurrence of an event and the trans- external events
mission of the data can be so small that
the system complies with strict real-time Composability
requirements. If a communication system allows inde-
The bus can be protected from unau- pendently developed subsystems to be
thorized access by a bus guardian. The bus integrated in an overall system, it is said
guardian prevents a defective node from to support composability. An important
interfering with network communication criterion when doing this is that the prop-
by transmitting messages outside the rele- erties that have been assured for the func-
vant transmit window. tionality of a subsystem are not adversely
These characteristics make it possible affected by adding other subsystems.
to create redundant, fault tolerant systems If this has been ensured, the checking
in which transmission errors can be reme- of system functionality is restricted to
died and faults in the network can picked subsystem checking that can be carried
up by network nodes that can provide the out by the constructor of the subsystem.
functionality without errors. If a communication system supports
composability, changes can be made to a
control unit without affecting the function-
ality of other control units. It is therefore
not necessary to recheck the entire system
after integrating a modified control unit –
it is sufficient to check that the individual
subsystems are operating reliably. Com-
14 Timer control posability therefore reduces the time and
cost of integrating new subsystems. This is
the only way to increase the complexity of
Occurrence of events
the electronics in the vehicle.
0
Max. output
Time slots delay
Time
SVA0011Y
Sensor measuring principles 55
왘 Overview of bus systems used in vehicles
CAN-C CAN-B LIN TTP

high-speed CAN low-speed CAN
Definition Controller area Controller area Local interconnect Time-triggered protocol
network network network
Bus type Conventional bus Conventional bus Conventional bus Conventional and
optical bus
Domains Drivetrain Comfort/ Comfort/ Safety-related
convenience convenience networking
Applications Engine management, Body and comfort Low-cost expan- Networking in safety-
transmission control and convenience sion of CAN bus for related environments
and ABS/ESP net- electronics net- simple applications such as brakes, steering,
working working in the comfort and railway signal boxes or
convenience electron- aircraft landing gear
ics area
Most frequently Linear bus Linear bus Linear bus Star topology
used topology
Data transfer rate 10 kbit/s to 1Mbit/s Max. 125 kbit/s Max. 20 kbit/s Unspecified,
typ. 10 Mbit/s
Max. number of nodes 10 24 16 Unspecified
Control mechanism Event-driven Event-driven Time-driven Time-driven
Bus lines Copper conductors Copper conductors Copper conductor Copper conductors
(twisted pair) (twisted pair) (single wire) (twisted pair)
Deployment in all vehicles in all vehicles in all vehicles Premium class vehicles,
aircraft, rail control
systems
Standard ISO 1198 ISO 11 519-2 LIN consortium TTA group
SAE classification Class C Class B Class A Drive-by-wire
MOST Bus Bluetooth Flexray

Definition Media oriented Proprietary name Proprietary name
systems transport (Danish king)
Bus type Optical bus Wireless Conventional and optical bus
Domains Multimedia and Infotainment Multimedia and Infotainment Deployment across all domains
Applications Transmission of control, Data transfers over short A network system for use
audio and video information distances, e.g. mobile phone in safety-related and simple
integration in the infotain- applications
ment system
Most frequently Ring topology Network topology (radio) Star topology
used topology
Data transfer rate Max. 22.5 Mbit/s Max. 3 Mbit/s (v2.0) Typ. 10 Mbit/s
Max. 723 kbit/s (v1.2) Max. 20 Mbit/s
Max. number of nodes 64 8 active (up to 256 passive) Theoretically up to 2,048
Max. 22 per passive bus/star
Control mechanism Time and event-driven Event-driven Time and event-driven
Bus lines Plastic or glass optical Electromagnetic radio waves Copper conductors
waveguides (twisted pair)
Deployment Premium class vehicles made All vehicles, connection be- Pilot application
by European manufacturers tween multimedia equipment
and infotainment system
Standard MOST cooperation Bluetooth SIG Flexray consortium
SAE classification Mobile Media Wireless Drive-by-wire
Table 1
56 Automotive networking | Cross-system functions
"VUPNPUJWFOFUXPSLJOH
Electrical and electronic systems in Cross-system functions

motor vehicles are often not indepen-
dent of each other but influence and If you examine the signals that are pro-
complement each other. For this reason, cessed in the individual systems, it be-
signal lines were used in previous injec- comes evident that many signals are
tion and ignition systems in order to needed in several control units. For exam-
simplify communication between these ple, the driving speed is evaluated in the
two systems. However, the increasing electronic stability program (ESP) for the
number of electronic systems rapidly vehicle dynamics control, in the engine
increased the demand for and the scope management for the automatic speed
of the information that was being ex- control (cruise control) and in the car
changed. The number of signal lines sound system for the speed-dependent
and plug connections that is required volume control.
increased accordingly, meaning that the The preparation of these variables from
technology that has so far been used was sensor signals requires computing power
approaching the limit of its capability. and therefore hardware and software re-
sources. It is therefore advisable for these
The solution was provided by the develop- variables always to be calculated in a con-
ment of serial bus systems, with which trol unit and transmitted to other control
large volumes of data from different units via a communication network.
sources can be transferred. A serial bus However, in order to implement cross-
system was first used in a vehicle in 1991, system functions, electronic systems can
when the CAN bus was used in the also exchange information and therefore
Mercedes-Benz 500E. influence each other. Intelligent sensors
The demand for additional driving are also considered to be electronic sys-
safety, convenience, economy and stricter tems that prepare the sensor signal in an
legal requirements on the environmental evaluation circuit and put the information
compatibility of motor vehicles can only on the data bus via a bus interface. Pre-
be achieved with the aid of additional elec- crash sensors detect a pending collision,
tronics. The number of electronic systems for example; the airbag control unit then
in vehicles is therefore increasing all the sends the door modules and the overhead
time (Fig. 2). control panel a request to close the windows
1 Automotive networking
SVA0037Y

Automotive networking | Cross-system functions 57
and the sliding roof. This protects the Requirements for

occupants from penetrating objects. bus systems
Another example of a system-encom-
passing function is the adaptive cruise Both financial (e.g. cable costs, component
control (ACC), in which the radar sensor, costs) and technical constraints must be
the engine management, the electronic taken into consideration when a bus sys-
stability program (ESP) and the transmis- tem is being selected. The most important
sion control communicate with each other. technical selection criteria are explained
Distance control from the vehicle in front in the following.
that is adapted to the flow of traffic is
made possible in this system by means Data transfer rate
of engine torque adjustments, automatic This variable specifies the volume of data
brake system intervention and gear that is transmitted during a time unit. The
selection. smallest unit of data is the bit, and the data
Coordination between the individual transfer rate is usually specified in bits/sec-
systems is therefore required for cross- onds. Alternative names for this expression
system functions. Large volumes of data are transfer rate, data rate or bit rate.
must be exchanged to do this. As well as The required data rate is dependent on
powerful components, a powerful commu- the application. A slower transfer rate is
nication system is also required, with a required to switch the air-conditioning
low-cost network that is suitable for auto- compressor on and off than to transfer
motive vehicles. Special serial databus sys- audio signals, for example.
tems have been developed for this purpose.
Interference immunity
The use of bus systems has the following Ideally, the data should be transferred
advantages in comparison to a solution without interference. However, this cannot
that uses conventional wiring: be guaranteed in a motor vehicle because
Reduced costs with less weight and in- of electromagnetic effects. The interfer-
stallation space because of fewer cables ence immunity requirements that are
in the wiring harness made depend on the safety relevance of
Better reliability and functional reliabil- the electronic systems concerned. Lesser
ity due to fewer plug-in connections requirements are made of comfort and
Simplification of vehicle assembly
during production Number of control units in the Mercedes S-class
2 that are networked via CAN
Multiple use of sensor signals
Simple connection of system compo- 60
nents to a bus
Easier handling of equipment and spe- 50
cial equipment variants in a vehicle
40
30
20
10
SVC0012-1Y
0
W140 W220 W221
(1991) (1998) (2005)
58 Automotive networking | Requirements for bus systems
convenience systems than the antilock Hard real-time requirement: the time
brake system (ABS), for example. specification must be strictly adhered
In order to meet these requirements, to. If the specified response time was ex-
mechanisms that detect transmission er- ceeded, the calculated result would not
rors are incorporated in the network probe able to be used. This can lead to seri-
tocols. A simple check can be carried out ous problems in safety-critical systems.
using the parity bit, which is calculated in
the transmitter and is transmitted together For example, if time allowances were ex-
with the useful data. This specifies whether ceeded in the ABS system, the incipient
the number of 1's in the transferred byte is locking of the wheels would not be de-
even or off. This information is checked by tected soon enough and the pressure in
the receiver. Single errors can be detected the master cylinder would not be reduced
using this method. in time. This would result in locked wheels.
Another method is the checksum check. The time allowances must also be strictly
If several data bytes are being transmitted, adhered to for many engine-management
the transmitter calculates a checksum system functions. Delays in transmitting
from the individual data bytes using a pre- injection and ignition signals could lead
defined formula and transmits this value. to engine judder and even misfiring.
The receiver also calculates the checksum These reactions must be avoided, since
of the data bytes that have been received they represent a potential danger. Hard
and compares it with the checksum that has real-time requirements must therefore be
been received. If a data transmission error made of these systems.
is detected, the received data is not used However, this does not necessarily mean
and a repeat transmission is requested. that the transmission of data via a bus sys-
tem also has to be subject to these hard
Real-time capability real-time requirements. Adherence to
A real-time system guarantees that its re- soft real-time requirements is usually suf-
sults are calculated within a fixed time in- ficient. If signals from other control units
terval. The duration of the time interval are needed for functions (e.g. a torque re-
depends on the application. The antilock duction request during a shift operation),
brake system (ABS) must react to the the bus system must transmit the data at
incipient locking of a wheel within a few a faster data transfer speed and with a
milliseconds (wheel speed reduction), smaller time delay so that the overall sys-
whereas response times of 100 ms are tem complies with the specified real-time
adequate for actuating the power-window requirements.
motor. Human beings cannot perceive
delay periods of less than 100 ms. Number of network nodes
Different demands are made of real-time The maximum number of nodes to be inte-
behavior depending on the application: grated varies for different areas of vehicle
Soft real-time requirement: the system operation. The number of nodes for com-
generally adheres to the specified re- fort and convenience systems may be high
sponse time, and if these times are occa- due to servomotor networking (e.g. seat
sionally exceeded, it does not produce adjustment) and intelligent sensors (e.g.
any serious effects (e.g. image jerking rain sensors). Several identical busses can
during picture transmission). be used if necessary.
Automotive networking | Classification of bus systems 59
Classification of bus systems Applications in the vehicle

Because of differing requirements, The overall vehicle system can be divided
bus systems can be subdivided into into four domains or functional areas from
the following classes. the point of view or electrics/electronics:
Drivetrain
Class A Chassis
Transfer rates Low data rates Interior and
(up to 10 kBit/s.)
Telematics
Applications Actuator and sensor
networking
Representative LIN In the drivetrain and chassis domains,
the emphasis is primarily on real-time
Class B
applications. In the interior domain, the
Transfer rates Average data rates
(up to 125 kBit/s.) main focus is on multiplex aspects in net-
Applications Complex mechanisms working. Mainly multimedia and infotain-
for error handling,
control unit networking
ment applications are networked in the
in the comfort functions telematics domain.
Representative Low speed CAN
Class C Real-time applications

Transfer rates High data rates The networking of these systems makes
(up to 1 MBit/s.) considerable demands of the performance
Applications Real-time requirements, capability of the communication system.
control unit networking
in the drive and running Crankshaft-synchronous processes or pro-
gear functions cesses within a fixed time frame with cycle
Representative High speed CAN times of a few milliseconds are typical.
Class C+ If the system response times are adequate
Transfer rates Extremely high data rates for the task in hand, it is described as hav-
(up to 10 MBit/s.)
ing real-time capability (e.g. rapid ignition-
Applications Real-time requirements,
control unit networking timing advance in the Motronic after a re-
in the drive and running quest from the traction-control system for
gear functions
reducing torque and therefore preventing
Representative FlexRay
the wheel from spinning).
Class D
Transfer rates Extremely high data rates
(> 10 MBit/s.)
3 Domains in the overall vehicle system
Applications Control unit networking

in the telematics and
multimedia functions
Representative MOST
Chassis Drivetrain
Interior Telematics
SVC0013-1Y
60 Automotive networking | Applications in the vehicle
The drivetrain and chassis systems are Multimedia networking

assigned to class C. These require fast Mobile communication applications com-
transfer rates in order to ensure the real- bine components such as
time behavior that is required for these Car sound system
applications. They also make considerable CD changer
fault tolerance demands. These require- Navigation system
ments are met by the event-driven CAN Driver-information systems
bus with a transfer rate of 500 kBaud Telephone
(high-speed CAN). Video system
Voice input
Examples: Internet, E-mail
Engine-management system (Motronic Back-up camera
or electronic diesel control, EDC)
Transmission control The networking of these components
Antilock brake systems, ABS makes it possible to have a centrally lo-
Vehicle dynamics control (e.g. electronic cated display and control unit for several
stability program, ESP) applications. Operating procedures can
Chassis control systems (e.g. active body be standardized in this way, and status
control, ABC) information can be summarized. Driver
Support systems (e.g. adaptive cruise distraction is therefore minimized.
control, ACC)
A distinction must be made between con-
Multiplex applications trol data and audio/video data in multime-
The multiplex application is suitable for dia networking. Transfer rates of up to
controlling and regulating components in 125 kBit/s are sufficient for control tasks
the body and comfort and convenience (such as CD changer control), meaning that
electronics area (class B), such as the low-speed CAN bus can be used, for
Displays example. The direct transmission of audio
Lighting or video data requires extremely high
Access authorization with anti-theft transfer rates of more than 10 MBit/s.
warning device The MOST bus is used for this purpose,
Air-conditioning for example.
Seat and mirror adjustment
Door module (power-window unit,
door-mirror adjustment)
Windshield wipers
Headlamp adjustment
The transfer rate requirements are not

as high for class B systems as they are for
class C systems. For this reason, low-speed
CAN with a transfer rate of 125 kBit/s or
single-wire CAN with 33 kBit/s. can be used.
If the transfer rate requirements drop
to less than 20 kBit/s, the low-cost LIN is
more frequently used. Applications are
mainly in the mechatronics area; examples
being the transfer of switch information or
the activation of actuators.
Automotive networking | Coupling of networks 61
Coupling of networks Examples of networked

vehicles
The network topologies and protocols that
are most suitable for requirements are Topology
used for the different applications. How- The topologies of the communication net-
ever, the different network protocols are works can differ considerably depending
incompatible, meaning that data cannot be on the vehicle equipment. Figure 5 shows
simply exchanged between networks. examples of how the network can be struc-
In this case, help is provided by a gate- tured for different vehicle classes. In some
way. A gateway can be compared to an in- cases, different car manufacturers use dif-
terpreter that receives the “data” from a ferent bus systems for communication.
discussion partner, translates it and passes
it to another discussion partner. Techni- Signal transmission
cally speaking a gateway is a computer Signal types
that reads in the data that is transmitted A wide variety of information can be trans-
by the networks and converts it into an- mitted in a communication network in a
other format. The use of gateways there- vehicle. Some examples are:
fore makes it possible to exchange infor- Engine operating conditions (e.g. engine
mation between different networks. temperature, engine speed, engine load)
A central gateway (Fig. 4a) or several Physical measurements recorded by
distributed gateways can be used (Fig. 4b) sensors (e.g. outside temperature)
to interconnect the bus systems. All bus Control signals for activating servo-
lines are routed to the central gateway. motors (e.g. power-window units)
In the other case, one gateway connects Control-element switch positions
two or more busses. (e.g. for the windshield wiper)
Multimedia data (audio and video) for
transmitting music and speech (e.g. from
radio stations or when handsfree talk-
ing with a cellular phone) and moving
pictures (e.g. when playing a DVD or the
displays from a reversing camera)
4 Gateway structures
a b
Diagnostics
LIN Diagnostics Gateway 2

(drivetrain)
FlexRay Central
CAN-C gateway Gateway 1 Gateway 3
CAN-B ECUn (telematics) (interior)
ECU1 ECU1 MOST Gateway 4
ECU1 ECU 1
ECU1 ECU2 ECU2 (chassis)
ECU1
ECU2 ECU 2
ECU2 ECU3 ECU3 ECU2
ECU3 ECU 3
ECUn ECUn ECUn Fig. 4
ECU4 ECU 4 ECU 1 ECU 1 ECU n a Network with
Chassis Drive- Interior Telematics
SVC0014-1E
train ECUn ECU n ECU n ECU 2 ECU 3 central gateway

b Network with
several distributed
gateways
62 Automotive networking | Examples of networked vehicles
5 Network topologies
a
Sensor CAN
Diagnostics
CAN Drive CAN
Instrument
cluster CAN
LIN
Comfort CAN
LIN
Infotainment CAN
Comfort CAN
LIN
BluetoothTM
b Distance control
CAN
Instrument
cluster CAN
Diagnostics CAN
MOST
Drive CAN
LIN LIN
Fig. 5 LIN
Typical network
topologies for
new-generation
SVA0028E
vehicles Comfort CAN

a Compact class
b Luxury class
Automotive networking | Examples of networked vehicles 63
Resolution (e.g. 1 second). The engine speed, on

The signals must be available with a suit- the other hand, can change extremely
able resolution. Switch positions can rapidly. The times at which the engine
simply be shown as a 1-bit value (0 for management carries out measurements
switch open, 1 for switch closed). Other and calculations depends on the crank-
signals such as digitized analog voltages shaft position and is performed once per
from the engine-temperature sensor or combustion cycle. At fast engine speeds,
the calculated engine speed are shown as this corresponds to a time interval of a
1-byte or 2-byte values, for example, de- few milliseconds, i.e. approx. 3.3 ms at an
pending on resolution requirements. One engine speed of 6,000 rpm for a 6-cylin-
byte can be used to represent 256 values, der engine. However, not every system
and two bytes can be used to represent that needs the speed information for its
65, 536 values (= 256?256). The resolution control and regulation functions depends
for sensor signals with a voltage range on this availability. The engine control
of 0…5 V is approximately 20 mV with unit therefore does not need to output the
a 1 byte representation (= 5 V/256). A speed information on the data bus as soon
resolution of 5 mV requires a 10-bit data as it is calculated. In this case too, data
representation. transmission takes place according to a
The conversion between the binary cyclic time frame. A time frame of 10 ms
value and the physical value must be uni- is normal in the engine control area. This
form so that the transferred signals rep- means that the speed information is trans-
resent the same physical value in all sys- mitted on the bus 100 times per second.
tems. A resolution of 30 rpm is sufficient
for accessing the ignition map as far as Multimedia data
the engine speed n is concerned. A value A multimedia-compatible digital bus
range of 0 to 255?30 rpm (= 7, 650 rpm) system is frequently used in luxury-class
and therefore the entire speed range can vehicles to transfer audio data – as an
consequently be represented by one byte alternative to using analog cables. In ad-
(8 bits). An incrementation of 30 rpm is too dition to improved audio quality, a bus
little for idle-speed control, on the other system of this type offers the advantage
hand. More bits are required to represent that various audio streams and the associ-
the signal with a higher resolution, pro- ated check commands can be transferred
vided that the same measuring range is in parallel. In the automotive industry,
being recorded. the optical MOST bus (Media Oriented
Systems Transport) has proven successful
Output as a the multimedia bus system.
In event-driven systems such as the ones The transfer of audio signals requires
that are primarily used in the automotive on the one hand synchronized transfer
area, the signals can be transferred on between the transmitter and one or more
the data bus when an event occurs. Ex- receivers and on the other hand a high
amples of such events are the operation data rate. In view of the fact that data
of switches for switching on the air-condi- transfer on the MOST bus takes place syn-
tioning system or the windshield wiper. chronously with a fixed clock-pulse rate,
Signals that represent the operating state synchronization is already assured by its
of the engine, for example, are not nec- transfer mechanism.
essarily associated with an event. The The digital transfer of CD-quality audio
engine temperature, which only changes signals, i.e. with a resolution of 16 bits and a
slowly, is cyclically measured in a fixed clock-pulse rate of 44.1 kHz, requires a con-
time frame by the engine control unit stant data rate of 1.35 Mbit/s for one stereo
channel. Up to 15 stereo channels can be driving speed to be read out in the work-
transferred in parallel with the current ver- shop via the connected diagnostic tester
sion of the MOST bus (MOST25 with a total (example: the correct assignment of the
transfer rate of up to 24.8 Mbit/s). wheel-speed sensors must be checked for
the ABS functional test).
Data transfer: examples
The following examples show which sig- Engine speed
nals are measured and evaluated in which The injection and (with a gasoline engine)
systems. the ignition timing are output with a reso-
lution of less than 1° of the crankshaft an-
Driving speed gle. In order to ensure real-time behavior,
The ESP control unit calculates the driving the crankshaft position must be recorded
speed from the wheel-speed sensors. This in the engine control unit. The engine-
variable is transmitted on the CAN-C bus speed sensor scans the crankshaft trigger
(drive CAN). The engine-management sys- wheel and relays the signal to the control
tem needs this value for the cruise control, unit, which calculates both the crankshaft
among other things, and the transmission position and the engine speed. This vari-
control unit determines gear changes from able is used to calculate the injection time
the driving speed. The adaptive speed and the ignition angle, for example.
control (ACC, Adaptive Cruise Control) The engine speed is a variable that is
needs the current driving speed to calcu- needed in many other systems. The engine
late the necessary distance from the ve- control unit therefore outputs it on the
hicle in front and use it as a setpoint value. data bus. The shifting points are defined
A gateway transmits the speed informa- in the transmission control unit depending
tion via another CAN bus (instrument clus- on the speed. The engine speed is needed
ter CAN) to the instrument cluster, which for the ASR function (acceleration slip
displays the value via a needle instrument. control) in the Electronic Stability Program
The CAN-B bus (comfort CAN) is also (ESP) – ASR intervention (torque reduc-
connected to the network via the gateway. tion) must not make the engine stall.
Some luxury class vehicles are equipped As in the previous example, the engine
with dynamic seats. The padding of the speed is transmitted to the diagnosis inter-
seats is inflated depending on the speed face and the instrument cluster (display on
and the acceleration, counteracting the rev counter).
the centrifugal force of the driver. This
increases comfort considerably when Turn signaling
cornering. The driver operates the turn-signal lever
The speed information is sent to the (Fig. 6, Item 1). A signal is relayed to the
Infotainment CAN via the gateway and re- steering column control unit via a discrete
layed to the car sound system. This allows line (2) depending on whether the driver
the volume to be adapted to the driving is indicating a right or left turn. This may
speed. The navigation system needs the be a resistance-coded signal, for example.
speed to calculate the position if the GPS The control unit evaluates the signal and
signal is missing (e.g. in a tunnel). detects that the driver is indicating a left
The diagnosis interface is directly con- turn, for example.
nected to the engine and transmission The comfort CAN relays this informa-
control unit via the serial K-line. All other tion to the vehicle power supply control
control units are connected to the diag- unit (3). The indication direction is defined
nosis interface via a virtual K-line that is on the basis of the received information
simulated on the CAN bus. This allows the (normal flash frequency, increased flash
frequency in the event of bulb failure). The Load management

front left and rear left turn-signal lights At low revs (idle speed) and when a con-
are then actuated via discrete lines (4, 5). siderable amount of power is being used
The vehicle power supply control unit also up by the electrical consumers that are
transmits the “left turn signaling” informa- switched on, the battery or alternator volt-
tion on the comfort CAN. The gateway (6) age can drop to a low value. The vehicle
relays the information to the instrument power supply control unit calculates the
cluster CAN. The indicator lamp then current status of the vehicle power supply
flashes on the instrument cluster (9). from the current battery voltage, the DF
If the vehicle has a trailer hitch, the in- signal from the alternator (alternator uti-
formation goes to the trailer-recognition lization) and the information about heavy
control unit (7) via the comfort CAN. This current consumers that are switched on
actuates the turn-signal lamp on the trailer with a short switch-on duration. The vehi-
(8) via cables. cle power supply control unit requests an
idle speed increase via the CAN bus if the
Wiper stage 1 vehicle power supply is insufficient at idle
The wiper switch (Fig. 7, Item 1) transmits speed. The engine control unit implements
a signal via a discrete line to the steering the request. If this action does not solve
column control unit (2), which evaluates the problem, the vehicle power supply
the information (e.g. wiper stage 1). The control unit switches off specific consum-
control unit transmits this information on ers such as the heated rear windshield, the
the comfort CAN bus. The vehicle power seat heating or the heated outside mirror.
supply control unit (3) picks up the infor- These consumers are connected to the
mation and relays it to the wiper motor (4) vehicle power supply control unit via dis-
via the LIN bus. The vehicle power supply crete lines.
control unit acts as a gateway between the
comfort CAN and the wiper LIN.
Fig. 6
1 Turn-signal lever
2 Steering column
control unit
3 Electrical-system
6 Data transfer during turn signaling 7 Data transfer during windshield-wiper operation
control unit
4, 5 Turn-signal lights
4
6 Gateway
4 7 Trailer-recognition
3 5 LIN control unit
M
3 8 Turn-signal light
Comfort CAN
on trailer
9 Instrument cluster
Comfort CAN
2
1
Instrument cluster
Fig. 7
CAN 9 1
2 1 Windshield-wiper
6
lever
2 Steering column
control unit
SVA0029E
SVA0030E
7 3 Electrical-system
control unit
8
4 Wiper motor
Multimedia application to its current reception status (the current

The signal sequences during the output of station name, for example) which are up-
a radio signal on the amplifier is described dated where necessary.
in the following as an example of the trans- At the amplifier, the head unit connects
fer of signals and multimedia data in a the stereo channel with a selected input
luxury-class vehicle infotainment system and uses corresponding check commands
with MOST bus (Fig. 5b). to set the properties for outputting the au-
The functions of the vehicle infotain- dio signal, e.g. the set volume. When all the
ment system are controlled via the cen- settings have been made, it instructs the
tral operating unit, the head unit (Fig. 8, amplifier to fade in the audio signal of the
Item 1). In current systems, this unit usu- radio signal, which is then output through
ally has a screen in the instrument panel the infotainment system’s speakers.
and a rotary/pushbutton controller in the Before this, the head unit has if neces-
center console, by means of which the user sary ensured that the output of a previ-
interacts with the system. ously output audio signal has been faded
To output a radio station, the head unit, out, the output of the associated device
which in most systems is responsible for stopped, and the associated audio channel
managing the audio channels of the info- removed.
tainment system, establishes on the MOST In parallel the CAN/MOST gateway (4)
bus a stereo channel of the required audio constantly transmits the driving-speed
quality between the radio tuner (2) and the data, which it receives via the comfort
amplifier (3). It now tunes in the requested CAN, via the MOST bus to the amplifier,
station at the radio tuner via correspond- which has requested a notification for this
ing check commands and if necessary information when the system was started.
makes further settings. Finally, the head The amplifier can use the speed together
unit connects the tuner output with the with further vehicle information to calcu-
previously created stereo channel. For the late additional settings for outputting the
information to be displayed in the user audio signal, e.g. adapting the volume de-
interface, the head unit receives from the pending on the current speed.
radio tuner corresponding data relating
8 Multimedia networking
4 2
Comfort CAN
MOST
Fig. 8 3
5 1
1 Head unit
2 Radio tuner
3 Amplifier 6 7
SVM0005E
4 CAN/MOST gateway
5 Screen Check commands
6 Control element Multimedia data stream
7 Speakers
Control of an automatic steel folding roof four CAN buses and a further four LIN sub-
Vehicles with automatic steel folding roofs buses. In this example, a total of 13 control
are currently very much in vogue. These units are involved in the control of the au-
roofs are automatically opened and closed. tomatic steel folding roof. The text below
The technical realization of this seemingly describes which functions are performed
simple mechanical function poses a huge by the individual control units to move the
challenge for the networking of the elec- roof.
tronics and control units.
The networking depicted in Figure 9 Roof control unit
shows a typical present-day mid-size ve- Power activation of the motors for the steel
hicle with the additional function of an au- folding roof is performed by this control
tomatic steel folding roof. At its maximum unit. This control unit also assumes the
equipment specification, this vehicle has role of complete monitoring of the move-
over 35 electronic control units (ECUs), ment process. Proximity-type sensors
which communicated with each other via which monitor, record and evaluate the
9 Control of an automatic steel folding roof
Instrument cluster
Central OBD socket
(WFS)
gateway
Central
processing unit Electrical-system Door control unit Dynamic headl.
(Radio, RadioNavi, control unit driver leveling contr.
RadioNaviNet,…)
Wiper
(master) Door control unit AWD
passenger control unit
Booster AMP Wiper
(slave) Door control unit Transmission
rear left control unit
Voice control RS/LS
Door control unit Airbag
SMLS control unit
rear right
TV tuner (steering column)
Comfort control unit
MF st.w. Selector lever
Telephone (Kessy)
interface box
Roof control unit IR Electric
(CSC) power steering
Telematics
NAR Sounder
PTC heating Steering-angle
NGS sensor
Parking aid Engine control unit
A/C control unit NOx
…
Tire-pressure Seat, driver
monitoring (memory) Brake
control unit
Trailer Seat, passenger Fig. 9
(ABS, ESP, …)
control unit (memory) Systems in blue boxes
SVA0047E
ESP
cluster are involved in the
Multifunction
Auxiliary heating control of the steel
control unit
folding roof
movement of the roof are used for this ated. This also applies to the sliding sun-
purpose. In order to inform the driver of roof, which is activated by the roof control
the current status while the roof is being unit.
opened and closed and to prevent any
damage to the vehicle or the surroundings, Trailer control unit
the roof control unit receives numerous The trailer control unit informs the roof
further parameters from other control control unit whether a trailer is hitched. If
units in the vehicle. this is the case, opening and closing of the
roof is disabled.
Instrument cluster
The instrument cluster receives via the Door control units
gateway the current status from the roof The status of the windows and doors is
control unit and from other participating interrogated by all four door control units
control units, and informs the driver ac- and transmitted to the roof control unit.
cordingly. Example: “Obstacle behind the Opening or closing of the roof is only
vehicle! Roof movement stopped”. begun if all the doors are closed and the
windows are in the correct positions. This
Gateway setting is automatically corrected if the
All the communication requests which are windows are not in the correct positions.
transmitted from one bus system to an-
other are routed via the central gateway. Comfort control unit
The comfort control unit informs the roof
Vehicle power supply control unit control unit of, among other things, the
The vehicle power supply control unit key position. Movement of the roof is en-
checks whether the vehicle battery has abled only if the correct key is inserted.
sufficient charge to enable the roof to be
moved. If necessary, the driver is informed Brake control unit
and no roof movement is performed. The brake control unit uses sensors to
record the wheel speeds and thereby iden-
Parking-aid assistant tify whether the vehicle is moving. The
Before the roof is opened, the parking-aid roof control unit receives the speed infor-
assistant monitors the area behind the mation and enables movement of the roof
vehicle for obstacles to ascertain whether only if the vehicle is stationary.
there is sufficient space available to un-
fold the roof. If an obstacle is detected, Future areas of FlexRay application
the driver is informed and movement Up to now the high-speed CAN bus (CAN
is stopped. In this event, the driver can C) has been used to network control units
decide for him-/herself whether he/she in the drivetrain and in the chassis area. In
wishes to continue moving the roof. In future X-by-wire systems, the mechanical
certain cases the parking-aid assistant can connections, e.g. between steering wheel
detect something which poses no danger and front axle (steer-by-wire) or brake
to the vehicle. pedal and wheel brakes (brake-by-wire),
will be replaced by electrical communica-
A/C control unit tion systems in conjunction with driving-
All windows are automatically closed dynamics control systems.
when the air-recirculation switch is actu-
Furthermore, vehicle architectures have ity Program), AFS (Active Front Steering),
up to new often used several CAN buses and ARC (Active Roll Control). Higher-
which are linked to each other via gate- level driving-dynamics control controls
ways to distribute the high data volume. the individual systems without limiting
Future architectures will use, for a fast, their functionality and prevents negative
powerful connection between several mas- interactions. Figure 10 shows the intercon-
ter computers (which assume, for example, nection of these systems with bus system
central functions in the safety and driver- and sensors.
assistance areas), a backbone bus with a The yaw sensor transmits the vehicle’s
high data rate to which in each case sub- acceleration values and yaw rate to ESP.
buses (e.g. drivetrain CAN and sensor CAN If a critical driving situation is detected
in the chassis area) are connected. by ESP using these data (e.g. vehicle over-
The new FlexRay bus system satisfies steer), situation-conditioned brake pres-
the demands which will be placed in future sures and engine-management interven-
on the vehicle architecture, such as, for tions are calculated. Higher-level global
example driving-dynamics control also evaluates
high data rates and guaranteed real- the driver-command steering angle from
time capabilities in the drive and chassis the steering-angle sensor and calculates
areas, a supplementary steering angle, which is
large date volume in the backbone, and converted by AFS. More effective and com-
high failure safety of safety-relevant fort-enhancing control can be achieved by
applications (e.g. X-by-wire) this interaction of the individual systems.
In this way, the vehicle can, for example,
through properties such as be stabilized by a corrective steering
high availability and redundancy by movement already at a very early stage
means of two physically independent such that braking interventions can be
channels, partially or even completely avoided.
high data rates with up to 10 Mbit/s per
channel,
data transfer with guaranteed latency,
and 10 FlexRay
synchronicity of all communication
users by means of a global time base. Global driving-dynamics control
Driving ESP
CAN dynamics FlexRay (active
FlexRay is currently still in the develop-
braking)
ment stage, although some initial systems
are already in volume production. The fol-
lowing example is a current approach to a Sensor AFS
technology (Active
FlexRay topology.
Front
Steering)
Excerpt from chassis domain
Concepts for global driving-dynamics Further ARC
compo- (Active
control are being developed to facilitate nents Roll
SVA0048E
control which is as cooperative and thus Control)

as effective as possible. Components of
these concepts are ESP (Electronic Stabil-
70 Bus systems | CAN bus
#VTTZTUFNT
CAN bus to damage. As a result of these different

requirements, buses with different data
In 1991 the CAN bus (Controller Area rates are used that offer an optimum cost-
Network) was the first bus system to be benefit ratio for the field of application
introduced to a motor vehicle in mass concerned. A distinction is made between
production. It has since established itself high-speed and low-speed CAN buses.
as the standard system in the automotive
sector, but the CAN bus is also commonly High-speed CAN (CAN-C)
used as a field bus in automation engineer- CAN-C is defined in ISO Standard 11898-2
ing in general. In imitation of other net- and operates at bit rates of 125 kBit/s to
work types, such as the local area network 1 MBit/s. The data transfer is therefore
(LAN), wide area network (WAN) or per- able to meet the real-time requirements
sonal area network (PAN), this bus system of the drivetrain.
was given the name, CAN. CAN-C buses are used for networking
the following systems:
Applications • Engine-management system (Motronic
The CAN bus is used in various domains in • for gasoline engines or EDC for diesel
the motor vehicle. These domains differ in engines)
the requirements they demand of the net- • Electronic transmission control
work. Due to the fast processes involved in • Vehicle stabilization systems (e.g. ESP)
the area of engine management, informa- Instrument cluster

tion is required much faster here than in
the area of comfort/convenience where Low-speed CAN (CAN-B)
the controlled systems are located further CAN-B is defined in ISO Standard 11898-3
apart and as such lines are more prone and operates at a bit rate of 5 to 125 kBit/s.
For many applications in the comfort/con-
venience and body area, this speed is suffi-
1 Networking of electronic control units by CAN cient to meet the real-time requirements
demanded in this area. Examples of such
applications are:
ECU 1 ECU 2 ECU 3 ECU 4 • Control of the air-conditioning system
• Seat adjustment
CAN-C • Power-window unit
• Sliding-sunroof control
• Mirror adjuster
• Lighting system
Gate-
way • Control of the navigation system
The CAN bus is finding ever more use in

Diagnosis vehicle diagnostics. Here, the electronic
interface control unit is connected directly to the
CAN bus and thus receives the information
Fig. 1 ECU 5 ECU 6 ECU 7 ECU 8 it needs for diagnostics immediately. Pre-
ECU Control unit
vious diagnosis interfaces (e.g. KWP2000)
ECU 1 to ECU 4
CAN-B are becoming less important.
On the CAN-C
SVC0016E
high-speed bus
ECU 5 to ECU 11 ECU 9 ECU 10 ECU 11
On the CAN-B
low-speed bus

Bus systems | CAN bus 71
Topology Incoming messages are processed by the

Bus topology transceiver and sent to the CAN controller
In the development of the CAN, particular on the RxD line.
focus was placed on eliminating the need The microcontroller, which runs the ap-
for a central control element for communi- plication program (e.g. Motronic), controls
cation. This approach is most effectively the CAN controller, prepares the data to be
supported by a bus topology in which all sent and evaluates the data received.
network nodes are connected to a bus and
each node is able to receive all information Logic bus states and coding
sent on the bus. The bus topology is most CAN uses two states for communication,
commonly selected during the conceptual dominant and recessive, with which the
design of the communications system. information bits are transmitted. The dom-
In addition to offering favorable electri- inant state represents a binary “0”, the
cal properties, the linear bus topology recessive a binary “1”. NRZ (Non-Return to
has the advantage that the failure of one Zero) is used as the encoding method for
station would not affect the functionality the data transmission. With this method,
of the data transmission system. Further- there is no compulsory return to zero be-
more, additional stations can be connected tween two transmission states of the same
to the system with little extra effort. value.
When it receives messages, the
Star topology CAN transceiver converts the signal level
The use of a central coupler makes it pos- back to logical states. In the process, a
sible to build star topologies. Active as well differential amplifier subtracts the CAN_L
as passive couplers can be used. The use level from the CAN_H level (Fig. 3). If lines
of a star topology achieves a high level of become twisted, disturbance pulses (e.g.
flexibility in adapting to the networking from the ignition system) have the same
task. effect on both lines. Differential data trans-
The coupler is used to build a star topol- fer therefore makes it possible to filter out
ogy and simply forwards the messages to interference on the line.
the individual segments. Since the signal
transit times remain unchanged, the cir-
cuit length of the star topology is the same
as that of a bus topology.
2 Network nodes in the CAN
Data transmission system
Network nodes
Network nodes
A network node (Fig. 2) comprises the mi-
crocontroller for the application software, Actuators Microcontroller Sensors
the CAN controller and the CAN trans-
ceiver (bus driver). The CAN controller is
responsible for the transmit and receive CAN controller
modes. It generates the bit stream for CAN TxD
CAN RxD
data communication from the binary data
CAN transceiver
to be transmitted and forwards it to the
transceiver on the TxD line. This amplifies CAN_H CAN_L
the signals, generates the voltage level re-
SVC0017E
quired for differential data transfer and CAN bus

transmits the processed bit stream serially
on the bus line (CAN_H and CAN_L).
Some transceivers also evaluate the volt- Single-wire line

age level on the CAN_H and CAN_L line The single-wire line is a means of reducing
separately. It would then be possible for manufacturing costs by dispensing with
operation to continue in single-line mode the second line. For this to be possible,
if one of the two bus lines were to fail as however, all bus subscribers must share
a consequence of a short-circuit or cable a common ground that would assume the
break. However, the bus subscribers would function of the second line. The single-
have to share a common ground that would wire version of the CAN bus is therefore
assume the function of the failed line. only an option for a communications sys-
tem of limited spatial dimensions.
Transmission agent and bus coupling The data transfer on the single-wire line
Two-wire line is more prone to interference radiation be-
For the CAN, any transmission agent on cause it is not possible to filter out distur-
which dominant and recessive states can bance pulses as it is with the two-wire line.
be transmitted is worthy of consideration. For this reason, a higher level increase is
An untwisted or twisted pair of wires, de- required on the bus line to improve the
pending on the ambient conditions, is usu- signal-to-noise ratio. This in turn has a
ally used, and the wires are either galvani- negative effect on interference radiation.
cally coupled or decoupled. The two bus The flank steepness of the bus signals
lines are designated CAN_H and CAN_L. must therefore be reduced in comparison
The two-wire line supports symmetrical to the two-wire line. This is accompanied
data transfer whereby the bits are sent on by a lower data transfer rate. Conse-
both bus lines and represented by differ- quently, the single-wire line is only used
ent voltages. This reduces sensitivity to for the low-speed CAN in the body elec-
in-phase interference because the interfer- tronics and comfort/convenience electron-
ence affects both lines equally and can be ics. Thanks to this feature, a low-speed
filtered out. Additional shielding of the CAN with two-wire line would still remain
lines reduces their own radiation emis- functional in the event of a line failure
sions, especially at high baud rates. (CAN_H or CAN_L).
3 Filtering out interference on the CAN bus
a V
3.5 CAN high level

(CAN_H)
2.5
CAN low level
1.5
(CAN_L)
0
a V
2.0 Differential
SVC0018E
signal
0
Voltage level Emission limits

The CAN transceiver converts the logical To permit correct evaluation of bit trans-
states 0 and 1 received by the CAN control- mission, the signal must arrive at each
ler into voltage levels that are fed to the node within the bit time without interfer-
CAN_H and CAN_L bus lines. ence, and yet still at the relevant sampling
The high-speed and low-speed CANs use point. Delays arise from the signal transit
different voltage levels for the transmis- time on the data bus. The maximum per-
sion of dominant and recessive states. missible bit rate thus depends on the total
The voltage levels of the low-speed CAN length of the bus. ISO 11898 specifies
are shown in Figure 4a, those of the high- the bit rate for a defined circuit length.
speed CAN in Figure 4b. The following recommendations exist for
In the recessive state, the high-speed longer lines.
CAN uses a voltage of 2.5 V on both lines. • 1 MBits/s for 40 m (specified)
In the dominant state, a voltage of 3.5 V is • 500 kBits/s up to 100 m
present on CAN_H and a voltage of 1.5 V is (recommendation)

present on CAN_L. • 250 kBits/s up to 250 m
On the low-speed CAN, a voltage of 0 V • 125 kBits/s up to 500 m
is present on CAN_H in the recessive state, • 40 kBits/s up to 1,000 m
and 5 V on CAN_L. In the dominant state,

voltages of 3.6 V and 1.4 V are present on It is possible to connect at least 30 network
the CAN_H and CAN_L respectively. nodes to the bus without the need for ad-
ditional measures.
Reflection-free termination
Reflections of the electrical signals at open
ends of lines would interfere with commu-
nication. To dampen these reflections, the
bus lines are terminated at each end with
a resistor of 120 V.
Alternatively, the terminating resistors
may be integrated into the electronic con-
trol units themselves.
4 Voltage level 5 Data transmission system
a V
CAN_L
5
3.6
Node 1 Node n
1.4 (max. 30)
CAN_H
0
recessive dominant recessive
Time t CAN_H
b V
CAN_H 120V CAN bus line 120V Fig. 4
3.5
a Voltage level of
2.5 CAN_L
CAN_L the low-speed CAN
1.5
SVC0019E
SVC0020E
(CAN-B)
b Voltage level of the
recessive dominant recessive
high-speed CAN
Time t
(CAN-C)
CAN protocol information. To fulfill this purpose, the

Protocol layers transport layer is responsible for such
For communications protocols, it is functions as arbitration or fault detection
standard practice to arrange thematically and signaling.
coherent tasks as a set of layers, which af-
fords a high level of flexibility in the imple- Physical layer
mentation of a bus system. With the CAN, The physical layer is the lowest level in the
both the CAN hardware and the CAN soft- transport stack. It consists of the physical
ware are subdivided into several layers components of the network, such as the
(Fig. 6). wiring and the voltages used to send the
information.
Application layer
The application layer represents the infor- Multimaster principle
mation in the form of data structures used The CAN protocol supports communica-
by the application. These sets of data tion between network nodes without the
which are to be transmitted are forwarded need for a central control unit. Each node
to the object layer for this purpose. may attempt to send messages at any time.
Whether this attempt is successful or not
Object layer essentially depends on two factors:
The task of the object layer is to manage • Is the bus free before the start of trans-
the messages. The functions of this layer mission?

are used to decide which message should • Has the arbitration phase been passed
be sent at which time. For incoming mes- successfully?
sages, this layer is responsible for message
filtering. This design ensures that, even if any num-
ber of nodes were to fail, it can still be de-
Transport layer termined whether a node is authorized to
The transport layer furnishes the object send.
layer with received messages and pro-
cesses the messages prepared by the ob- Content-based addressing
ject layer for sending in such a way that Unlike other networks, the CAN does not
the physical layer is able to transmit this address the individual network nodes but
the messages that have been sent. Each
6 Protocol layers message has a unique marker, or identifier.
The identifier classifies the content of the
message (e.g. engine speed or power-win-
Application layer dow position). A station is therefore able
to broadcast a message to all other stations
(multicast or broadcast method). These
stations read only those messages whose
Object layer
identifiers are stored in their acceptance
list (message filtering, Fig. 7). In this way,
Transport layer
each station decides for itself whether or
not it needs a message sent on the bus.
The identifier has 11 bits (standard format,
CAN 2.0 A) or 29 bits (extended format,
SVC0021E
Physical layer
CAN 2.0 B). With 11 bits in the standard
format, it is possible to distinguish be-
tween 2,048 different CAN messages;
in the extended format, this number rises ority (lowest binary value of the identifier)
to over 536 million. is assigned first access, without any data
The advantage of this addressing loss or delay (non-destructive protocol).
method is that the network nodes do not The arbitration principle permits the
require any information about system dominant bits transmitted by a given sta-
configuration and are thus free to operate tion to overwrite the recessive bits of the
fully independently of each other. This re- other stations (Fig. 8). Each station outputs
sults in a highly flexible complete system, the identifier of its message onto the bus
which makes it easier to manage equip- bit by bit, with the most significant bit first.
ment variants. If one of the ECUs requires During this arbitration phase, each station
new information which is already on the wishing to send data compares the level
bus, all it needs to do is call it up from the present on the bus with the level it actually
bus. It is possible to integrate additional possesses. Each station that attempts to
stations into the system (provided they are send a recessive bit but encounters a dom-
receivers) without having to modify the inant bit loses the arbitration process.
existing stations. The station with the lowest identifier, i.e.
the highest priority, makes its way onto
Controlling bus access the bus without having to repeat the mes-
Arbitration phase sage (non-destructive access control).
If the bus is unoccupied (recessive state) The transmitters of lower-priority mes-
and messages are available for sending, sages automatically become recipients of
each station is free to initiate the sending the message just sent by another station.
of its message. The message begins with a They repeat their attempt to send as soon
dominant bit (start-of-frame bit), followed as the bus is free again.
by the identifier. When several stations Without this access control, bus colli-
start to transmit simultaneously, the sys- sions would result in faults. To guarantee
tem responds by employing “wired-and” unequivocal bus arbitration, therefore,
arbitration (arbiter = logical AND operator) it is not permissible for more than one
to resolve the resulting conflicts over bus node to send a message with the same
access. The message with the highest pri- identifier.
7 Addressing and message filtering 8 Bit arbitration

(acceptance check)
CAN CAN CAN CAN 1

Station 1 Station 2 Station 3 Station 4 Bus line
0
Fig. 7
1 Station 2 transmits
Accept Provision Accept Station 1
0 Stations 1 and 4 accept
the data
1
Selection Send Selection Selection Station 2
message 0 Fig. 8
1 Station 2 gains first
Station 3 access (signal on the
Reception Reception Reception 0
bus = signal from
UAE0742-2E
UAE0284-3E
Station 1 Station 3
loses the loses the station 2)
arbitration arbitration
Bus
0 Dominant level
1 Recessive level
With this access method, maximum-pri- • Error frame:

ority messages only have to wait for the If a station detects a fault or error, it
transfer of the message currently being communicates this to the other stations
sent and, with bit times of 130 (CAN 2.0 A) using an error frame.
or 150 (CAN 2.0 B), they have the lowest • Overload frame:
latency. At a data transfer rate of This can be used to create a delay be-
500 kBits/s, this equates to 260 µs or tween a preceding and subsequent data
300 µs. The higher the load on the bus, frame or remote frame. The transmitting
the greater the data transfer's temporal node reports that it cannot currently
offset becomes for messages of lower pri- process another frame.
ority, and thus the greater the uncertainty
over when a message to be sent will arrive For data transfer on the data bus, a mes-
at the recipient. sage frame is created. It contains the infor-
In order that all messages have a chance mation to be transmitted arranged in a
of accessing the bus, the data transfer rate defined sequence. CAN supports two dif-
must be matched to the number of bus ferent formats of frame, which are speci-
subscribers. fied in CAN 2.0 A and CAN 2.0 B. The most
important difference between these two
Priority assignments frame formats is the length of the identi-
The direct consequence of the arbitration fier. A CAN 2.0 A frame has an 11-bit iden-
process is that the identifier also has the tifier, while a CAN 2.0 B frame has a 29-bit
role of prioritizing the frame during trans- identifier divided into two parts (11 bit
mission in addition to identifying the and 18 bit).
frame content. An identifier correspond- Both formats are compatible with
ing to a low binary number has high prior- each other and can be used together in
ity and vice versa. Message priorities are a network. The frames compliant with
derived from the speed at which the mes- CAN 2.0 A and B are shown in Figure 9.
sage content changes or from the impor- They have a maximum length of 130 bits
tance of the message to safety consider- (standard format) or 150 bits (extended
ations, for example. It must not be possible format).
for messages to have the same priority.
Message format
The message transfer on the CAN bus is 9 CAN message format
based on four different frame formats:

• Data frame: Start of frame
The transmitted message contains data Arbitration field
(e.g. current engine speed) that is pro- Control field
Data field
vided by the transmitting station (data CRC field
source). ACK field
• Remote frame: End of
frame
Stations can request the data they
Inter-
need from the data source (example: frame
the windshield wiper requests how wet space
1
the windshield is from the rain sensor). IDLE 1* 12* 6* 0...64* 16* 2* 7* 3* IDLE
The data source responds by sending 0
UAE0285-3E
the relevant data frame. Data frame

Fig. 9
0 Dominant level Message frame
1 Recessive level
In CAN 2.0 A and CAN 2.0 B data frames, Control field

the sequence of transferred information In a CAN 2.0 A frame, the control field
is arranged in the same way. The start-of- comprises the IDE bit (Identifier Extension
frame bit is followed by the arbitration, Bit), which is always sent as dominant
control, data and CRC fields. A frame is here, followed by a reserved bit for future
completed by the ACK field and the end- extensions, which is sent recessively. The
of-frame. remaining four bits in this field define the
number of data bytes in the next data field.
Start of frame This enables the receiver to determine
The bus is recessive in idle state. The start whether all data has been received.
of frame, represented by a dominant bit, The structure of a CAN 2.0 B frame is
indicates the start of a transmission and, practically identical. However, since the
together with the process of bit stuffing, IDE bit already belongs to the arbitration
serves to synchronize all stations. field, a further reserved bit for future ex-
tensions takes its place and this reserved
Arbitration field bit is sent recessively.
With a CAN 2.0 A frame, the arbitration
field comprises the 11-bit identifier and Data field
a control bit, the RTR bit (Remote Trans- The data field contains the actual mes-
mission Request). sage information comprised of between
With a CAN 2.0 B frame, the arbitration 0 and 8 bytes. A data field in which the
field comprises the 11-bit identifier, fol- length of the data is expressed in 0 results
lowed by the SRR bit (Substitute Remote in the shortest possible data frame with
Request) and the IDE (Identifier Extension a length of 44 or 64 bits. A frame like this
Bit). Both bits are sent recessively and can be used to synchronize distributed
thereby ensure that a CAN 2.0 A frame processes. A number of signals can be
always takes priority over a CAN 2.0 B transmitted in a single message (e.g. en-
frame if it has the same 11-bit identifier. gine temperature and engine speed).
The second 18-bit identifier follows. The
RTR bit completes the arbitration field. CRC field
The RTR bit indicates whether the trans- The CRC field (Cyclic Redundancy Check-
mitted frame is a data or remote frame. sum) contains a 15-bit checksum (frame
The RTR bit is dominant in the data frame check word) across the preceding frame
and recessive in the remote frame. If sta- from the start bit to the final bit of the data
tion A, for example, happens to send a field. The 16th bit (CRC delimiter) in this
message by data frame and station B re- field is recessive and closes the checksum.
quests this message by remote frame at The checksum is a means of detecting
the same time, the arbitration conflict can- possible transmission interference.
not be resolved in this situation by means
of the message identifier: the RTR bit is the ACK field
decisive factor for access authorization. Unlike all preceding fields, the ACK field
First of all, station A wins arbitration (acknowledgment) is not set by the sender
with the sending of the dominant RTR bit of the frame but by a different node that is
and continues to transmit the message. able to acknowledge receipt of the frame
Station B, which has requested precisely directly after the data field. This field
this message, prepares to receive and comprises the ACK slot and the recessive
is able to read the other data sent by ACK delimiter. The ACK slot is also trans-
station A. mitted recessively by the sender and over-
written as dominant by a receiver upon the Cyclic redundancy check

message being correctly received. Here, it For every transmitted message, the trans-
is irrelevant whether the message is of any mitter calculates a check sequence from
significance for the particular receiver in the start-of-frame, arbitration, control and
the sense of an acceptance check (message data field. A 15-bit checksum is derived
filtering). Only correct receipt is con- from the bit sequence by means of a gener-
firmed. This signals to the sender that no ator polynomial. The checksum is then
malfunction occurred during data transfer. used to detect errors in the data transfer.
With CRC generation, a defined genera-
End of frame tor polynomial is used to carry out poly-
The end of frame marks the end of the nomial division across a given frame
message and comprises seven recessive range. The remainder forms the checksum.
bits. With seven bits of the same value in Once the CRC field has been received,
succession, the stuffing rule is deliberately the receiver is able to check that the frame
broken (see Error detection). has been transferred correctly by carrying
out its own polynomial division of the re-
Interframe space ceived frame range using the generator
The interframe space comprises a succes- polynomial and checking whether the re-
sion of three recessive bits which serve to ceived checksum matches the calculated
separate successive messages. After a total remainder.
of 10 of these recessive bits, the stations
are permitted to begin transmitting within Frame check
the network again. Until then, the bus re- The frame check involves all bus subscrib-
mains in idle state. ers, senders as well as receivers, checking
Only data and remote frames need to the sent/received data frame for compli-
include the interframe space. Error and ance with the predefined frame structure.
overload frames can be sent immediately The CAN protocol contains a number of
after the last frame. This enables immedi- fixed-format fields (start and end-of-
ate signaling of errors and problems. frame, delimiter) which are checked
by all stations.
Transmitter initiative
The transmitter will usually initiate a data ACK check
transmission by sending a data frame. It is With the ACK check, a receiver acknow-
also possible for a receiving station to call ledges that the frame has been received
in data from a sending station by deposit- correctly by sending a dominant bit in the
ing a remote frame. The network node ACK slot. The sender of the frame can then
that is able to deliver the information re- verify whether a message was transferred
quested makes the information available correctly. An absence of this bit indicates
in response to this query. that a transmission error has been de-
tected.
Error detection
The bus line may suffer interference, e.g. Monitoring
electromagnetic interference. To eliminate The sender of a frame continuously moni-
the risk of faulty behavior, the transmitted tors the bus level. It is able to detect a
data must be checked for correctness. transmission error by comparing the
A number of control mechanisms for sent bit and the sampled bit.
detecting errors are integrated in the
CAN protocol.
Bit stuffing Fault confinement in the event of failures

Compliance with bit stuffing is checked by Defective stations could weigh heavily on
means of the code check. The stuffing con- bus traffic if they were to frequently send
vention stipulates that in every data frame erroneous messages or interrupt the trans-
or remote frame a maximum of five suc- mission of correct messages by repeatedly
cessive bits of equal state may be sent be- sending an error frame. The CAN protocol
tween the start of frame and the end of localizes station failures by means of
the CRC field. As soon as five identical bit statististical error analysis. A station rec-
states have been transmitted in succes- ognizes the probability of its own malfunc-
sion, the sender includes a bit with the option by how often it aborts messages be-
posite state. The receiving station clears fore other stations send an error frame.
all of these inserted bits after the message The protocol's first measure in this case
has been received (destuffing). is to prevent a station such as this from
This measure permits detection of line continuing to abort transmissions. In an
faults, e.g. short-circuit or burst interfer- emergency, the station shuts down auto-
ence. Rare signal changes would affect matically.
the possibility for synchronization in the
nodes.
Error handling
If a CAN controller detects a fault or for-
mat error, it interrupts the current trans-
mission by sending an error frame com-
prising six successive dominant bits. This
breaks the stuffing rule that prohibits this
type of bit sequence. If the sender detects
that its message has been interrupted by
an error frame, it stops transmitting and
makes another attempt at a later time.
This effect prevents other stations from
accepting the erroneous message and
thereby ensures consistency of data
across the entire system.
Hardware If messages on the bus are managed in

CAN controller of a subscriber extended format (29-bit identifier), the
The CAN controller of a subscriber gener- CAN controllers must be compatible with
ates a frame from the data to be sent con- the CAN 2.0 B specification. There are
taining all the fields required for compli- controllers that only support CAN 2.0 A
ance with the CAN protocol. It then con- and send and receive messages in the
verts the frame into a bit stream. standard format, but that generate errors
CAN only specifies the physical, trans- with messages in the extended format.
portation and object layers. The interfaces Other controllers are able to tolerate the
that are provided in the application layer extended format without generating an
can be arranged differently to suit the field error. These modules can be used in a CAN
of application. In order to provide proper together with controllers that administer
CPU support (electronic control unit mi- the sending and receiving of messages in
crocontroller for application software) for the extended format.
a wide range of different requirements,
semiconductor manufacturers have Basic CAN
brought onto the market products that In modules with basic-CAN implementa-
provide a variety of capabilities. They dif- tion, only the basic function of the CAN
fer neither in the frame format they pro- protocol for the generation of the bit
duce, nor in their error-handling methods, stream is implemented in the hardware.
but solely in the type of CPU support re- For the management of the messages to be
quired for message administration. sent and received, there is an intermediate
10 Basic-CAN module 11 Full-CAN module
Microcontroller Microcontroller
Output Input
buffer buffer
CAN
controller
Control
unit
System bus interface
Transmission Test for: Control
control unit acceptance, unit
faults
Transmit Receive
buffer buffer
Basic CAN
controller Transmission Test for:
Transmit Receive control unit acceptance,
buffer buffer faults
CAN TxD CAN TxD

CAN RxD CAN RxD
CAN transceiver CAN transceiver
CAN_H CAN_L CAN_H CAN_L

SVC0023E
SVC0022E
CAN bus CAN bus

buffer to which the local computer (appli- Modules without local computer
cation software microcontroller) has ac- A further category of CAN module is one
cess (Fig. 10). Since the buffer capacity that is supplied without a local computer.
is limited, the computer must read the These SLIOs (Serial Linked Input/Output)
received data before new messages are are able to input and output data via ports.
received. Message filtering also takes They are therefore suitable for making
place in this computer. A part of the sensors and actuators bus-compatible at
computer's capacity is therefore used for low cost, but they do need a master that
CAN management. Since the computers controls them.
do not usually have sufficient processing
capacity, modules with basic CAN are pri- Transceiver
marily suitable for low bit rates, or for the The bit stream generated by the CAN con-
transmission of fewer messages but at troller is made up of binary signals. They
higher bit rates. do not yet correspond to the required volt-
The advantage of these modules, in com- age levels of the CAN bus. The CAN-bus in-
parison to modules with full CAN, is the terface module, or transceiver, generates
smaller chip surface and the lower manu- the differential signals CAN_H and CAN_L
facturing costs. and the reference voltage Uref from the
binary data stream.
Full CAN
Full CAN implementation is the protocol of Sleep mode
preference in cases where a station has to The CAN comfort bus must remain ready
manage several messages at high bit rates for operation even with the ignition
and the local computer has no free capac- switched off so that functions such as the
ity for communication tasks. They contain radio, power windows or parking lamp
several “communication objects”, each of may continue to operate. The bus sub-
which contains the identifier and the data scribers must therefore be supplied by
of a particular message. During the initial- terminal 30 (permanent positive). After
ization of the CAN module by the local terminal 15 has been switched off (ignition
computer, it is decided which messages off), a CAN node may enter sleep mode
the CAN controller should send and which (standby) to relieve the vehicle electrical
received messages it should process fur- system of as much load as possible. The
ther. Received messages are only accepted transmitter part of the transceiver module
(message filtering) if the identifier matches is switched off in this condition to mini-
one of the communication objects. mize the power consumption in this mode
CAN controllers with full-CAN imple- of operation. However, the receiver part
mentation relieve the burden on the local remains active and checks whether mes-
computer by performing all of the commu- sages are being sent on the bus. In this
nication including message filtering in the way, the CAN controller, which also enters
controller (Fig. 11). standby mode, is able to react to a wake-up
The CAN controller can be coupled to message and fully activate the CAN node.
the microcontroller in the electronic con-
trol unit as a stand-alone module by the
address/data bus. Powerful microcon-
trollers have the CAN controller integrated
on-chip. This type of bus coupling is the
more cost-effective and thus the more
common solution.
Data transfer sequence Once the engine-speed information has

Using the transfer of engine speed as an been stored in the transmit buffer, the
example, it shall be demonstrated how CAN controller begins to compile the CAN
data is transferred on a CAN with full-CAN frame. Over the RxD line, the CAN con-
module. troller now monitors the bus. If the bus is
free, it immediately starts to generate the
Transmission bit stream and directs it to the transceiver
The engine management's application soft- on the TxD line; if the bus is occupied, it
ware calculates the engine speed from the waits. From the bit stream, the transceiver
signal from the engine-speed sensor. This creates a signal with the requisite voltage
value is calculated once for every combus- level.
tion cycle. The measured value enters both
the receive buffer and the transmit buffer Reception
of the microcontroller (Fig. 12a). The message originally sent by the engine
The engine-management microcon- management and forwarded on by the
troller is coupled to the CAN controller transceiver is received by all stations
by a parallel interface. The contents of connected to the bus (Fig. 12b). The mes-
the transmit buffer are cyclically passed sage reaches the CAN controller on the
(e.g. every 10 ms) to the transmit buffer RxD lines in the form of a bit stream.
of the CAN controller. A flag notifies the At the first stage (monitoring layer),
CAN controller that a message is ready the CAN controller checks the incoming
to be sent. With this transmit instruction, message for errors using the CRC check-
the engine management has completed its sum. If the message is free of errors, each
part of the task.
12 Data transfer sequence
a b
Engine control unit ABS control unit Instrument cluster
Microcontroller Microcontroller Microcontroller
n n
Rev
counter
output
CAN
area n n
CAN controller CAN controller CAN controller

Transmit Transmit Transmit
buffer buffer buffer
RX TX RX TX RX TX
Transceiver
SVC0024D
SVC0024E
CAN bus
station responds with an acknowledge CAN is also widely used in industrial auto-
(ACK check). mation. These applications are supported
At the second stage, or acceptance layer, by an alliance of companies in the “CAN in
the message undergoes message filtering. Automation” users group (CiA).
Each station checks whether the received Bosch has concluded contracts with its
identifier is addressed to the particular licensees that guarantee that any CAN im-
station and whether the message is re- plementations will be able to communicate
quired in the application software. If not, with each other. Users will be able to rely
the message is rejected. Otherwise, it on the interaction of any CAN modules.
makes its way to the receive buffer. A flag
notifies the application software that a Characteristics
new message is ready for processing. • Standardized in accordance with
The instrument cluster, for example, ISO 11898
calls up the available message, processes • Prioritized communication
the engine-speed information and calcu- • Data transfer rates: up to 1 MBits/s
lates triggering signals for the actuator of • Data capacity: up to 8 bytes per message
the rev counter. • Real-time response: the data protocol is
sufficient for the real-time requirements
Standardization in the motor vehicle
The International Organization for Stan- • Non-destructive bus-access method
dardization (ISO) and SAE (Society of Au- • Low power consumption
tomotive Engineers) have issued CAN stan- • Flexibility of configuration
dards for data exchange in automotive ap- • Simple and economical design with
plications: twisted line pairs

• For low-speed applications up to • Very high reliability of data transfer
125 kBit/s: ISO 11 519-2 and 11 898-3 • Fault detection and signaling
• For high-speed applications faster than • Localizing of failed stations
125 kBit/s: ISO 11 898-2 and SAEJ 22 584 • Handling of intermittent and permanent
(passenger cars) or SAEJ 1 939 (commer- faults
cial vehicles) • Short-circuit resistance
• An ISO Standard for diagnosis via CAN • The number of nodes is theoretically un-
has also been published as ISO 15 765 limited. However, a limit arises in prac-
tice from the capacitive load of the bus
Standardization makes it possible for and the increasing latencies of messages
components of different manufacturers when a high number need to be sent.
to function together. No adaptations are
required.
84 Bus systems | LIN bus
LIN bus The LIN bus is suitable for low data rates
of up to 20 kBit/s and is typically limited to
Overview a maximum of 16 bus subscribers.
The increasing use of mechatronic systems The electrical interface can be created
in the motor vehicle gave rise to the idea easily and cost-effectively in the network
of designing a cost-effective bus system nodes. As far as the nodes are concerned,
as an alternative to the low-speed CAN. a distinction is made between the master,
In 1998, several automotive manufacturers which is generally an electronic control
founded a consortium with the aim of de- unit connected to a superordinate bus sys-
veloping a specification for a serial bus for tem, and the slaves. These are intelligent
the networking of sensors and actuators in actuators, intelligent sensors or, quite sim-
the body electronics area. ply, just switches with additional hardware
It was believed that a bus system with for the LIN-bus interface.
simple bus protocol and a simple sequence The bus subscribers are usually ar-
control would make it possible to use even ranged in a linear bus topology and con-
low-capability microcontrollers without nected to each other by a single-wire line.
additional hardware for the communica- This topology, however, is not explicitly
tion interface. specified.
The workgroup's resulting LIN bus spec- Communication on the LIN bus takes
ification was introduced into mass produc- place in a time-synchronous manner,
tion with the Mercedes-Benz SL as early as whereby the master defines the time grid.
2001. Consequently, there arises a strictly deter-
The name, LIN (Local Interconnect ministic LIN bus response.
Network), is derived from the fact that Figure 1 shows an example of a LIN net-
all electronic control units are located work as a subbus in the roof/wiper area of
within a demarcated installation space the motor vehicle. Here, the bus comprises
(e.g. in the door). The LIN, therefore, is a central electronic control unit, as the
a local subsystem for supporting the vehi- master, and the four slaves: mirror, garage-
cle network by means of superordinate door opener, rain/light sensor and wiper
CAN networks. actuator. The master also functions as a
gateway to the Chassis CAN, the Body CAN
and the Diagnostics CAN.
1 LIN bus with master and slave nodes
Slave 1 Slave 2 Slave 3 LIN slave 4

Mirror Garage-door Rain/light Wiper
opener sensor actuator
Diagnostics
CAN Master
Central ECU
Roof/wiper LIN
Body CAN Chassis CAN

SVL0001E
Bus systems | LIN bus 85
Applications Data transmission system

The LIN bus as a means of networking me- The LIN bus is designed as an unshielded
chatronic systems can be used for many single-wire line. The bus may adopt two
applications in the motor vehicle for which logical states:
the bit rates and variability of the CAN bus • The dominant level corresponds to the
are not essential. Examples of LIN applica- electrical voltage of approx. 0 V and
tions: represents logical 0
• Door module with door lock, power-win- • The recessive level corresponds to
dow drive and door-mirror adjustment battery voltage UBat and represents
• Control of the power-sunroof drive unit the logical 1 state
• Control of the wiper motor for the wind-
shield wiper The recessive level is characterized by a
• Sensor for rain and light detection resistance to battery voltage of 1kV in the
• Air-conditioning system (transmission master and 30kV in the slaves.
of signals from the control element, Due to circuitry variations, there may be
activation of the fresh-air blower) differences in the voltage levels. A stable
• Headlight electronics data transfer is ensured by the defining of
• Control of motors for seat adjustment tolerances for sending and receiving in the
• Theft deterrence areas of the recessive and dominant levels.
• Garage-door opener The tolerance zones on the reception side
are wider (Fig. 2) to make it possible for
valid signals to be received despite inter-
ference radiation.
The data rate of the LIN bus is limited
to a maximum of 20 kBit/s. This is the out-
2 Tolerance ranges for transmitters and receivers come of a compromise between the de-
mand for high edge steepness for easy syn-
Transmitter chronization of the slaves on the one hand,
UBat
recessive
and the demand for lower edge steepness
80% for improved EMC characteristics on the
other. The standard bit rates of 2,400 Bit/s,
Voltage level
9,600 Bit/s and 19,200 Bit/s are recom-

mended. The minimum permissible value
20% for the bit rate is 1 kBit/s to prevent time-
dominant out conflicts. The edge steepness itself
0
Rise time Release time is defined in the LIN specification as
Time t 1 to 3 V/µs.
The maximum number of nodes is not
specified in the LIN specification. It is
Receiver
UBat theoretically limited by the number of
available identifiers. In practice, the num-
recessive
ber of subscribers is restricted to 16 due
Voltage level
60% to the maximum permissible total capacity

40% of the bus system.
dominant
SVL0002E
0
Time t
Bus access LIN protocol

Access to the LIN bus is determined by the Frame
master-slave principle. Each message is The information transferred on the
initiated by the master. The slave has the LIN bus is embedded in a defined frame
possibility to respond. The messages are (Fig. 3). A message initiated by the master
exchanged between the master and one, always begins with a header. The response
several or all of the slaves (point-to-point, (message field) contains different informa-
multicast, broadcast). tion depending on the type of message.
The following relationships are possible If the master wishes to transmit control
in the communication between master and instructions for a slave, it will populate the
slave: response with data to be used by the slave.
• Message with slave response: the master If the master is transmitting a data request,
sends a message to one or more slaves the slave that it is addressing will populate
and asks for data (e.g. switch states or the response with the data requested by
measured values) the master.
• Message with master instruction: the
master gives a slave a control instruction Header
(e.g. switch on a servomotor) The head is composed of the following
• The master initiates communication parts:
between two slaves • The synchronization break
• The synchronization field
No arbitration or conflict management is • The identifier field
required because, with the master-slave
access control, the master alone controls Synchronization
access to the data line. To guarantee consistent data transfer
between master and slaves, a synchroniza-
tion takes place at the beginning of each
frame. First of all, the beginning of a
frame is unambiguously marked by the
synchronization break (SynchBreak).
3 LIN message format
LIN frame
Header Response
Synch Synch Ident Data Data Data Data Data Data Data Data Check-
Break Field Field field 0 field 1 field 2 field 3 field 4 field 5 field 6 field 7 sum
SVL0003E
The SynchBreak comprises at least 13 con- Of the 64 possible messages, 32 may only
secutive dominant levels and one recessive contain two data bytes, 16 may only con-
level. tain four data bytes, and the remaining 16
At the end of the synchronization break, eight data bytes each.
the master sends the synchronization field The last bits in the identifier field con-
(SynchField) consisting of the bit sequence tain two checksums, which are used to
01010101. The slaves are then able to ad- check the identifier for transmission er-
just themselves to the time basis of the rors and any resulting incorrect message
master and thus synchronize. assignments.
The synchronization method described
permits a loose specification of the timing Data field
of the bus subscribers. The clocking of Once the header sent by the master node
the master should not deviate more than has been transmitted, it is time for the
±0.5 % from the nominal value. The clock- transfer of the actual data to begin. The
ing of the slaves is permitted to deviate by slaves know from the transmitted identi-
up to 15 % before synchronization as long fier whether or not they are being ad-
as the synchronization reduces the devia- dressed and, if they are, they reply with
tion to a maximum of 2 % before the end their response in the data field.
of the message. Several signals can be packed into one
In this way, the slaves can be built with- frame. In this case, each signal has pre-
out an expensive quartz oscillator, e.g. cisely one generator, i.e. it is always writ-
using a simple RC circuit. ten by the same node of the network.
During the data transfer of the bytes,
Identifier it is always the least significant bit (LSB)
The third byte in the header is used as the that is output first. Each byte (8 bits) is
LIN identifier. As with the CAN bus, a con- preceded by a start bit and followed by
tent-based addressing method is used – a stop bit, which means that each byte
the identifier therefore provides informa- involves the transmission of ten bits.
tion about the content of a message (e.g. The purpose of the start and stop bits is
engine speed). Based on this information, to resynchronize the nodes and thereby
all nodes connected to the bus decide prevent transmission errors.
whether they would like to receive and The data response of the slaves is veri-
process the message further or simply fied by means of a checksum.
ignore it. This process is known as accep-
tance filtering. LIN description file
Six of the eight bits of the identifier field The configuration of the LIN bus, in other
determine the identifier itself. Their per- words the specification of network sub-
mutations give rise to a possible 64 differ- scribers, signals and frames is managed
ent identifiers (ID). They have the follow- in the LIN description file, or ldf. For this
ing meanings: purpose, the LIN specification provides for
• ID = 0 to 59: transmission of signals. an appropriate configuration language.
ID = 60: master request for the com- From the ldf, a set of C codes or header
mands and diagnosis files is automatically generated using suit-
• ID = 61: slave response to ID 60 able tools. These are used as the basis for
• ID = 62: reserved for manufacturer- implementing the master and slave func-
specific communication tions in the electronic control units con-
• ID = 63: reserved for future extensions nected to the bus.
to the protocol
The ldf is therefore a means of configuring Example: air-conditioning control

the entire LIN network. It represents a The LIN bus is typically used in the control
common interface between the vehicle of the air-conditioning system. The operat-
manufacturer and the supplier of the ing and display unit functions as the bus
master or slave modules. master. This is where the software for the
open and control algorithms are stored.
Message scheduling One of its tasks is to adjust the speed of
The scheduling table in the ldf defines the the fresh-air blower. The controlled vari-
sequence and time grid in which messages able for this is the actual temperature in
are sent. Frequently needed information is the passenger cell and the desired temper-
sent more often. Once the table has been ature set by the driver. The electronic con-
worked through, the master begins with trol unit receives the interior temperature
the first message again. The sequence in from a temperature sensor located in an
which the table is worked through may appropriate place in the interior. The de-
change depending on the operating status sired temperature is set at the control unit,
(e.g. diagnostics active/inactive, ignition e.g. using a sensor ring.
on/off). From the input variables, the electronic
The transfer grid of each message is control unit calculates that the blower
therefore known. This deterministic be- speed needs to be increased to a value of
havior is guaranteed by the fact that, with 200 rpm, for example. The electronic con-
the master-slave access control principle, trol unit transmits a message containing
all messages are initiated by the master. the master instruction to the LIN bus at a
defined time interval. The identifier in this
Network management example would be “set the blower speed”.
The nodes of an LIN network can be forced In this subbus, this instruction corre-
into sleep mode to minimize the no-load sponds to identifier 25, for example.
current of the entire electronics and elec- Following the header containing this iden-
trical system in the vehicle. Sleep mode tifier, the master transmits a numerical
can be achieved in two ways: value in the data field that equates to the
• The master sends the go-to-sleep com- physical value of 200 rpm. Each slave pos-
mand with the reserved identifier 60 sesses a list of identifiers that are relevant
• The slaves enter sleep mode of their own for this node. The fresh-air blower is the
accord if no data transfer has taken one and only slave that responds to the
place on the bus for a relatively long “set blower speed” identifier and executes
period (4 seconds) the master's request.
At speeds below idle, the fresh-air
Both the master and the slaves are able to blower must be switched off. At this low
wake up the network. To do so, they must speed, the heavy load could cause the en-
send the wake-up signal. This comprises gine to stall. The LIN bus is coupled to the
data byte 128. After a break of 4 to 64 bit CAN bus via the gateway and thus receives
times (wake-up delimiter), all nodes must the current engine speed on a continuous
have been initialized and be ready to re- basis. If the speed falls below the specified
spond to the master. engine-speed threshold, the LIN master
sends the message containing the “set
blower speed” identifier where the data
field contains the value 0. The fresh-air
blower switches off in response.
4 LIN networking (example)
LIN slave 1 LIN slave 2 LIN slave 3 LIN slave 4

Fresh-air Windshield PTC auxiliary PTC auxiliary
blower heating heater, right heater, left
LIN master 1
ECU for air-con-
ditioning system
LIN slave 1
CAN
Sliding sunroof
motor
LIN master 2
SVL0004E
Roof module
Intelligent actuators can send up-to-date Summary

operating-status information to the actuat- The essential characteristics of the LIN bus
ing unit. The fresh-air blower records the are:
speed by means of a sensor and sends it • Single-master/multiple-slave concept
back on the LIN bus as a numerical value. • Master/slave access control
Due to the master-slave access method, • Independent synchronization of the
the value is only contained in a message slave possible even without quartz
with slave response initiated by the mas- • Deterministic signal transmission
ter. The feedback signal with the current • Communication in the form of very
engine speed makes it possible to control, short messages

and therefore accurately maintain, the • Character-based transmission (UART)
desired value. • Bit rate max. 20 kBit/s
• Data transfer over an unshielded single-
wire line
• The reference potentials of the data line
are battery voltage and ground
• Maximum bus length 40 m
• Maximum number of nodes 16, typically
fewer than 12
90 Bus systems | Bluetooth
Bluetooth for transmitting data for multimedia

applications.
Overview In the meantime, the Bluetooth SIG has
Mobile communication is gaining ever expanded to involve around 2,000 com-
more importance in all domains. For con- panies from the telecommunications, data
venient communication between various processing and automotive engineering
devices, a wireless link is indispensable. fields.
Infrared connections, as often used in
the past, tended to be manufacturer- Applications
specific, they required a direct line of Bluetooth is an industry standard for
sight and imposed constraints on the the networking of mobile multimedia
area of movement. Only with a standard- devices, such as a car sound system, cell
ized, wireless link could mobile devices phone, headset, PDA (Personal Digital
of different manufacturers communicate Assistant), PC and peripheral equipment
with each other without problems. (Fig. 1). Bluetooth stands for the simple
The development of the Bluetooth exchange of data between portable ter-
standard began in 1994 at the telecom- minals and the wireless transmission of
munications company, Ericsson. As part audio and video signals for entertain-
of an initial study, Ericsson investigated ment and information.
the possibilities for completely replacing The short-range wireless connectivity
the cable connections between the cell of Bluetooth eliminates the need for con-
phone and additional devices. In cooper- nection cables between the cell phone
ation with other industry partners, the and the hands-free system: no more
Bluetooth SIG (Special Interest Group) tangling of cables when making a call.
was formed in 1998 to create a uniform, It is in this area that Bluetooth has been
internationally-accepted standard. The the most well received.
SIG set itself the aim of specifying a wire- Meanwhile, a large number of con-
less technology with low manufacturing sumer electronics devices, be they cell
costs, low energy consumption and ro- phones, PDAs, notebooks or car phones,
bust resistance to interference. The now offer an address book function.
wireless interface had to be suitable
1 Bluetooth networking of mobile multimedia devices (example)
1 2 3
Fig. 1
1 Cellular phone
2 Headset 5 4
3 Personal digital
assistant (PDA) FM2 EUROPE 1
SVB0001Y
4 Laptop computer
5 Car sound system
with navigation unit
Bus systems | Bluetooth 91
With Bluetooth, it is possible to conve- the specification, great emphasis was

niently synchronize the entries stored placed on backward compatibility with
in the various address books. The car the previous versions.
driver is spared the laborious task of
typing the phone number stored in the Transmission technology
cell phone or PDA or inputting an ad- Frequency band
dress into the navigation system. Bluetooth operates in the worldwide-
Bluetooth also makes it possible to license-free 2.4 GHz ISM band (Industrial
listen to the radio through a headset in Scientific Medicine). This is available
the rear of the vehicle and to control practically worldwide without approval
the car sound system using a PDA. – apart from certain national limitations –
and is reserved for applications in indus-
Bluetooth is predominantly used for try, science and medicine. For this rea-
the transfer of data in the multimedia son, the possibility of interference from
domain. However, this technology can garage-door openers, microwave ovens
quite easily be used for diagnostic and or other appliances which operate using
service purposes. Workshops or break- the same frequency band has to be ac-
down services would be able to conve- cepted. To minimize interference, the
niently read vehicle information during best possible use should be made of the
fault diagnosis and record the condition frequency spectrum. Devices that use
of the vehicle. the ISM band therefore implement
spread-spectrum techniques.
Bluetooth versions
The first specification (Bluetooth 1.0) of Frequency-hopping method
the Bluetooth standard was adopted in Bluetooth divides the 2.4 GHz band
July 1999. In December 1999 this specifi- into 79 channels at intervals of 1 MHz
cation was replaced by Bluetooth 1.0b, (f = 2,402 + k, k = 0…78). The band is
which contained improvements and clar- terminated at both the upper and lower
ifications for interoperability. February end by two guard bands with a width of
2001 saw the release of the Bluetooth 1.1 3.5 MHz (upper guard band) and 2 MHz
specification, which fixed fundamental (lower guard band). The transmission
interoperability problems between the
first devices and chips and is used in
most Bluetooth devices today. Three 2 Frequency-hopping method
years later, in November 2003, version

1.2 of the specification was published.
This used the IEEE (Institute of Electrical
and Electronics Engineers) language def-
inition for technical documentation and
introduced adaptive frequency hopping
Frequency f
(AFH) as a technical advancement that

improves resistance to interference in
the frequency band. The most significant
modification was introduced with ver-
sion 2.0 of the specification in November
2004. With the enhanced data rate (EDR), Time slots Time t
SVB0002E
it was possible to triple the bit rate from

1 MBit/s to 3 MBit/s and reduce power
consumption. With each new version of
uses a combined frequency-hopping/ 8 differential phase-shift keying (8DPSK),

time division duplex (FH/TDD) method: which modulate the phase of the carrier.
the frequency-hopping spread spectrum p/4DQPSK represents one of four com-
(FHSS, Fig. 2). The channels are switched binations of a bit pair (00, 01, 10 and 11)
1,600 times per second (i.e. in 625 µs and so transfers a maximum of two bits;
slots). The slots are assigned in accor- 8DPSK can represent any combination
dance with the TDD method, i.e. the of three bits. This is used to increase the
transmitter and receiver are authorized bit rate from 1 MBit/s (net 727 kBit/s) to
to send alternately. The frequency-hop- 2 MBit/s (net 1,446 kBit/s) with GFSK,
ping method guarantees optimal and or 3 MBit/s (net 2,169 kBit/s) with EDR.
uniform use of the ISM band and makes
Bluetooth resistant to interference from Power classes
other transmitters in the same frequency Bluetooth devices are categorized into
band (e.g. WLAN). In addition, it also different power classes. Manufacturers
offers some security because a hacker are free to decide which power class they
would not know the destination of a fre- implement. The present specifications
quency switch or indeed the time frame define three power classes:
for which a particular frequency would • Class 1 with 100 mW (20 dBm) trans-
be valid. This makes interception of a mission output for a range of 100 to

connection much more difficult. 150 m
The introduction of adaptive fre- • Class 2 with 2.5 mW (4 dBm) transmis-
quency hopping (AFH) in the Bluetooth sion output for a range of 10 to 25 m

1.2 specification further increased • Class 3 with 1 mW (0 dBm) transmis-
Bluetooth's resistance to interference sion output for a range of 10 m

and reduced the effect of Bluetooth itself
as a source of interference on other de- Topology
vices in the 2.4 GHz ISM band. The num- Bluetooth supports the automatic config-
ber of effective channels is dynamically uration of ad-hoc networks. This means
adaptable to the conditions in the fre- that two or more devices are able to
quency band. Frequencies used by other spontaneously form a network with no
transmitters are then removed from the prior knowledge of each other. Bluetooth
list of 79 possible channels. is based on the master-slave principle.
In principle, any Bluetooth device may
Modulation method be a master or a slave. The master is a
Bluetooth uses the Gaussian frequency- special device that coordinates commu-
shift keying (GFSK) method of modula- nication between the devices.
tion. This is a special variant of fre-
quency-shift keying (FSK) supplemented
with a Gaussian low-pass filter. GFSK
varies the frequency and delivers 1 bit
per keying interval. Bluetooth has a sym-
bol rate of 1 MBit/s. With the enhanced
data rate (EDR) introduced with the
Bluetooth 2.0 specification, it has been
possible to increase this data rate to
3 MBit/s. This is made possible by the
use of two variants of phase-shift keying
(PSK): the p/4 differential quaternary
phase-shift keying (p/4DQPSK) and the
Piconet The piconet is defined by the frequency-

A network of Bluetooth devices is known hopping sequence, i.e. the sequence of
as a piconet (Fig. 3). For the devices switching between the 79 possible chan-
within a piconet, there are two roles: nels. The hopping sequence is specified
master or slave. Which device assumes by the master and is calculated from the
which role is not decided until communi- master's device ID (a worldwide unique
cation is established. A device sets up a 48-bit identifier) and its internal clock.
piconet and transmits its device ID and The slaves synchronize with the master's
the value of its internal clock. This de- clock and its hopping sequence.
vice becomes the master. A master is able to manage up to seven
Any device that is not participating in active slaves. Each active slave receives
a piconet but is not switched off remains a 3-bit active member address (AMA).
in standby, which is marked by a small Active devices can send data or even
current draw. simply remain connected. By deregister-
The first stage in the creation of a ing, a device can switch back to standby.
piconet involves the potential master A Bluetooth device may also switch to
searching for other devices in its range any of three energy saving modes: hold
(inquiry scan). Special frequency-hop- mode, sniff mode and park mode. The
ping sequences (inquiry sequences) are devices continue to be synchronized with
available for this purpose. Devices that the hopping sequence of the piconet. The
wish to enter into communication with lowest energy saving mode is the park
other devices respond with an inquiry mode, where the device relinquishes its
response message. By the end of the in- AMA and receives in its place an 8-bit
quiry sequence, the information about all parked member address (PMA). With the
devices willing to communicate is avail- PMA, the device is still a subscriber in
able. The potential master is now able to the piconet but makes way for another,
address a specific device by means of a active device. Parked devices remain
special paging sequence. Paging is a pro- synchronized with the hopping sequence
cess whereby a fixed connection is es- of the piconet and can be addressed and
tablished between a master and a slave. reactivated by the master using the PMA.
The slave contains all the information it
needs to synchronize with the master. Scatternet
In a piconet, only one slave is ever able
3 Piconet to communicate with the master simulta-
neously. When more slaves enter the net-
work, the data rate per device falls very
quickly. This is where scatternets pro-
vide a solution. A scatternet consists of
S up to ten piconets in which the individ-
P P ual piconets partly overlap. In this way, it
is possible to set up even larger networks
M
in which all network subscribers may
S S connect to one another. The scatternet
in Figure 4 comprises two piconets. One
SB
of the Bluetooth devices is a member of
Fig. 3
both piconets. Through this device, the M Master
SVB0003E
two networks are able to exchange data. S Slave

However, this device, which is assigned P Park mode
SB Standby
to more than one piconet, can only func- It is conventional for the transmissions
tion as master in one of these networks. of the master to begin in the even time
A device that is assigned to two picon- slots and those of the slaves to begin
ets in a scatternet must synchronize with in the odd time slots. This characteris-
the network with which it wishes to com- tic conforms to the regular time-divi-
municate. It cannot remain connected to sion multiplexing method (TDD here)
both networks at the same time. Before whereby the master uses one half of the
the device leaves the old piconet, it must time slots and the slaves the other half.
inform its master that it will be unavail- In addition to packets that may occupy
able for a certain period of time. The one time slot, Bluetooth provides for
remaining network subscribers can packets that may occupy three or five
continue to communicate. time slots (multi-slot packets). As soon
A master, too, may leave its own as a master or slave sends a packet three
piconet and become a slave in another. or five time slots long, this transmitter
All communication in its old network, will remain on the same frequency.
however, is broken until it returns and There is no frequency change within
assumes its master function again. a packet (Fig. 6).
A master cannot become master of a Once the packet has been transferred,
second piconet. If this were to happen, the frequency hops as determined by the
both piconets would behave in an identi- hopping sequence (independently of the
cal manner and form a single network. transmission process). In the illustration,
for example, the frequency hops from
Physical data channel fk to fk+3 after the transmission of a 3-slot
A data channel is represented by a hop- packet). The reason for this behavior is
ping sequence between the 79 possible that not every station may have received
frequencies in the ISM band. Any device the transmission and cannot therefore
that is actively participating in a piconet respond specifically to the transfer of
must also hop at the same carrier fre- data in several time slots. For this rea-
quency (frequency fi). As soon as a mas- son, all stations that are not involved in
ter has sent data at frequency fk, a slave the transmission always continue to hop
can respond at frequency fk+1. This pro- in the sequence specified by the master.
cedure is illustrated in Figure 5.
4 Scatternet 5 Timing of data transfer
Piconet 1 Piconet 2
S SB f(k) f(k+1) f(k+2)

Master
P P S
M M
S S P Slave
P S 625 ms
Fig. 4
M Master
SVB0004E
SVB0005E
S Slave
P Park mode
SB Standby
Physical connections which transfer data from the user level –

Bluetooth supports both circuit-switched the applications.
and packet-switched data channels. For
circuit-switched, synchronous communi- Data packets
cation, Bluetooth offers a synchronous A Bluetooth data packet generally
connection-oriented link (SCO), and an comprises three fields:
ACL (Asynchronous Connectionless Link) • Access code
for packet-switched, asynchronous • Header and

communication. Between two Bluetooth • Payload
devices, there may only ever be one
ACL link but there can be up to three A packet may simply comprise the access
SCO links. The ACL links are the founda- code, or the access code and a header,
tion for the data transfer. The SCO links or the access code, header and payload
are purely voice channels with a data (Fig. 7).
rate of 64 kBit/s. The voice is transmitted A packet begins with an access code.
by continuously variable slope delta If a header follows, the access code has
modulation (CVSD) or pulse-code modu- a length of 72 bits, otherwise it is 68 bits
lation (PCM). A Bluetooth master can long. The access code is used for the
maintain several different data channels synchronization and identification of a
between its slaves even at the same time. piconet, for device polling and for device
calls. There are three different types of
Logical data channels access code:
Logical channels refer to different types • Channel access code (CAC), the purpose
of channels streamed over a physical of which is to synchronize and identify

connection. The data transmitted over a piconet. The CAC is sent with each
a physical channel has different logical packet transmitted over the piconet.
meanings. Bluetooth distinguishes be- • Device access code (DAC), which is used
tween two categories of channel: link to transfer specific identifiers during
channels, which are used for the ex- the device call (e.g. when paging).
change of control information between • Inquiry access code (IAC), which is
master and slave; and user channels, used for device polling.
6 Frequency hopping with multi-slot packets 7 Structure of a packet
Access code Header Payload

f(k) f(k+1) f(k+2) f(k+3) f(k+4) f(k+5) f(k+6)
72 (68) bits 54 bits 0 to 2,745 bits
625 ms 4 preamble 3 AMA
f(k) f(k+3) f(k+4) f(k+5) f(k+6) 64 synchron- 4 type

ization
1 flow
(4) appendix
f(k) f(k+5) f(k+6) 1 ARQN
1 SEQN
SVB0007E
SVB0006E
8 checksum
The header of a Bluetooth packet is Bluetooth architecture

54 bits long and contains information The Bluetooth architecture is complex.
about the link (link control information). This is inherently associated with the
It is made up of the address, packet type, idea behind general cable emulation.
flow control, error monitoring and The architecture is designed to enable
checksum fields. The actual length of new protocols and applications to be
the header (i.e. the sum of the bits in adapted for the use of Bluetooth. Accord-
the header) is 18 bits. The header is pro- ingly, the Bluetooth protocol stack is
tected by forward error correction (FEC). very extensive (Fig. 8 shows a simplified
The bits of the header are sent three version of the protocol stack).
times (1/3 FEC), giving a header length of The Bluetooth technology can be sub-
54 bits. A receiver can then simply take divided into three logical sections:
the majority decision; each bit triplet is • Hardware
resolved to the value with majority in • Host stack
the triplet. • Applications
The payload of a Bluetooth packet can

be up to 341 bytes in size (1,023 bytes The hardware layer and the host stack
with EDR). The data field of the payload are defined in the Bluetooth Core Specifi-
consists of up to three segments: the cation; the applications are part of the
payload header, the payload itself and, Bluetooth Profile Specification.
in certain circumstances, a checksum.
Bluetooth uses 16 different packet Hardware-side protocols
types, the common feature of all being The hardware is represented by the
the access code and the header. The Bluetooth chip. This is where the radio
structure of the payload depends on the layer, baseband and link management
packet type concerned. Along with pack- are contained. The radio layer deals with
ets for ACL and SCO links, there are also the radio transmission, used frequencies,
packet types for polling the slaves, syn- modulations and transmission output.
chronizing the hopping sequence and The baseband represents the link-estab-
acknowledging data transfers. lishment mechanisms, the packet struc-
ture and the timing. The link management
Bluetooth device addresses establishes and manages the link between
A Bluetooth device is assigned a world-
wide unique Bluetooth device address. 8 Protocol stack
The address is derived from the IEEE-

802 standard and implemented in the de- Applications
vice by the manufacturer. The 48-bit ad-
OBEX
dresses are subdivided into three parts
• LAP (Lower Address Part, containing SDP TCS BIN RFCOMM
24 bits)
Audio L2CAP
• UAP (Upper Address Part, containing
8 bits) HCI (Host Controller Interface)
• NAP (Non-significant Address Part,
Link Manager Protocol LMP
containing 16 bits) (link management)
Baseband
The LAP and UAP fields form the signifi-
SVB0008E
cant part of a Bluetooth device address. Radio (radio layer)

two devices and also implements security In addition to L2CAP, there are also au-
and authentication functions. dio and control protocols that govern the
The host controller interface (HCI) is handling of audio data and control data.
the interface between the hardware of a Audio applications, for example, can
Bluetooth module and the host-side pro- make direct use of the baseband layer
tocols. The HCI is a hardware abstraction once the audio signals have been coded
and realizes various interfaces for con- accordingly.
trolling the Bluetooth hardware and
transferring data (e.g. via USB or UART). The service discovery protocol (SDP) is
another important host-side protocol.
Host stack It serves to identify and scan for services
The host stack uses the HCI to control with specific properties and to describe
the Bluetooth hardware and transfer services within range of a Bluetooth de-
data. The format of this data is deter- vice. Bluetooth devices are meant to be
mined by the upper protocol layers. The able to interact on an ad hoc basis with
logical link control and adaptation proto- other devices in different environments.
col (L2CAP) implements the abstraction It is therefore necessary to know which
of the hardware's properties, adapts the services are made available by which de-
upper layers of the protocol stack to the vices within range. All devices wishing to
capabilities of the baseband and hides provide services must use an SDP server;
transmission details such as the connec- for all other devices, an SDP client is suf-
tionless or connection-oriented trans- ficient. The SDP prepares the services
mission type. The L2CAP layer essen- available on a device in a service data-
tially has three main functions: base. The service information in the pos-
• It can receive packets with a length of session of the SDP server is stored in a
up to 64 kB from the upper layers and service record. The service record con-
decomposes them into smaller data sists of a list with service attributes that
packets (segments) for processing in describe the properties of the service
the lower layers if necessary. At the more precisely and is identified by a
other end, the segments are recom- 32-bit service record handle.
posed back into packets.
• It manages the multiplexing and de- Two further protocols in the protocol
multiplexing of several packet sources. stack form the fundamental basis for the
If a packet is being recomposed, the interoperability of Bluetooth devices.
L2CAP layer determines to which pro- The RFCOMM (Radio Frequency Com-
tocol of the upper layers the packet is munication) cable emulation protocol
forwarded. above the L2CAP layer simulates up to
• L2CAP offers functions for negotiating 60 virtual serial interfaces derived in ac-
quality of service and configuration pa- cordance with the ETSI 07.10 standard.
rameters. This means, for example, that As a result, almost any software that pre-
the maximum size of the payload can be viously expected a serial interface can
negotiated so that a device with limited work with Bluetooth.
resources would not be overwhelmed TCS BIN (Telephony Control Protocol
by overly large packets. Using the con- Specification – Binary) is employed as
figuration parameter for the quality of the means of controlling telephone and
service, it is possible to define the telephony functions. This is a bit-ori-
properties of the data transfer: best ef- ented protocol for establishing voice
fort (best attempt but with no guarantee and data connections between Bluetooth
of data transfer) or guaranteed. devices. The use of appropriate, indus-
try-wide standards has made it possible, The Bluetooth profiles form the hierar-
here too, to ensure wide-ranging com- chy (Fig. 10). The GAP forms the basis
patibility with legacy applications. and describes all the essential functions
that a Bluetooth device must fulfill at the
Applications lowermost level. These include, for ex-
Many more protocols have been adapted ample, the functions for establishing and
for the Bluetooth standard (adapted pro- managing the link, the supported operat-
tocols) and can be found in the protocol ing modes and the security of a link.
stack. Internet applications can, for ex- The SDAP defines the access interface
ample, continue to use TCP/IP through for the service discovery protocol (SDP)
the point-to-point protocol (PPP) or the with which devices can discover or poll
Bluetooth network encapsulation proto- the services offered by other devices.
col (BNEP). For the exchange of vCalen- SDP builds on the GAP. All Bluetooth de-
dars and vCards, it is possible to use the vices must implement these two profiles.
object exchange protocol (OBEX) cov- The SPP is used by most other
ered by the IrDA standard. Bluetooth profiles. One exception to
this is the telephony control profile.
Profiles This is always used when Bluetooth is
The uppermost layer contains the appli- being used for cable emulation or if a
cations and profiles. Profiles represent serial data connection is to be used.
“standard solutions” for a particular us- The SPP builds on the GAP and uses
age scenario. The Bluetooth specification the RFCOMM protocol.
currently brings together 13 different The GOEP defines the fundamental
applications described as profiles. Each functions necessary for the exchange
profile uses a certain choice of protocol; of complex objects. It defines a client-
in principle, application profiles make server relationship for the exchange of
a different protocol stack available for data. Like the SPP, the GOEP provides
each application. Profiles describe the the foundation for further profiles.
vertical slice through the Bluetooth pro-
tocol stack; the protocols represent the
horizontal layers (Fig. 9). Within the pro-
files, the required and optional functions
of the layers are defined. These stan-
dardized profiles make it possible to en- 9 Profiles and protocols
sure interoperability between different

devices.
The four fundamental profiles of the
Applications
Bluetooth specification are:
• Generic access profile (GAP)
• Service discovery application profile
(SDAP)
Protocols
• Serial port profile (SPP) and

• Generic object exchange profile
(GOEP)
SVB0009E
Profiles
10 Profile hierarchy
Generic access profile (GAP)
Service discovery Telephony control profile

application (SDAP) Intercom (IntP) Cordless telephone (CTP)
Serial port profile (SPP)
Dial up network (DUNP) Generic object exchange profile (GOEP)
Fax (FaxP) File transfer (FP)
Headset (HSP) Object push (OPP)
SVB0010E
LAN access (LAP) Synchronisation (SP)
▶ Origin of the name, Bluetooth
The name, Bluetooth, originates from

King Harald Gormsen, a Danish Viking.
Harald had the nickname, Blatand.
In the Middle Ages, Bla actually meant
dark, while Blatand made reference
to the dark figure of Harald Gormsen.
Bluetooth, therefore, is not a direct
translation of Blatand.
In the 10th Century, Harald Gorm-

sen united large parts of Denmark
and Norway. The Bluetooth wireless
system developed 1,000 years later
unified a wide range of different infor-
mation, data processing and wireless
mobile devices. This is why it was
named after King Harald.
100 Bus systems | MOST bus
MOST bus The MOST Cooperation creates and ad-

ministers the specifications on which the
Introduction MOST bus is based. Furthermore, it de-
The MOST bus (Media Oriented Systems fines the requirements for implementation
Transport) was specifically developed for of MOST devices and provides appropriate
the networking of infotainment systems in compliance tests through accredited test
motor vehicles. In addition to traditional houses.
entertainment functions, such as radio
tuners or CD players, these infotainment Type of use
systems (which are becoming an increas- At present, the MOST bus is almost exclu-
ingly common feature of modern vehicles, sively used for the networking of infotain-
and premium class vehicles in particular) ment systems in motor vehicles. It can cur-
also offer video functions (DVD and TV), rently be found in over 35 vehicle models
route guidance capabilities and access to of various manufacturers, especially in
mobile communications and information. the premium and mid-range classes (as at
2006).
Requirements
The functions of an infotainment system Features of the MOST bus
place high demands on a bus system. The MOST bus supports the logical net-
The transmission of multimedia data, working of up to 64 devices, at which point
both audio and video, requires a high constraints associated with the chipsets
data rate and a synchronization of the may be encountered.
data transfer between source and sink In its current version, the MOST bus
as well as between sinks themselves. offers a data rate of 24.8 MBit/s (MOST 25).
For the transmission of informational Versions with higher data rates of 50 MBit/s
data, e.g. detailed information about the (MOST 50) and 150 MBit/s (MOST 150) are
music tracks on an mp3 player, and the already available as a development model.
sending of software updates, it is neces-
sary for the data transfer to be flexible Data transmission channels
with varying and, at certain times, For the transmission of data, the MOST bus
equally very fast data rates. supports the following channels and their
Likewise, the requirements for use in a different attributes:
motor vehicle, and the associated require- • The control channel is used for the sim-
ments of electromagnetic compatibility ple transmission of control commands,

(EMC), must also be fulfilled. for the signaling of device statuses and
for the sending of messages necessary
MOST Cooperation to system management. With MOST 25,
It was with these requirements in mind the control channel has a gross band-
that the MOST bus was developed by the width of 705.6 kBit/s.
MOST Cooperation. This was founded in • For the transmission of multimedia data,
1998 by BMW, DaimlerChrysler, Harman/ the MOST bus has a flexible number
Becker and Oasis SiliconSystems (now part of synchronous channels that are able
of SMSC). By 2006, the MOST Cooperation to carry both audio and video data.
had 16 automotive manufacturers and In MOST 25, there are a maximum of
67 suppliers and tool makers among its 15 audio channels available in stereo
membership. quality.
Bus systems | MOST bus 101
• On the asynchronous channel, data is Each device has a bypass. When this is
sent in packets. It is therefore suitable closed, the device forwards the signal
for transferring information that does directly and is therefore invisible to the
not require a fixed data rate but does MOST system. This service is useful,
need a high data rate at certain times. for example, when the system is starting
Examples could be the transfer of track and a device requires more time for initial-
information of an mp3 player or a soft- ization, or in the event of a temperature-
ware update. With MOST 25, the asyn- dependent shutoff.
chronous channel has a gross bandwidth There are also other possible structures
of up to 12.7 MBit/s. for a MOST system, e.g. a star topology.
Internally, however, these must always
The available bandwidth can be flexibly be arranged to form a (logical) ring.
distributed between the asynchronous With a star structure, this can be achieved
channel and the synchronous channels through use of a central hub, for example.
by means of the “boundary descriptor”,
which can also be shifted to the transit Device model
period if relevant preconditions are met. The MOST standard defines the model
shown in Figure 2 for a MOST device for
Topology which the following elements are required:
A MOST system is arranged in a ring struc- At the lowermost level, the physical
ture whereby a device is connected to its layer provides access to the transmission
predecessor or successor in the ring by agent. Departing from the original defini-
an input or output respectively (Fig. 1). tion that provided for optical transmission
One of the devices acts as the “timing by POF cable (Plastic Optical Fiber), there
master” and generates the data frames are today various physical layers available
for data transfer with which the other with optical and electrical transmission as
devices synchronize. well as various speeds of transfer.
1 Ring structure of the MOST system 2 Model for a MOST device
Timing master MOST device

position 0
NetBlock Appl.
Rx Tx
Network service
e
MOST network
Tx
Rx tion 1 ve
Tim sition 2
slav
po ng s
Tim
interface controller
si
i
Rx
Tx
ing
po
MOST physical
la
UVM0001Y
UVM0002Y
layer
The MOST network interface controller Transmission agent

(NIC) is a hardware controller that is re- The MOST bus was originally defined as
sponsible for controlling the physical layer an optical transmission agent that used a
and implements the basic transfer ser- plastic optical fiber (POF). The optical sig-
vices. While in the original version of the nal used here with a wavelength of 650 nm
NIC many control tasks had to be carried (in the red wavelength range) is generated
out by the main processor of the device, by an LED on the transmit side, which is
the current version (Intelligent Network designated Tx-FOT (FOT = Fiber Optic
Interface Controller, INIC) is, for the most Transceiver). On the receive side, Rx-FOT,
part, able to handle tasks autonomously the optical signal is converted back into an
and already implements substantial parts electrical signal by a PIN photodiode.
of the network service. Typical controllers The POF cables used for automotive
for the MOST are the OS8104 (NIC) and the applications (Fig. 3) consist of a 980 μm
OS81050 (INIC), manufactured by SMSC. thick optical core insulated by 20 μm thick
The MOST network service is the driver optical cladding with a low refractive in-
layer through which applications and dex. In total, the optical conductor there-
system services have access to the NIC. fore has a diameter of 1 mm. The optical
The lower layer of the network service fiber is insulated with a black buffer, which
(layer 1) offers basic communication and is in turn surrounded by a protective cable
management functions, while the layer sheathing. This gives the cable a total dia-
above (layer 2) offers functions for sup- meter of 2.3 mm.
porting the development of applications, POF lines offer the following benefits
such as a translator for the application when used in the motor vehicle:
protocol or support for the notification • No electromagnetic interference radia-
service. tion
The applications of the device are imple- • Insensitivity to interference irradiation
mented on top of the network service. • Lower weight than equivalent shielded
In a MOST system, the interface of an ap- electrical lines

plication is realized as a function block • More flexible routing in comparison to
(FBlock). Each device must at least imple- equivalent shielded electrical lines
ment a special FBlock, the NetBlock,
which is required for management func-
tions within the MOST system. In addition,
a device will usually implement one or 3 POF cable
more applications that can be used as

FBlocks of applications on other devices.
A device of a vehicle infotainment system
is often integrated with several functions,
Cable sheathing
e.g. those of a radio tuner and those of an
amplifier, which can be represented in the Buffer
MOST system as independent FBlocks.
2.3 mm
1 mm 1.5 mm Optical fiber

UVM0003Y
To date, modern vehicles are equipped • An area for the control channel (with
exclusively with MOST 25, with POF data MOST 25, two bytes per data frame)
transmission via POF connections. In addi- • An area for the synchronous channels,
tion to this, however, further transmission whereby one or more bytes are always
techniques have been defined for the assigned to a specific synchronous
MOST: channel
• The optical transmission by glass fiber • An area for the asynchronous channel
cable (PCS) with laser diodes (VCSEL),
which offers a greater damping reserve, In the case of MOST 25, the synchronous
supports higher speeds and is less sensi- and asynchronous area demand a 60-byte
tive to high temperatures. share of the data frame. The distribution
• An electrical transmission by copper between the synchronous channels and
cable, which is also less sensitive to the asynchronous channel is determined
temperature and is more economical by by the value of the boundary descriptor
comparison, requires additional shield- with a resolution of 4 bytes. The synchro-
ing measures at higher bandwidths. nous area must have at least 24 bytes
This has consequences for costs and (6 stereo channels). This means that
cabling. 24 to 60 bytes are permitted for the
synchronous area and 0 to 36 bytes are
Data transfer permitted for the asynchronous area.
Data transfer on the MOST bus is orga-
nized into data frames, which are gener-
ated by the timing master with a fixed data
rate and passed on by subsequent devices
in the ring.
Data frames
The timing master normally generates
data frames with a cycle of 44.1 kHz,
and in rarer cases 48 kHz. The cycle is
defined by the system manufacturer
(i.e. the vehicle manufacturer in conjunc-
tion with its device suppliers) to suit the
predominant media formats in the system. 4 Data frame
The size of the data frames, by extension,

determines the bus speed of a MOST sys- Administrative area: 1 byte
(preambel, boundary descriptor)
tem. With MOST 25, the size of a data
frame is 512 bits.
Administrative area: 1 byte
A data frame is made up of the following (frame control, status and parity)
ranges (Fig. 4):
• Areas for administrative information
at the start and end of the data frame,
the tasks of which include the synchro- Syn- Asyn- Control
chronous chronous channel:
nization of the data transfer, determina- channels channel 2 bytes
tion of the devices' ring positions and
Total of 60 bytes: distribution
the transmission of the boundary by boundary descriptor
UVM0004E
descriptor
1 data frame for MOST 25: 64 bytes
System events • Group addresses: each device may also

The fundamental system events described be assigned a group address in the
below indicate the current status of the 0x300 to 0x3FF range, making it possi-
data transfer in the MOST system. These ble to address all devices in a given
are monitored by the NIC and forwarded group at the same time.
by the network service to the applications, • Broadcast address: the MOST addition-
which must then respond accordingly. ally supports a special broadcast ad-
• Lock, stable lock: a lock occurs as soon dress, hexadecimal 0x3C8, which makes
as communication is established in the it possible to address all devices in the
ring, i.e. the correct device receives the MOST system.
correct data frames. A stable lock is set
when the system has been in lock status Transmission of control commands
for a specific period. The control channel permits the transmis-
• Unlock, critical unlock: an unlock de- sion of control messages with a maximum
notes the loss of the lock, i.e. that no length of 32 bytes. These have to be shared
correct data frames are being received. among consecutive data frames. With
If the system is in the unlock state for MOST 25, the 16 data frames required
a certain period, the status switches to here are also known as a block. The format
critical unlock. of the control messages is defined by the
• Network change event: if the number of protocol described in the next section.
devices in a MOST system changes be- The sending of messages on the control
cause one device has opened or closed channel is supported by the NIC. It pro-
its bypass, this is signaled by a network vides the necessary services for data pro-
change event. tection and channel arbitration, i.e. for de-
termining the next opportunity for a mes-
Addressing sage to be sent. The reliable transmission
Devices on the MOST bus are addressed by of control messages is safeguarded by the
means of a 16-bit address. The following use of a cyclic redundancy code (CRC) and
different addressing types are supported: an automatic repeat request.
• Logical addresses: each device has a log- The control channel's gross bandwidth
ical address that is unique in the system of 705.6 kBit/s is reduced to a net data rate
and is set and distributed when the sys- of 406 kBit/s following the deduction of ap-
tem starts. The logical addresses can proximately 3 % for system administration
be set dynamically based on the ring overheads (two of 64 control messages)
position, for which, beginning with the and less the extra overhead to account
timing master, the hexadecimal range for arbitration and data protection. Since
0x100 to 0x13F is reserved. Or they arbitration only allows a particular device
can be defined statically by the system to populate every third available control
manufacturer in a range reserved spe- message, an individual device has a maxi-
cifically for this purpose. mum data rate of 135 kBit/s at its disposal
• Physical addresses: for system adminis- for sending.
tration services, e.g. determination of To enable the sending of messages of
logical addresses, each device also has a more than 32 bytes, the network service
physical address that is derived from its supports the segmentation and deseg-
current ring position. The physical ad- mentation of application messages up to
dresses are assigned the range 0x400 to 65,556 bytes in size on the control chan-
0x43F. nel. This service, also known as the appli-
cation message service (AMS), differs from to occupy a synchronous channel

the control message service (CMS), which sends an appropriate system message
only allows 32-byte control messages to request (ResourceAllocate) to the tim-
be sent. ing master. A corresponding message
(ResourceDeallocate) is used to free the
Transmission of multimedia data channel.
In the MOST system, the transmission of
multimedia data takes place on the syn- Transmission of packet data
chronous channels. The data transfer is On the asynchronous channel, data is sent
controlled by relevant control commands in the form of packets. The asynchronous
on the control channel. A synchronous channel currently supports two modes:
channel can be assigned a specific band- a slower 48-byte mode, where 48 bytes
width, which is achieved with a resolution are available in each packet for the net
of one byte of a data frame. A stereo audio data transfer but which places less heavy
channel with a resolution of 16 bits re- demands on device implementation, and
quires four bytes, for example. In the a 1,014-byte mode that is more complex
case of MOST 25, a maximum of 60 bytes to implement.
(depending on the value of the boundary With MOST 25, between 0 and a maxi-
descriptor) are available for the synchro- mum of 36 bytes of a data frame can be
nous channels, which equates to 15 stereo assigned to the asynchronous channel,
audio channels. which corresponds to a maximum gross
The sinks and sources for multimedia bandwidth of 12.7 MBit/s. For the 1,014-
data are each assigned to an FBlock, which byte mode, this produces a maximum net
provides the necessary functions for their bandwidth of almost 11 MBit/s, and a data
management. An FBlock may therefore rate of almost 3 MBit/s for the 48-byte
contain several sources and sinks, which mode. In practice, however, substantially
are numbered consecutively with a source lower data rates are achieved mainly due
and sink number respectively. to restrictions in device implementations.
An FBlock has functions that supply in- The transmission of data packets on the
formation about the quantity and type of asynchronous channel is also directly sup-
sources and sinks that it makes available ported by the NIC. Likewise, a packet on
(SyncDataInfo, SourceInfo and SinkInfo). the asynchronous channel contains a CRC
Furthermore, each FBlock with a source to permit detection of simple transmission
has an “Allocate” function with which it errors. Unlike the control channel, how-
requests a synchronous channel and asso- ever, there is no issuing of automatic re-
ciates its source to it. Its “Deallocate” peat requests on the data link layer.
function, on the other hand, is used to To ensure reliable transfer and flow con-
free up the channel again. trol in the large data volumes typical of the
Similarly, an FBlock with a sink has a asynchronous channel, it is conventional
“Connect” function for connecting the sink to use yet another transport protocol that
to a particular synchronous channel, and is implemented in a higher driver layer.
a “DisConnect” function for its disconnec- This may either be the MOST high protocol
tion. Several sinks may also be connected (MHP), developed specifically for the
to a particular channel at the same time. MOST, or the popular TCP/IP protocol,
With MOST 25, the channel assignment which is set up on an appropriate adapta-
for synchronous channels is managed tion layer known as the MOST asynchro-
by the timing master. A source wishing nous medium access control (MAMAC).
Administrative functions not necessarily have to be the case. The

The MOST standard defines the following other devices in the MOST system in this
administrative services that are required relationship are known as network slaves.
for the operation of a MOST system. The network master manages an image
of the current configuration of a MOST
Configuration status system in the central registry. In addition
A valid configuration status in the MOST to the addresses of all devices, this also
system is the prerequisite for communica- contains all FBlocks implemented by them
tion in the application layer. The configu- (down to the system FBlocks such as the
ration of the MOST system is managed by NetBlock). A network slave searching for
a single device, the network master. An ap- the device address for a particular FBlock
plicative communication is only permitted may request it from the network master.
once this device has signaled a conflict- If a network slave frequently accesses
free configuration status. other devices, it may, for faster access, also
• OK configuration status: the network store a local copy of the central registry –
master transmits the OK configuration a decentral registry – which it must update
status as soon as a conflict-free configu- accordingly in the event of any change in
ration is achieved. the configuration.
• NotOK configuration status: if a conflict To build the central registry, the net-
arises, e.g. because two devices have work master carries out a network scan
the same address, the network master when the system starts and as soon as
transmits its NotOK configuration status. a network change event has occurred.
In response to this, all devices initialize During this scan, it queries the NetBlocks
their addresses and communication set- of all devices for the FBlocks implemented
tings. by this device. If it detects a conflict while
• New and Invalid configuration statuses: doing so, e.g. duplicate device address or
these configuration statuses indicate duplicate instance of an FBlock, it takes
that a new application (FBlock) has measures to resolve the conflict. To do
been registered in the system or that a this, the network master may initiate a
previously registered application has recalculation of addresses, convert the
dropped out. InstId of an FBlock or, if necessary,
ignore a device.
NetBlock
It is mandatory for each device to imple- Connection master
ment an FBlock called the NetBlock, which The connection master manages the syn-
covers a range of different administrative chronous connections existent in the MOST
functions (e.g. for address initialization) system at a given time. It makes available
and supplies information about the device a table containing information about all
and the FBlocks that it implements. the connections currently present in the
MOST system (SyncConnectionTable).
Network master It also makes it possible to query the band-
The network master is implemented by width remaining for further synchronous
a special device in a MOST system and connections (AvailableChannels).
is responsible for system configuration. On request, the connection master
In present-day systems, the network mas- builds connections between a specific
ter is usually realized by the headunit source and sink (BuildSyncConnection) or
(i.e. the control element) of the infotain- removes them (RemoveSyncConnection)
ment system. This device often assumes by sending the necessary commands to
the role of timing master too, but this does the associated FBlocks.
It is even possible to dispense with the

1 Elements of a control message
connection master completely if a central
Field Size Description
entity in a MOST system, e.g. the headunit,
DeviceID 16 bits Device address
is responsible for the administration of the
FBlockID 8 bits FBlock identifier
synchronous channels.
InstID 8 bits Instance of the FBlock
FktID 12 bits Function identifier
MOST application layer OPType 4 bits Type of operation
The MOST standard defines a suitable pro- Length 16 bits Length of the data
tocol in the application layer for the trans- field
mission of control commands, status infor- Data 0 to Data field
mation and events. This protocol makes it 65,535 bytes Table 1
possible to address a specific function of
an application interface (i.e. of an FBlock) Function block (FBlock)
that is provided by any device within the A function block (FBlock) defines the in-
MOST system. For example, it is possible terface of a specific application or system
to start playback of a CD on a separate service. An FBlock is assigned an address
CD player or to query the number of the comprising an 8-bit FBlockID, which spec-
track currently playing in it. While the ifies the type of FBlock, and an additional
MOST protocol is mainly applied to the 8-bit InstID. The latter is a means of distin-
control channel, it can be transferred guishing between several instances of
across the asynchronous channel if FBlocks of the same type in a MOST sys-
necessary. tem.
An FBlock that is controlled by an asso-
Overview ciated application, and the device that im-
The protocol for MOST control messages plements the FBlock, are known as the
provides for the following elements of a slave in this relationship (e.g. an applica-
control message (Table 1): tion that addresses an external CD player
• The address of a device in the MOST sys- through the associated FBlock). The appli-
tem (DeviceID) cation that controls the FBlock is called
• An identifier for an FBlock implemented the controller.
by this device (FBlockID) and its in- The functions that an FBlock possesses
stance in the MOST system (InstID) are defined by the function catalog. In the
• The identifier for the requested function case of system services, the function cata-
within the FBlock (FktID) logs are defined by the MOST standard.
• The type of operation (OpType) that The MOST standard even defines the inter-
should be applied to this function, e.g. faces for common applications from the
the setting or querying of a property of area of vehicle infotainment systems (e.g.
the FBlock for an amplifier or a CD player/changer).
• A data range containing the parameters However, these interfaces tend to be ex-
of the function call (Data) and a corre- tended by the system manufacturer. Other
sponding length value (Length) proprietary FBlocks are defined in their
entirety by the system or device manufac-
turer.
Functions and operations Data field

A function block is made up of several The data field of the control message con-
functions that are addressed with a 12-bit tains the parameter values for a function
FktID. Depending on the type of function, call or its results. The data field is inter-
the function may be assigned differently preted as specified in the definition of the
predefined operations, which are de- function concerned and may contain one
scribed by a 4-bit OpType identifier. In the or more parameter values.
case of functions, a further distinction is
made between methods and properties. The MOST standard defines the following
A method describes an action that an parameter types:
FBlock is able to execute, e.g. starting the • Boolean and BitField for individual
station scan of a radio tuner. A method boolean values (1 byte) or sequence

can be initiated by the controller using a of individual bits (1, 2 and 4 bytes).
Start or StartResult operation. In response, • Enum for enumerations.
the slave returns a corresponding result • Unsigned and signed byte, word and
message if applicable (for StartResult). long for integer values with or without
If this result is not made available within sign and a size of 1, 2 or 4 bytes.
a specific period, the controller is notified • String for character strings. These
that the method is still being processed by are zero terminated and contain a
means of a processing message. In the end, description of the character coding
a method may even be terminated by the in the first byte.
sending of an abort message. • Stream, ClassifiedStream and
A property describes a particular attri- ShortStream for various types of

bute of the FBlock, such as the number of byte sequences of any length.
the track currently playing. A property
such as this can be requested (Get opera- A parameter type for real numbers does
tions) or set (Set or SetGet operations) not exist. The MOST standard provides for
by the controller. As the result of such a real numbers represented as fixed-point
query, the slave responds with a status values by the integer parameter types.
message, if applicable (for Get and SetGet), More complex data types are not defined
containing the current contents of the in the standard either. These have to be
property. realized by function classes instead.
Furthermore, the MOST standard pro-
vides for a notification service with which
a controller may register with a slave for
certain properties. If the value of one of
these properties changes, all controllers
that have registered for notification are in-
formed of the change by a status message.
Function classes Applications

To standardize the way in which functions As well as defining the lower layers neces-
are defined, the MOST standard specifies sary for data transfer, the MOST standard
a series of function classes for proper- defines the interfaces for typical applica-
ties. These determine which properties tions from the area of vehicle infotainment
the function has and which operations are systems, e.g. a CD changer, amplifier or
permissible. radio tuner.
In addition to simple function classes The FBlocks defined by the MOST
(with a single parameter for numerical Cooperation are summarized in a func-
values or texts), there also exist complex, tion catalog and listed with their current
composite function classes. These are versions in the next section. The function
used to represent composite data struc- catalog defines in detail all functions for
tures such as records and arrays. They can the FBlocks contained in them as well
also be nested at a specific nesting depth, as the permissible operations and their
which makes it possible to create two- parameters.
dimensional arrays or arrays of records, For the function catalog, there is a
for example. These nested data structures machine-readable description available in
are, for example, used to represent the XML format that can be used to import the
phone book of a mobile phone. catalog into various MOST tools.
With multidimensional function classes, In addition to the function catalog’s
the two parameters PosX and PosY are description of all functions of an FBlock,
used to address a particular position in the developments are underway to specify
first or second dimension, where 0 always the dynamic behavior for the use of an
stands for all elements of this dimension. FBlock. This involves the use of message
This makes it possible to access not only a sequence charts (MSCs) compliant with the
specific element but also an entire line of a MSC 2000 standard but with minor, MOST-
twodimensional array. specific enhancements.
Functions required by various applica-
tions, e.g. the aforementioned functions
for managing sources for multimedia data,
are defined in the GeneralFBlock. This is
used as a function compilation, i.e. when a
new FBlock is defined, all necessary func-
tions are copied from the GeneralFBlock to
the new FBlock.
Standardization The following application interfaces are

The MOST standard is maintained by the standardized by the MOST Cooperation.
MOST Cooperation, which also publishes The following versions are currently valid:
the corresponding specifications. The • AudioAmplifier (FBlockID: 0x22),
specifications are available through the Version 2.4.2: a simple amplifier.

home page of the MOST Cooperation • AuxIn (FBlockID: 0x24), Version 2.4: an
(www.mostcooperation.com). interface for connecting MP3 players or

The following versions of the MOST data carriers with pieces of music.
specifications are currently valid (as at • MicrophoneInput (FBlockID: 0x26),
2006): Version 2.3.1: a microphone input.

• MOST Specification, Version 2.4. • AudioTapePlayer (FBlockID: 0x30),
• MOST Dynamic Specification, Version 2.3.1: a cassette player.
Version 1.2. • AudioDiskPlayer (FBlockID: 0x31),
• MOST Specification of Optical Physical Version 2.4: a CD player or CD changer.
Layer, Version 1.1. • DVDVideoPlayer (FBlockID: 0x34),
• MOST Specification of Advanced Optical Version 2.4.1: a DVD player or DVD
Physical Layer, Version 1.0. changer.
• MOST Specification of Electrical • AmFmTuner (FBlockID: 0x40), Version
Physical Layer, Version 1V1. 2.4.2: a radio receiver for FM/AM.

• MOST High Protocol Specification, • TMCTuner (FBlockID: 0x41), Version
Version 2.2. 2.3.1: a special receiver for traffic-

• MAMAC Specification, Version 1.1. message signals (TMC).
• TVTuner (FBlockID: 0x42), Version
The MOST Cooperation defines within 2.3.2: a television receiver.

the framework of the compliance process • DABTuner (FBlockID: 0x42), Version
requirements placed on MOST devices 4.0: a receiver for digital radio (DAB).
which must be satisfied by all devices in • SDARS (FBlockID: 0x44), Version 2.4: a
order to be able to carry the MOST logo. receiver for satellite radio.
MOST compliance is tested and awarded • Telephone (FBlockID: 0x50), Version
by test institutes which have received ap- 2.3.2: a telephone module or a connec-
propriate accreditation from the MOST tion to a cellular phone.
Cooperation. The compliance require- • GeneralPhonebook (FBlockID: 0x51),
ments are governed by the following speci- Version 2.3.1: access to a phonebook,
fications, which describe the compliance e.g. that of a cellular phone.
process itself and the compliance tests • NavigationSystem (FBlockID: 0x52),
for the physical layer, for system mecha- Version 1.11: an interface for a navi-
nisms (Core) and for application interfaces gation system; but only unofficially
(Profile): approved by the MOST Cooperation.
• MOST Compliance Requirements, • GraphicDisplay (FBlockID: 0x60),
Version 2.0. Version 2.3: an independent display

• MOST Compliance Test of Physical Layer screen.
Specification, Version 1.0.
• MOST Core Compliance Test Specifica-
tion, Version 1.1.1.

• MOST Profile Compliance Test Specifica-
tion, Version 1.0.

Bus systems | TTP/C 111
TTP/C Areas of application

TTP/C was originally developed for use
Overview in safety-relevant systems in the automo-
The Time-Triggered Protocol for SAE tive field. Because the core functions of
Class C networks (TTP/C) is a time-con- TTP/C were formally verified and a TTP/C
trolled protocol in which the network us- network can be put together economically
ers transfer data in special time intervals compared with other protocols with simi-
determined in advance. In other words, all lar reliability properties, TTP/C has also in
network users have a global time defini- the meantime come to be used in aircraft,
tion which is determined by means of a a field which is subject to very stringent
time-synchronization protocol. requirements with regard to safeguarding
TTP/C was developed specially to satisfy functionality. TTP/C is also used in train
real-time requirements in distributed, control systems.
fault-tolerant systems. It was developed
specifically with a view to its use in the Fault-tolerance strategy
automotive field, but has in the meantime The fault-tolerance strategy is used to
also come to be used in other fields, such describe which faults can be detected
aviation and railroads. and handled. TTP/C is able to detect and
According to the requirements of the handle each individually occurring fault
Society of Automotive Engineers (SAE) for (Single Fault Hypothesis). In addition, a se-
Class C protocols, TTP/C is a protocol for ries of multiple fault scenarios is handled.
high-speed networks (>125 kbit/s bit rate) However, some important multiple fault
which can be used for real-time commu- scenarios, such as a complete temporary
nication in check systems. For this reason, breakdown of communication for example,
particular emphasis was placed during cannot be detected on the node level. Ad-
development on the fields of fault detec- ditional methods of fault detection and
tion and tolerance, robustness against mal- handling are required on the application
functions, composability, and guaranteed level for this purpose.
latencies. The TTP/C network must consist of at
TTP/C is based on the principle of least four real member nodes to ensure
Time-Triggered Architecture (TTA), a that all single faults are detected.
framework for developing highly reliable, The network ensures on the architecture
distributed real-time systems and the ap- level that a faulty network node cannot
plications that run on them. prevent any properly functioning network
The development of TTP/C dates back node from transmitting its message. This is
to the MARS project (Maintainable Real- done by bus guardians, which permit bus
Time System), which was conducted in access only at defined points in time.
1979 at the Technical University of Berlin.
Further development followed within the Time-Triggered Architecture
framework of further projects, such as Time-Triggered Architecture (TTA) forms
TTA (Time-Triggered Architecture) or the the basis on which the TTP/C protocol was
Brite-EuRam III project “Safety Related specified (Fig. 1). The core element of TTA
Fault Tolerant Systems in Vehicles (X-by- is the network node, which comprises a
Wire)”, in which systems for activating host processor and a TTP/C controller,
safety-relevant automotive components, referred to in the following as the control-
such as brakes or steering systems, via ler. The network node is the “Smallest Re-
electronic networks without a mechanical placeable Unit” (SRU) in a TTP/C system.
fallback level were examined. The host processor processes inputs from
112 Bus systems | TTP/C
sensors and activates actuators. The Com- Communication Network Interface (CNI)
munication Network Interface (CNI) forms The CNI functions as a temporary firewall
the interface between the host processor because it makes available a memory area
and the controller and provides memory via which data received and to be sent are
areas via which the host processor and the exchanged between the host processor
controller can exchange information. and the controller. This interface repre-
The controller is in the end connected to sents the host processor’s sole possibility
the TTP/C bus, by way of which it connects of transferring information via the net-
the node with other nodes. The group of work. It is therefore not possible for the
all the nodes connected to a TTP/C bus and host processor to influence the communi-
the bus forms a cluster. cation sequence. In particular, the point in
time at which information is provided by
Host Processor the host processor has no influence on the
The host processor executes the applica- network’s communication sequence.
tion, the actual function of the network
node. In order to support the application, TTP/C Controller
an operating system, which also for the The controller (Fig. 2) operates entirely
most part provides the FT-COM layer, runs independently of the host processor. Its
on the host processor as well as the ap- most important components are the proto-
plication. The FT-COM layer introduces a col processor, a local bus guardian, and the
further abstraction level of the technical Message Descriptor List (MEDL).
realization of communication. The applica- The function of the protocol proces-
tion can provide data for transfer via the sor is to prepare information provided
interfaces of the FT-COM layer, thanks to by the host processor in the CNI in such a
which the application designer does not way that they can be transferred as TTP/C
have to deal with deeper concepts, such as frames.
the memory areas in the CNI. The controller is connected via two in-
terfaces with the remaining components
of the network node. The connections to
the host processor and to the transceivers
(drivers) are established via the CNI and
the Logical Line Interface respectively.
1 TTA network
Node Node Node Node

Sensors,
actuators Host Host Host Host
processor processor processor processor
Application
level
CNI Data CNI CNI CNI
communication
TTP/C
communication Controller Controller Controller Controller
subsystem
Channel 0
SVT0001E
TTP/C bus
Channel 1
Message Descriptor List (MEDL) Strictly speaking, the designation Message

The MEDL contains all the check informa- Descriptor List is incorrect. Information
tion required by the controller to send and blocks within a frame are designated as
receive data. It is a list adapted for each messages in TTP/C (Fig. 3). However, this
controller containing, for example, infor- structure exists only from the perspec-
mation on the position and size of transmit tive of the host processor, not from the
and receive slots. perspective of the controller, which only
The information stored in the MEDL processes frames. The designations has
includes the points in time at which the persisted, however, because it was intro-
controller frame may be transmitted and duced in earlier specifications.
at which frames of other controllers are
expected which are processed by the host Local and Central Bus Guardians
processor. The primary function of the bus guardian
CNI memory positions in which informa- is to limit transmissions of a node to the
tion from the host processor is to be pro- defined time windows. For this purpose,
vided for transfer and in which received it accesses the information of the MEDL
information from the controller is to be in which the transmit slots of the node are
stored are also defined in the MEDL. stored.
The MEDL also contains further in- There are two different types of bus
formation needed for the operation. For guardians. Local bus guardians are con-
example, the identification of frames used nected between the controller and the
to synchronize the internal clocks and the TTP/C bus and prevent bus access out-
synchronization times. side defined transmit slots. Central bus
guardians operate in a central coupler.
If exclusively central bus guardians are
2 TTP/C node used, however, only faults are excluded for
the nodes which are directly connected to
the couplers. Connected bus topologies,
on the other hand, cannot be completely
I/O Interface
protected.
Hostprozessor
3 Message components
Communication Network
Interface (CNI) Application data:
Supplied/processed by the host processor
Protokoll- Messages:
MEDL Data areas interpreted
prozessor
by the host processor
Bus Guardian Controller Message Message Message CRC

TTP
1 2 3
Logical Line
Interface Protocol overhead Protocol overhead
Frame type, explicit C-state Frame CRC
Driver Driver
Frame:
SVT0002E
SVT0003E
Bit stream transmitted via the channel

TTP/C Bus
In addition to monitoring bus access, the Network

central bus guardian can also assume the Network Size
function of detecting and containing SOS TTP/C is designed for networks of up to
faults (Slightly-Off Specification). These 64 nodes. A TTP/C network can consist of
are faults in which a node continually vio- a minimum of two nodes, but at least four
lates slightly one or more communication nodes are required to satisfy the Single
parameters, such as the voltages used or Fault Hypothesis. A problem in the data
the timing of the transmit slots. The upshot transfer lies in the danger that a transfer is
of this is that some network nodes which not received uniformly by all the network
were designed with greater tolerances nodes. If there are at least four real mem-
process these messages, but other reject ber nodes (nodes with sole access to one
them as faulty. As a result of SOS faults, it transmit slot) in a TTP/C network, TTP/C
becomes problematic to establish whether can identify reliably whether the message
data have been correctly transmitted or transmission was successful or whether
incorrectly received. faults occurred during transmission or
reception.
Logical Line Interface
The Logical Line Interface (LLI) is the Topology
interface between the controller and the All TTP/C networks have two independent
transceiver (driver). The LLI establishes physical channels via which a network
the structure of the logical information in node simultaneously transmits information
which the driver must accept information redundantly. This configuration is required
to be transmitted and in which it must pro- in order that the Single Fault Hypothesis,
vide received information. This makes it i.e. the handling of the failure of any com-
easier to adapt controllers to a new trans- ponent, can be satisfied. However, it is also
mission medium since only the drivers possible for nodes which do not fulfill any
have to be modified; however, the format safety-relevant functions to be connected
of the interface to the rest of the system only to one of the two channels.
remains unchanged. Since a controller always forwards
messages via the LLI to its drivers, it is
Driver possible to adapt TTP/C to very different
The driver is the transceiver for the trans- topologies. Bus, star and hybrid topologies
mission medium. It converts the physical from any combinations of bus and star to-
states of the TTP/C bus into logic states, pologies are described (Fig. 4) in the TTP/C
which can be processed by the controller. Specification (Version 1.1).
Bus topology Star topology

In a bus, all the network nodes are con- If a star topology is used, the network
nected in series to central cables. If a bus nodes are connected to each other via cen-
topology is used, the structure dictates tral couplers. Here, each TTP/C channel is
that no central bus guardians can be used. brought together in its own coupler.
However, the overall length of the cabling Star topologies facilitate the use of cen-
can be reduced to a minimum for this pur- tral bus guardians, which are able to pro-
pose. vide additional protection against faults.
4 Network topologies Hybrid topologies of bus and star

Even hybrid technologies composed of bus
and star topologies can be built. For ex-
a ample, it is possible to cascade several star
Node Node Node Node topologies into a multi-star or to connect a
sub-bus to a star topology.
Transmission Media
Within a TTP/C network, communication
takes place via two channels. The channels
are designated Channel 0 (Ch0) and Chan-
b
Node Node nel 1 (Ch1), or Channel A (ChA) and Chan-
SC nel B (ChB).
Because the TTP/C Specification is
SC formulated very openly with regard to
Node Node the transmission medium, it is possible
to build TTP/C networks using very dif-
ferent media. Both electrical and optical
media can be used here. In particular, it is
c also possible to design one channel to be
Node Node optical and the other to be electrical. Dif-
ferences in the latencies of the different
SC SC
media can be compensated for by way of
Node Node
SC SC the configuration in the MEDL.
The achievable bit rate also varies to-
Node Node gether with this flexibility; however, the bit
rate itself is fixed during operation within
a network and cannot be altered.
It is possible with current implemen-
d tations to achieve bit rates of 5 Mbit/s
Node Node Node for asynchronous data transfer during
system starting and of up to 25 Mbit/s for
SC synchronous data transfer during normal
operation. Prototypes in the laboratory Fig. 4
SC a Bus
environment has achieved bit rates of up
b Star
to 1 Gbit/s.
Node Node Node c Multi-star
d Star/bus
SVT0004Y
combination
SC Coupler
Communication cycle FTUs are used so that safety-relevant

TTP/C controls bus access by means of a information can also be provided in the
TDMA process (Time Division Multiple event of the failure or one or more net-
Access, Fig. 5). Here, the available band- work nodes. Here, all the nodes combined
width is divided into time windows (slots) in a FTU are able to process a task and use
in which in each case an established net- in a cycle alternately the transmit slot of a
work node can utilize the network’s full round allocated to the FTU.
bandwidth.
TTP/C distinguishes between rounds Passive Nodes
and cycles in the sequence of slots. All A node transfers data only when current
rounds contain the same sequence of data are available for transfer, otherwise
slots whose size can be independently it remains silent. A node which does not
established. A cycle is a defined sequence transfer any data in its transmit slot is
of rounds in which the individual rounds a passive node. This is the standard be-
differ only in the content of the individual havior when the controller of the node
slots, but not in the sequence of size of the establishes that its host has not updated
slots. the “Host Life Sign”, by means of which
In round, each member node of the an application signals that it is active. The
cluster transmits one frame precisely in controller then assumes that the applica-
one slot. A distinction is made here be- tion cannot generate any valid data and
tween two types of member node: real does not transmit.
and virtual. Real member nodes are nodes This mechanism is also used for monitor
which have their own transmit slot which nodes, which merely monitor the commu-
only this node can access. Virtual member nication of the other nodes, but themselves
nodes consist of a group of nodes which do not influence the communication.
use a common transmit slot. This structure
is used to build up redundant structures
(Fault Tolerant Unit, FTU) which make the Time control
failure of nodes more tolerable. The time window of a slot can be subdi-
vided using the performed activities into
different phases (Fig. 6). Before a node
can transmit information, it needs time in
order to read in the transmission param-
5 Schematic bus-access diagram eters or to prepare the transceivers. This
phase is called the Pre-Send Phase (PSP).
This is followed by the time interval in
A Node C Frame Ch0 Frame
which information is actually transmitted
B Round slot D Frame Ch1
via the network, the Transmission Phase
(TP). After a message has been received,
A 0 1 2 3 0 1 2 3 0
a node needs time in order to prepare the
information. This takes pace in the Post-
B 0 1 2 3 4 5 6 7 0
Receive Phase (PRP). Because the time
C
needed for PSP and PRP is not fixed, the
D
PRP is followed by the Idle Phase, a phase
Node slot Time
through which the length of a node slots
TDMA round TDMA round is stretched to the time interval defined in
SVT0005E
the MEDL.
Cluster cycle
A TDMA slot is fixed at points in time COM layer, also an application, which is
which can be observed using the transmis- processed on the host processor.
sion. But because this is not applicable to The lower layers are implemented in
all the activities of a controller during a the controller, where data communication
slot, a distinction is made between node takes place between the host processor
slot and TDMA slot. and the controller via the CNI interface.
For a node, the slot begins with the PSP. In the controller, the top layer is the Pro-
But because no signal is transmitted in this tocol Service Layer; under this is the Data
phase, external observers lack a precise Link Layer followed at the lowest level by
point in time at which the node slot begins. the Physical Layer.
The first precise point in time which can
be observed from an external source is the Physical Layer
start of the TP. This point is also called the The Physical Layer comprises the require-
Action Time (AT). A TDMA slot therefore ments which are placed on the physical
comprises the time interval which passes transmission of information. Often the
between two ATs. Therefore, there exists connections to be used, the cable types or
in a TDMA slot only the TP and the Inter- the voltage levels are established for this
frame Gap (IFG), in which the PRP, Idle purpose in a protocol specification.
Phase and PSP are combined. Efforts are indeed underway to formu-
A node slot, on the other hand, begins late a specification for a TTP/C physical
with the PSP, which is followed by the TP, layer, but the TTP/C protocol specification
the PRP and the Idle Phase. merely establishes requirements which
must be fulfilled by the physical layer
TTP/C Protocol used. This has the advantage that the pro-
TTP/C systems use the TTA framework tocol is very flexible with regard to the
(Time-Triggered Architecture) as the basis networking and bit rate. It is thus possible,
for which – with the exception of the appli- for example, for networking to use both
cation – finished solutions can be used. optical and electrical media, which can
The TTA protocol stack (Fig. 7) can be also be combined with each other.
divided hierarchically into three areas. However, the system designer is faced
The top layer is the host layer, the ex- with the problem of initially having to
ecuted application. Under this is the FT- identify and evaluate possible solutions.
6 Slot timing 7 Protocol stack
Host
Executed application
Layer
Node Dslot Node
slot i-1 Node slot i slot i+1 FT-COM CNI
Redundancy management FT-COM
DTP DPRP DIdle DPSP Message permanence Layer
TTP CNI
PRP Idle PSP TP PRP Idle PSP TP Communication services Protocol
Safety services Service
Action Time
Action Time
IFG Higher-layer services Layer

(AT)
(AT)
Frame structure Data Link

Endianess of bus Layer
TDMA slot i-1 TDMA slot i TDMA
slot i+1 Bit synchronization Physical
Bit coding/decoding Layer
SVT0007E
SVT0006Y
Layers processed by the host processor

Layers processed by the controller
Essentially, three requirement are placed FT-COM Layer

on the physical layer: The FT-COM Layer offers functions for
• Two independent physical transmission realizing replicated, distributed applica-
channels must exist. tions. These include redundancy control
• The transmission channel must permit and the decision as to whether replicated
distributed broadcast transmissions, i.e. messages of other nodes are correct.
transmissions which are received by all In order to allow applications standard-
the network users. ized access to communication data, the
• The delay which occurs during the FT-COM Layer offers an interface similar
transmission of information via a me- to CNI in which the data for the Host Layer
dium (Propagation Delay) must be applications are provided.
known.
Host Layer
Data Link Layer The Host Layer comprises the applications
The Data Link Layer contains the functions which run on the host processor. In addi-
which are needed to transmit frames be- tion to the actual applications, these in-
tween nodes. These include the structures clude the operating system and the check
which are required for access to the trans- structures (control loops).
mission channels and for the transmission In order to give the communication sys-
of frames. In particular, the Data Link tem as modular a structure as possible, the
Layer establishes the format of the mes- applications of this layer should access the
sage frames. FT-COM interface and not the CNI. This
makes it possible, for example, for an ap-
Protocol Service Layer plication which previously ran on a real
The Protocol Service Layer is responsible member node to be switched with minimal
for implementing the protocol services of- complexity to a series of virtual member
fered. Here, a distinction is made between nodes.
communication services, safety services
and higher-level services.
The functions of the communication ser-
vices are the safeguarding of reliable data
transfer, startup of a cluster, reintegration
of member nodes in the cluster, transmis-
sion acknowledgment, and fault-tolerant
clock synchronization.
The safety services include the manage-
ment of node membership, avoidance of
cliques, and the independent bus guardian.
The higher-level services include func-
tions for delayed mode changes, distrib-
uted alarm, external clock synchroniza-
tion, or reconfiguration of a node.
Protocol services The internal clocks of the network nodes 1)
Jennifer Lundelius-
Cluster Startup of a cluster can run with different levels of
Welch,
All the nodes of a cluster act without syn- precision. Only nodes which have precise
Nancy A. Lynch
chronization during cluster startup; bus clocks should be used for time synchro-
access is therefore not yet controlled via nization. These nodes receive the flag as 2)
TDMA. Because there is always a trans- “Master Clocks” in the MEDL, whereby Hermann Kopetz,
mission delay caused by the medium the SYF (Synchronization Frame Flag) Wilhelm Ochsenreiter:
used when messages are distributed in a is set for their frames. Nodes with more “Clock Synchroniza-
tion in Distributed
network, two unsynchronized nodes of a imprecise internal clocks act as “Slave
Real-Time Systems”,
cluster can start simultaneously with the Clocks” and use the clock information of
IEEE Transactions on
message transmission if the transmission the master clocks. There must be at least Computers, 36(8):
delay is greater than the duration of the four master clocks in a cluster to ensure 933-940, August 1987.
startframe. This can occur during system that clock synchronization is Byzantine-
startup if several nodes are authorized to fault-tolerant. Byzantine faults are faults in
initiate synchronization (coldstart nodes). which different nodes interpret the same
The situation can arise where different transmission differently.
cliques arise within one cluster of which The nodes of a cluster are synchronized
only the nodes of one clique are synchro- by measuring the Action Time (AT) in
nized, but not all the nodes of the cluster. each transmitted slot. If the SYF flag of a
In order to avoid this, all the nodes re- frame is set and the received frame was
ject the first correctly received coldstart faultlessly received, the AT of the frame is
frame. The coldstart nodes wait for a time stored in a stack four deep. In other words,
interval (startup timeout) before trying the last four correction values determined
again to transmit a coldstart frame. Be- are stored in the stack.
cause the size of the startup timeout must The actual clock synchronization takes
be different for all the coldstart nodes in place in frames, for which the ClkSyn
the cluster and differ at least by the size flag (Clock Synchronization) is set in the
of the transmission delay, collisions can MEDL. If this is the case, the mean value
no longer occur during the second cluster is formed from the stack values, where
startup. This process is also known as the highest and lowest values are ignored.
“Big Bang”. The internal clock is now corrected by the
calculated mean value.
Synchronization TTP/C also supports the inclusion of
In a communication system which controls external reference clocks (e.g. GPS) in the
bus access by means of TDMA, it is impor- system. Time Gateway Hosts are nodes of
tant for the individual users to have as pre- this type which offer an external time ref-
cise an idea as possible of a global time. erence. These nodes transmit the external
Synchronization in a TTP/C system uses correction value as application data in a
for the operation time only the offset cor- frame. The remaining nodes of the cluster
rection by means of which the onset of a store this value in the CNI and include it in
point in time is synchronized to all the in- the calculation of the AT correction values.
ternal clocks of the nodes of a cluster. For
this purpose, TTP/C uses a variant of the
Welch-Lynch algorithm 1) which was de-
scribed by Kopetz and Ochsenreiter 2).
Membership Service Transmission acknowledgment

TTP/C uses Membership Service to ensure TTP/C offers an implicit transmission ac-
that all the nodes of a cluster at each point knowledgment. Because the member vec-
in time have the same consistent view as tor of the next sender in the round is also
to which nodes are present in the network transmitted as part of the frame, the origi-
and which are not. For this purpose, each nal sender can check whether its transmis-
node carries a member vector in which it sion was correctly received.
notes whether the last received transmis- But because the successor may be faulty,
sion of a node has run correctly. If a node the sender also observes the transmission
receives a message incorrectly, it changes of the second successor. If the latter iden-
the entry of the message sender to “incor- tifies in its member vector that it has also
rect”. received the transmission incorrectly, this
By comparing the member vectors of means for the sender that a transmit fault
the following frames, a calibration is made has occurred at its end. If the second suc-
with the remaining nodes of the cluster. cessor has received the message correctly,
the first successor detects that a receive
Clique Avoidance fault has occurred.
In order to achieve correspondence within Measures can now be taken to handle
the cluster with regard to the current state both fault types.
of the cluster, it is necessary to prevent the
formation within the cluster of subgroups Controller State (C-State)
or cliques whose view of the current state In order for the applications executed on
differs from the real state of the cluster. the host processors to be able to deliver
To prevent the formation of cliques, consistent results, it is necessary that
each node logs how many transmissions all the nodes of a cluster share common
were successfully and how many were in- knowledge of the state of the cluster. This
correct. knowledge is summarized in the C-state.
Before it can transmit a frame itself, Each node may indeed determine its C-
the Clique Avoidance Algorithm is used. state independently, but the cluster agrees
The controller checks whether more suc- by comparison with the other nodes to the
cessful transmissions were received than view of the majority of the users.
incorrect transmissions. Only if this is the The controller state is influenced by the
case does the node set both counters to values for global time, current position in
zero and transmit its frame. If this is not the TDMA round (MEDL position), current
the case, the controller switches to the cluster mode, requested mode change
“Freeze” state, in which the controller is (DMC), and the member vector.
deactivated and no data are transmitted.
From this state, it can be placed by the
host processor in the initialization mode
again.
Cluster Mode the case, the controller assumes that there

In many real-time systems, operation has been a communication fault and begins
and check functions can be divided into with a cluster startup.
phases which are mutually exclusive. For
example, the operation of a motor vehicle Frame format
could be divided into the phases “Driv- Because TTP/C regulates bus access via
ing” and “Stationary”. In order to support TDMA, all messages are transmitted in
the different requirements placed on the broadcast mode and a priori knowledge
exchange of information in the different of the communication sequence and its
phases, it is possible to define in a TTP/C content is present in the MEDL, it is pos-
network modes in which the transmitted sible to keep very low the overhead of the
information in a frame can vary. However, protocol, i.e. the volume of information
the order of the slot in a TDMA round is needed to process the frame.
the same in all modes. Two different frame types are used in
For the purpose of correct interpreta- the operation of a cluster (Fig. 8). Cold-
tion, it is therefore necessary when modes start Frames, also known as Initialization
are used for all the nodes of a cluster to be Frames or I-Frames, occur exclusively
in the same mode. during system startup. Normal Frames,
also known as N-Frames, are used for
Deferred Pending Mode Change (DMC) communication during normal operation.
A change of cluster mode can be initi-
ated by applications if the controller is
permitted in the MEDL to change to the
requested new mode.
Because the change of cluster mode
comes into force only at the beginning of 8 Frame format
the following cycle, the information for

the deferred pending mode change is re-
Header
corded in the C-state.
I/N Mode Change
If another change of cluster mode is re- Frame Flags
1 bit 3 bits
quested by another application later in the
cluster cycle, this request overwrites the
previous request. A special case is CPM
(Clear Pending Mode change), in which an I-frame
application prevents the requested change Header C-state CRC
of cluster mode.
Reintegration of a node
A node must synchronize itself with the N-frame (explicit C-state)
cluster if it has failed due to an incorrect Application data
Header C-state CRC
transmission from the member vector or up to 240 bytes
if it has been re-initialized. To this end, it

waits for a frame in which an explicit
C-state was transmitted, e.g. an initializa- N-frame (implicit C-state)
tion frame. The controller adopts this Application data
C-state and uses it for its transmission. Header CRC
up to 240 bytes
SVT0008E
In order for this to be able to function,

a frame with an explicit C-state must be
transmitted every two rounds. If this is not
I-Frame C-state
An I-frame begins with the TTP/C header. The C-state consists of the global time,
This is followed by the current controller information on the MEDL and the member
C-state and finally the CRC value of the vector of the cluster. The MEDL position
frame. corresponding to the Round Slot Position
in the current cluster mode, the current
N-Frame cluster mode and the requested cluster
To ensure that Membership Service can mode change (DMC) is transmitted as in-
function, it is necessary for the C-state to formation on the MEDL.
be included in the frame structure. This An explicit C-state has a size of 96 bits
can occur explicitly or implicitly in the (six words of two bytes each).
case of N-frames.
If the C-State is explicitly transmitted, CRC checksum
the header is followed by the controller TTP/C uses the CRC checksum only to
C-state. This is followed by application identify transmission faults. Transmission
data and then the CRC value of the frame. faults are not corrected. Transmission
If the C-State is implicitly included in the faults are detected by a process whereby
frame, the frame consists of the header, each node determines a CRC checksum for
application data and finally the CRC value the received message and compares this
of the frame, the determination of which with the transmitted checksum.
was initialized with the current C-state. A polynomial is used to determine the
Download frames are a special form checksum. The TTP/C Specification in this
of the N-frame; these are accepted by all respect does not stipulate a specific poly-
the receiving nodes, regardless of cluster nomial; instead, it requires the polynomial
association and C-state. To facilitate this to facilitate a Hamming distance of at
behavior, a constant value which is known least 6. In other words, it must be possible
to all the nodes is used for these frames to within one transmission for at least five in-
determine the CRC. correctly transmitted bits to be detected.
The CRC calculation is initialized by the
Frame Fields Schedule ID, where the ID is split into two
Header parts. One part is used for initialization
The header of all the TTP/C frames is for Channel 0 and the other for Channel 1.
identical. Firstly, one bit signals the frame This prevents a node which does not have
type, whether an I-frame or an N-frame is the correct MEDL or which was connected
transmitted. crosswise to the channels from transmit-
This is followed by three bits, by means ting or receiving successfully.
of which a change of current cluster mode The C-state for initializing the CRC cal-
can be requested. These bits are set to culation is also used in the transmission of
zero if no change has been requested. frames with an implicit C-state. Thus, only
those frames whose CRC checksum was
Application data determined with the same C-state are ac-
The information of the application is trans- cepted as valid.
mitted in this area. The size of this field The calculated CRC value is transmitted
can be determined for each slot of a round at the end of a frame in a 3-byte field.
individually in the MEDL and must not
exceed 240 bytes per slot.
Composability Standardization
Networking technology has changed The development of TTA stretches back
greatly over the years. Where initially over a period of 20 years. Today, develop-
stand-alone systems (e.g. activation of a ment of the Specification is coordinated
turn-signal lamp via switches and relays) by the TTAgroup, a cross-industry consor-
were used, the changeover was quickly tium which was founded by the companies
made to cooperative systems in which Airbus, Audi, Delphi, Honeywell, PSA Peu-
individual components exchange informa- geot Citroën, Renault, and TTTech.
tion with each other. The Specification of TTP/C has been in-
However, the interaction of components fluenced in particular by Hermann Kopetz,
is increasingly making it difficult to ensure who between 1979 and 1982 was Profes-
the correct fulfillment of a function. For sor of Computer Process Control at the
this reason, the aspect of composability, Technical University of Berlin and since
i.e. the possibility of combining separately 1982 has run the Institute for Technical
developed subsystems into an overall sys- Information Technology at the Technical
tem without having to verify the function University of Vienna. Professor Kopetz
of the overall system again, is becoming also directed the MARS project and was
increasingly important. substantially involved in each further de-
In order to ensure that this functions, velopment of the protocol.
it is necessary to ensure that a subsystem
autonomously fulfills its functions, but also Characteristics
that this is the case when the subsystem is • Support of communication for safety-
integrated in an overall system. relevant functions.
TTP/C facilitates the implementation of • Guaranteed transmission properties.
composability through the communication • Guaranteed detection of all single faults.
parameters which are stored in the MEDL. • Fault-tolerant communication units

(FTU) for fault-tolerant provision of
Verification of the Specification data.
The functionality of TTP/C has been thor- • Local and central bus guardians
oughly analyzed. The core algorithms possible.
of the protocol specification have been • Low protocol overhead.
formally verified with regard to certain as- • Data efficiency of 85 % for transmission.
pects of consistency, stability and safety. • Supports composability with regard to
Furthermore, the function of TTP/C sys- time response and ranges of values.
tems has been tested using failure-injec- • Functions for the most part formally
tion experiments. This involved the use of verified.

both physical and software faults to check • Consistent view by all nodes of the clus-
the fault tolerance and fault-detection ter state thanks to Membership Service
properties. and Clique Avoidance Algorithm.
• Clear structuring of the protocol inter-
faces.
• No restriction of the bit rate from the
Specification. TTP/C components were

tested with a bit rate of up to 1 Gbit/s.
• Experience from product use.
124 Bus systems | FlexRay
FlexRay bus access is controlled by the use of mini-

slots. Minislots are small time windows
Overview in which a defined message can be trans-
FlexRay is a field that was designed to ferred. This access method is also known
support open and closed-loop control as FTDMA (Flexible Time Division Multi-
technologies in the automotive sector. ple Access).
Development focused on the suitability With messages transmitted in the static
for use in active safety systems in particu- component of the cycle, the transfer prop-
lar. FlexRay therefore offers transfers with erties can be assured in accordance with
guaranteed compliance with transfer known methods. In the dynamic compo-
properties, high bit rates and a fault- nent, messages are sent as required. In this
tolerant design. case, messages are prioritized based on
The development of FlexRay, managed the message ID. The division between the
by the FlexRay Consortium, goes back to static and dynamic component is freely
cooperation between German automotive configurable but it cannot be modified
manufacturers, BMW and DaimlerChrysler. later during system operation. The same
In 1999 these two companies began to applies to the slot lengths in the static
compile requirements for a new communi- range, which, while configurable, must re-
cation system. In the process, develop- main constant once operation is underway.
ment would incorporate experience FlexRay systems can be equipped with
gained from the byteflight system of two transmission channels, with each
BMW and the prototype development channel having separate lines. This makes
of DaimlerChrysler. it possible to create a redundant data
The aim of FlexRay is to provide a sys- transfer or increase the available band-
tem with high transfer rates that will work width for specific applications. The latter
in a deterministic and fault-tolerant man- can be used both for the parallel transmis-
ner while being as flexible as possible to sion of information of two network nodes
use and expand. The main fields of appli- and for the faster transfer of information
cation of the FlexRay are drivetrain sys- of one node.
tems (drive) and active safety systems with FlexRay operates at a maximum bit rate
no mechanical fall-back level (x-by-wire). of 10 MBit/s, which is achievable in opti-
However, the areas of passive safety sys- mal environmental conditions. A rate of
tems and comfort/convenience and body up to 20 MBit/s is achievable if two chan-
electronics are also supported. nels are used without redundant access.
To support these very different domains, In addition to the bit rate of 10 MBit/s,
FlexRay uses two different types of bus ac- which is the only bit rate defined in Speci-
cess. For deterministic transfer properties, fication 2.1 dated 15th December 2005,
it offers time-controlled bus access. In the there are endeavors to support lower bit
case of applications whose transfers place rates of 2.5 or 5 MBit/s.
less demanding requirements on transfer
properties, it is desirable to make as effec-
tive use of the available transfer capacity
as possible. To combine these two ap-
proaches, communication takes place in
cycles. In each cycle, there is firstly a static
transmission component in which bus ac-
cess is controlled by TDMA (Time Division
Multiple Access). There then follows a dy-
namic transmission component in which
Bus systems | FlexRay 125
Areas of application Active star topology

The areas of application of the FlexRay are An active star topology can be viewed as
very diverse. By offering redundant, time- a point-to-point link between a number
controlled and fault-tolerant transfers, of nodes and an active coupler. In this to-
not only is FlexRay suitable for use in pology, therefore, only the distance from
active safety systems with no mechanical a node to the coupler is of importance.
fall-back level and in the area of the drive- This cannot be more than 24 m.
train, but its optional, dynamic transmis- Apart from the connection lengths and
sion range means that it is able to support the number of connectable nodes, the
application in the areas of passive safety most important difference between an
systems and networking in the body and active and a passive star topology is in
comfort/convenience domains. the achievable bit rates. With the use of
With the high bandwidth of up to passive couplers, the achievable data
20 MBit/s for non-redundant transfers, rate (approx. 1 MBit/s) is below the rate
it is even conceivable, in theory, that that can be achieved with active couplers
FlexRay could be used for the transfer (up to 10 MBit/s).
of audio or highly compressed video.
Cascaded star topologies
Topology Cascaded star topologies are the product
Point-to-point of several active star topologies linked
The simplest configuration of a FlexRay together. In a FlexRay system, a topology
system is the direct link between two like this is limited to three cascaded star
network nodes. The maximum distance topologies.
between the two nodes is 24 m in this
configuration. Hybrid topologies
Hybrid topologies are a blend of topolo-
Bus topology gies from bus and star topologies, e.g. the
Between 4 and 22 nodes can be connected connection of several bus topologies to a
in a bus topology. The maximum distance star topology instead of individual nodes.
between any two nodes in a bus topology
may be 24 m. Two-channel topologies
Since it is possible to implement both
Star topology channels of a FlexRay system indepen-
A star topology is another architectural dently, different topologies can be used
variant of the FlexRay network. Both ac- for each channel. For example, one chan-
tive and passive couplers are supported. nel could be realized as an active star
topology, the other as a bus topology.
Passive star topology
3 to 22 nodes can be connected in a pas-
sive star. Here, too, the distance between
any two nodes cannot be greater than
24 m.
Hardware mission. The information received by the

FlexRay controller of a subscriber communication controller is forwarded to
A FlexRay node (Fig. 1) comprises the the actuators responsible for processing.
host processor, a communication control-
ler (CC) and one bus driver (BD) for each Communication controller
channel to which the node is connected. The communication controller (CC) real-
Optionally, each node may contain a izes all aspects of the FlexRay system
bus guardian (BG) for monitoring each relating to the protocol. Its tasks include
bus driver and the bus guardian may re- scheduling, synchronizing with other net-
ceive information through an additional work nodes, creating a macrotick signal,
host processor. creating a bit stream from the information
It is optional whether a network node of the host or controlling bus access.
is connected to one or both channels;
here, at the very most, the requirements Bus driver
from the field of application of the network The bus driver (BD) assumes the role of a
node take effect. transceiver. It converts logical information
into physical voltages that are carried on
Transmission agents copper wires, and vice versa. The role of
The transmission agent for a FlexRay sys- the BD also includes a protective function
tem is of a twisted-pair cabling design, against electrostatic discharge (ESD).
where shielded (STP) or unshielded (UTP) Furthermore, the BD monitors the states
cabling may be used. of the BP and BM and is thereby able to de-
Each of the two FlexRay channels con- tect whether the bus is affected by physical
sists of two wires designated bus plus (BP) faults.
and bus minus (BM). The BD also processes wakeup signals
and, optionally, is able to control the
Host processor power management of the network node
The host processor gathers information because functions of the node can be
from the sensors, which is forwarded to switched off.
the communication controller for trans-
1 Network node architecture
Communication data
Host Configuration data & Communication controller
status information
Synchronization
Communication data
Communication data
signals
Configuration data & Bus guardian
Control signal
Control signal
status information (optional)
Control Status
signals signals
Control data &
status information Bus driver
UVF0001E
Control signal (optional)

Power supply
Bus guardian Local bus guardian

The bus guardian (BG) is a device that will The local bus guardian (node-local bus
only permit transmissions if the connected guardian) is implemented as a stand-alone
network node is authorized to send a mes- CC inside the network node that monitors
sage. With this system, it is theoretically operation of a CC (Fig. 2). Its synchroniza-
possible for faulty network nodes to inter- tion process is independent of that of the
fere at the very most with their own trans- CC's; in particular, the CC and BG have
mission but not with the transmissions of separate oscillators.
other network nodes. In addition, it is also possible for the BG
The FlexRay protocol only provides the and CC to be controlled by dedicated host
framework for use of a BG: its use is not a processors that communicate with each
mandatory specification. Specification 2.0 other. This is, however, not mandatory.
only described a local bus guardian that As a consequence of direct control lines
would operate inside the network node (bus-guardian enable, BGE) between the
and monitor access to one of the channels. BG, CC and BD, information can only be
A critical aspect of this concept was that transferred from the CC to the BD if the
correct operation of the bus guardian in a BG permits this transfer to take place.
faulty network node could not be guaran-
teed as a consequence of the integration Central bus guardian
of the bus guardian in the node and the The central bus guardian (CBG, Fig. 3) was
permission for the bus guardian to access a new feature introduced with Specifica-
the oscillator of the network node. tion 2.1. It operates in the coupler of a
Specification 2.1 contains a refined star topology and monitors the operation
concept for the local bus guardian and of all connected nodes or networks. Faults,
introduces the concept of a central bus such as short-circuits on buses or network
guardian. nodes, that transmit information like a
“babbling idiot” with no consideration of
the transmission window, and thus inter-
fere with orderly communication, can be
detected by a CBG and suitable counter-
2 Local bus guardian 3 Central bus guardian
ECU Node 1 Node 2 Node 3 Node 4 Node 5

Host1 e.g. SPI Host2
CC BGSM BG (CC)
BGEout_A
BGEout_B
TxENin_A
TxENin_B
BGEin_A
BGEin_B
TxEN_A
TxEN_B
RxD_A
RxD_B
RxD_A
RxD_B
TxD_A
TxD_B
CBG CBG
A B Subbus
BD BD
UVF0002E
UVF0003E
Channel A Optional Optional Optional

Node 6 Node 7 node 1 node 2 node 3
Channel B
measures may then be taken. As a result, the bus driver looks out for the transmis-
the effects of these faults can be confined sion of a wakeup signal, in response to
to small areas of the network. which the node is set to wakeup mode.
Operating modes Sending and receiving

In addition to normal operation, a FlexRay A bus driver connects a FlexRay network
system also supports the complete initial- node to the channel that contains a re-
ization of the network from sleep to ceiver and transmitter. If a controller is
standby (Fig. 4). In order for this system connected to both channels, it follows that
start to function, the BDs of all network there must be two receivers and transmit-
nodes have the capability to detect wakeup ters available in the network node.
signals on the bus and to start the remain-
ing other components of the dedicated Protocol
network node. Protocol layers
Special network nodes that may be con- The FlexRay protocol is built on five core
nected to the starter motor of a vehicle, mechanisms:
for example, generate the wakeup signal. • Coding and decoding
The system start is split into the wakeup • Control of bus access (media access
and startup phases. The wakeup phase in- control, MAC)

volves the activation of the network nodes • Processing of frames and symbols
by the wakeup signal. In the startup phase, • Clock synchronization and
the network nodes are initialized and syn- • Schedule monitoring by the bus
chronized. guardian
If the function of a node is not required
for the time being, the node can be set to Unlike in other network protocols, with
a standby mode in which all operations FlexRay there are interfaces between all
of the coding and decoding process are core mechanisms. This requires a process
stopped. This reduces the energy con- that coordinates and synchronizes the
sumption of the node. In sleep mode, the changes in the core mechanisms. This is
power consumption of a network node is the task of the protocol operation control
at its minimum. In this mode, all functions (POC).
of the network node are deactivated: only
4 Wakeup sequence
local wake POC state leave coldstart

wakeup event channel A inhibit mode
Node A
wakeup/coldstart node power config ready wakeup wakeup ready integration listen coldstart listen
channel A, B off/reset listen send
wake channel B
Node B
wakeup/coldstart node power off/reset config ready wakeup wakeup ready coldstart listen
channel A, B listen send
Node C
non coldstart node power off/reset config ready integration
channel B listen
Channel A wakeup
UVF0004E
pattern
wakeup
B pattern
FlexRay gives the host system the capabil- Coding on the physical layer
ity of influencing all five core mechanisms For coding, FlexRay uses a NRZ method
directly. To arrange this, communication (Non-Return to Zero) in which the two
takes place through an interface – the con- identical transmission states are not di-
troller host interface (CHI) – between the vided by a range in which the voltage falls
network node and host. back to a zero value. This type of coding
Concentrating on the main interfaces, makes it necessary to have mechanisms
the task areas can be arranged as a proto- with which a network node can divide the
col stack (Fig. 5). At the top level is the ap- transmitted states. To this end, FlexRay
plication, which forwards its commands to adds a sequence behind each transmitted
the CHI. Under the CHI is the POC, which byte: the byte start sequence (BSS). From
in turn has access to the MAC, clock syn- this sequence, each node can detect when
chronization and frame & symbol process- a byte has been transmitted and, with this
ing. In the next lower layer are the pro- information, encode the individual bits
cesses for coding and decoding. The low- of the transmission.
ermost level is represented by the transfer By applying different voltages to the
characteristics of the physical layer. two wires of one channel, it is possible
The bottom three layers are in nodes to create four bus states, which are desig-
that are connected to both channels: nated Idle_LP, Idle, Data_0 and Data_1
double presence because these functions (Fig. 6).
have to be fulfilled for each channel sepa- A bus state is identified by measuring
rately. More precisely speaking, this the differential voltage. Here, the bus volt-
means that for one node connected to age (uBus) is made equal to the difference
both FlexRay channels there is double between uBM and uBP (voltages at BP and
availability of synchronization, MAC, BM).
frame & symbol processing, coding uBus = uBP - uBM
and decoding, and physical layer.
5 Protocol stack 6 Bus states and voltages
Application V
Controller host interface Idle_LP Idle Data_1 Data_0
Protocol operation control

uBP uBus
Media Clock Frame & 2.5
access synchro- symbol
control nization processing
Voltage
Coding and decoding uBM

UVF0005E
UVF0006E
0
Physical bus
Time
With this method, the data transfer is pro- Generation of a frame bit stream
tected against external electromagnetic Before a node can transmit a frame con-
interference because these act equally on taining the data of the host, the frame is
both wires and are canceled out in the dif- converted into a bit stream. To this end,
ference. the frame is first decomposed into individ-
Idle_LP (LP = low power) is the state in ual bytes. The start of a frame is populated
which a very low voltage of -200 mV and with a transmission start sequence (TSS)
200 mV is present at BP and BM. This state, followed by a frame start sequence (FSS)
for example, is used to identify the start of (Fig. 7). From the bytes of the frames, an
a transmission. extended byte sequence (EBS) is then gen-
In Idle state, a voltage of 2.5 V with a tol- erated whereby each frame byte is pre-
erance of 500 mV is present at BP and BM. ceded with a byte start sequence (BSS).
To set the channel to Data_0 state, at The 24-bit checksum (CRC) for this
least one transmitting node must apply bit sequence (TSS+FSS+EBS) is now calcu-
a negative differential voltage of -600 mV lated and appended to the bit sequence.
to the channel. To finish the bit stream, a frame end
For the channel to be set to Data_1 state, sequence (FES) is appended.
at least one transmitting node must apply If the frame belongs to the dynamic seg-
a positive differential voltage of 600 mV ment, a dynamic trailing sequence (DTS)
to the channel. can be additionally appended to prevent
If the transmission of an information another node from beginning its transmis-
signal is neither blocked by the bus guard- sion on the channel prematurely.
ian nor the communication controller, a
HIGH bit is signaled by the Data_1 channel Communication cycle
state and a LOW bit by Data_0. Each FlexRay cycle (Fig. 8) contains a
static segment that is transmitted as the
first part of the cycle. The static segment
contains a fixed number of transmission
ranges, the “static slots”.
Optionally, there may be a dynamic seg-
ment in the FlexRay cycle and this is trans-
ferred in second place. Each dynamic seg-
ment contains a fixed, freely configurable
7 Bit stream (dynamic segment) number of “minislots”.
Also optional is the symbol window,
MAC MAC which is transmitted as the third element
FSP MAC FSP
of the cycle. It can be used for the trans-
mission of an individual symbol and has
CODEC a b cd the same size as a static slot.
BSS BSS BSS DTS
To terminate the cycle, there is a phase –
FSS FES
network idle time (NIT) – in which the bus
High
is set to idle state. In general, the length
TxD
Low of the NIT corresponds to the remaining
TSS 1st byte last byte macroticks not used by the static and dy-
sequence sequence namic segment or symbol window. This is
1* gdBit
High not the case if it was detected during syn-
UVF0007E
MSB LSB
TxEN chronization that an offset correction is
Low required through which the length of the
NIT can be increased or reduced.
Static segment To identify the end of a transmission,

The static segment comprises a fixed num- a phase in which the channel is idle (dy-
ber of equally sized transmission windows namic slot idle phase) is added to the end
(static slots), which are sent simultane- of a dynamic slot. The phase in which the
ously on both channels. Exactly one frame transmission takes place is called the dy-
can be sent in each static slot. Both the size namic slot transmission phase.
and quantity of the static slots in a static Another difference between the dy-
segment are configured during integra- namic and static segment is that the coun-
tion. ters of slots transferred are incremented
Bus access in the static segment is synchronously on both channels in the
controlled by a TDMA method whereby, dynamic segment. In the dynamic seg-
in each slot, the frame with the corre- ment, the counter of both channels is in-
sponding frame ID is sent. Frame ID 1 cremented independently in line with
is therefore sent in slot 1, frame ID 2 in the current transmission status (Fig. 9).
slot 2 and thus frame ID n in slot n.
Dynamic segment 9 Slot counter
Unlike in the static segment, the transmis-

Slot counter channel A minislot
sions in the dynamic segment are not of
gNumber-
fixed length. To control bus access, “mini- m minislot 1 m+1 m+2 m+3 m+4 m+5 OfMinislots
Channel B Channel A
slots” are used that hold the same fixed frame ID m frame ID m+3 frame ID m+5
number of macroticks for all nodes con- t
nected to the network. Macroticks are
frame ID m+3 frame ID m+7
ranges in which exactly one frame can
m m+1 m+2 m+3 m+4 m+5 m+6 m+7 m+8
be transmitted.
“Dynamic slots” (transmission windows Slot counter Dynamic slot Dynamic slot
channel B with without
of variable size) are now created based on transmission transmission
the minislots. The maximum size of a dy-
Dynamic segment containing
namic slot is limited by a fixed number of
UVF0009E
gNumberOfMinislots minislots
minislots specified during configuration. Transmission may only start within the first
pLatestTx minislots of the dynamic segment
8 Communication cycle
t
Communication
cycle level
static segment dynamic segment symbol window network
idle time
Arbitration
grid level
static slot static slot minislot minislot
action point action point action point
Macrotick
level
macrotick
UVF0008E
Microtick
level
microtick
Symbol window For this purpose, selected messages from

The symbol window makes it possible to the static part of the cycle are used (sync
send a collision avoidance symbol (CAS) frames) and are transmitted by the net-
or a media access test symbol (MTS). work nodes that are connected to both
These two symbols are identical and channels.
help to prevent collisions during the In a distributed communication system,
system-startup phase. each network node generally needs its
The third symbol defined in the FlexRay own internal clock for setting transmission
protocol – the wakeup symbol (WUS) – is timing correctly. As a consequence of tem-
not permitted to be sent in the symbol win- perature and voltage fluctuations and pro-
dow. It serves only to generate a wakeup duction-related precision tolerances, these
pattern (WUP), which is used during the internal clocks can soon deviate from an
system startup. abstract, system-wide global time. In the
case of a system such as FlexRay, which
Network idle time (NIT) controls bus access by means of time slots,
During the NIT, a network node corrects the synchronization of all network nodes
the ascertained time deviations of its inter- plays an especially important role because,
nal clock. It also gives the network node without a standardized, global time preset
the opportunity in this phase to make im- for all nodes, the specified transmission
plementation-dependent adjustments and windows cannot be maintained and may
changes to settings relating to the commu- break down.
nication cycle.
Clock control hierarchy
Clock control In a FlexRay system, clocking is controlled
In both the static and the dynamic part of on three levels (Fig. 10). The bottom level
the communication cycle, FlexRay reverts is represented by the microticks, which
to the identifiers of the transmission win- are derived directly from the oscillator
dow in which a node may transmit. If this clock of the network node. A microtick
method is to be successful, it is necessary is therefore not a system-wide variable,
for all nodes connected to the network but only relevant to the node concerned.
to have synchronous time information. At the next level up, the clock is con-
trolled by means of macroticks – ranges
in which exactly one frame can be sent.
10 Timing Within fixed tolerances, macroticks form
the lowermost system-wide time unit.
t
Communication Each network node determines the dura-
cycle level 0 1 2 3 4 5 … 62 63 tion of a macrotick in microticks. The
number of microticks per macrotick does
vCycleCounter cCycleCountMax
not have to be the same for all macroticks.
gdCycle The top level is characterized by a clock
Macrotick control based on the communication cycle,
0 1 2 3 4 5 … n-1 n
level which always contains the same, fixed
gMacroPerCycle -1
vMacrotick number of macroticks.
gdMacrotick
Microtick 0 1 2 3 4 5 …
UVF0010E
level
vMicrotick pdMicrotick
Time synchronization Gradient correction

A synchronization method is required in With gradient correction, deviations in
order to guarantee a standardized time the transmission frequency of a node are
across the entire system. The basic para- continuously corrected during the com-
meters for timing are the number of munication cycle. The node here checks
microticks per communication cycle the transmission frequencies of all other
(pMicroPerCycle), the duration of a com- nodes. By means of the FTAA, these
munication cycle (gdCycle) and the dura- values are used to adjust the node's
tion of a microtick (pdMicrotick), where own transmission frequency.
the relationship between each is deter-
mined by the following equation: Offset correction
Offset correction is a way to correct devia-
gdCycle
pMicroPerCycle = round( _____________ ) tions in the communication-cycle phase
pdMicrotick
by including an offset in the NIT. With this
During operation of a FlexRay network, method, the node similarly employs the
the clocks of the oscillators of various net- FTAA to determine its phase difference
work nodes will not be equal or even sim- in relation to the other nodes of the net-
ply remain constant. They will fluctuate work and uses it to determine the offset
in response to outside influences. This required to shorten or extend the NIT
deviation is balanced by two corrective phase. Since all nodes use this approach,
measures: the gradient correction and the all nodes can initiate the transmission of
inclusion of an offset in the communica- a cycle in a synchronous manner.
tion cycle.
Timing
Fault-tolerant average algorithm Transmissions in the FlexRay system
Since an individual node can only observe are timed by action points, which are
the differences between the expected time specially-marked macrotick boundaries.
at which a transmission should have been A transmission in a static slot begins once
actioned and the actual time at which the a static-slot action point has been reached.
transmission took place, a fault-tolerant The position of this action point in a mac-
average algorithm (FTM), as described by rotick is defined system-wide.
J. L. Welch and N. A. Lynch, is used in both In a dynamic slot, a transmission begins
corrective measures. The measured time once a minislot action point has been
differences of all transmitted sync frames reached and ends at a subsequent minislot
are sorted by their value. If more than action point. This method ensures that
seven sync frames were received, the two there is always an idle period between
highest and two lowest values are removed individual transmissions.
from the list. If between three and seven
sync frames were received, the highest and
the lowest value are removed. For fewer
than three sync frames, the average is
derived from all measured differences.
Welch, Lynch,
“A New Fault-Tolerant
Algorithm for Clock
Synchronisation”,
Information and
Computation, vol.77,
No.1, April 1988
Frame format The startup frame indicator reveals that

FlexRay uses the same format in both this frame is a startup frame. Startup
the static and dynamic part of the frame. frames are used in the network's starting
The format can be broken down into three phase and may only be sent by special
parts: header, payload segment and trailer coldstart nodes.
segment (Fig. 11). The indicators are followed by the
frame ID, payload length, header CRC and
Header segment cycle count.
The header segment contains a total of The frame ID has a length of 11 bits and
five bytes, at the start of which is a series corresponds to the number of the slot in
of indicators. which the frame is sent. In one cycle, each
The reserved bit is intended for future frame ID exists only once on each channel.
modifications to the protocol and is sent as The payload length expresses the size
a logical “0”. of the information sent in the payload seg-
The payload preamble indicator indi- ment. Since the payload in a FlexRay frame
cates whether or not the payload segment is always sent in 2-byte words, a 62 in the
contains a network-management vector payload length field indicates, for exam-
(NMVector). An NMVector enables the host ple, that 124 bytes of payload are being
processor to transmit data directly without transmitted in the payload segment.
first being processed and prepared by the In the static segment, the payload length
CC. field always contains the same value
The null frame indicator indicates a null because the size of information here is
frame that contains no usable information. constant for all frames. Of course, this
The sync frame indicator indicates that property does not apply to frames in the
this frame is intended to be used for sys- dynamic segment.
tem synchronization. This indicator may In the header CRC, an 11-bit CRC value
only be set by network nodes that operate is transmitted. This is formed from the
as sync nodes. polynomial
x11 + x9 + x8 + x7 + x2 +1.
Since the information in the header seg-

ment is constant for all frames in the static
segment, the header CRC value can be cal-
11 Frame format culated offline and assigned to the network
node through configuration settings.
Reserved bit If a frame in the dynamic segment is
Payload preamble indicator
Null frame indicator always of the same size, it is possible here
Sync frame indicator also for the CRC value to be calculated off-
Stratup frame indicator
line.
Header CRC
Covered Area The final field in the header segment
is the cycle count. This field contains the
Pay- count number of the cycle in which the
Frame load Header Cycle Data Data Data Data CRC CRC CRC
ID length CRC count 0 1 2 n
sending network node is currently active.
11 bits 7 bits 11 bits 6 bits 0…254 bytes 24 bits
Header Payload Trailer

Segment Segment Segment
UVF0011E
FlexRay Frame 5 + (0 … 254) + 3 bytes

Payload segment development of FlexRay, depending on

The payload segment has a maximum the obligations that they have accepted.
length of 254 bytes, which are transmitted The uppermost group is represented by
in 2-byte words. the core partners, which first and foremost
The payload segment customarily trans- include the founding members. Alongside
fers the payload that is further processed the core partners are the premium associ-
by the host processors. Optionally, how- ates and associate memberships.
ever, an NMVector may also be sent in the By 2004 General Motors, Bosch and
payload segment, or a 16-bit message ID. Volkswagen had joined the FlexRay Group
With the assignment of message IDs, as core partners. In 2004 Freescale took
it is possible to send several data blocks in over Motorola's position in the group of
a frame. For this purpose, the message ID core partners.
is placed at the front of the other data by
the host processor as application data. Characteristics
• Deterministic transfers possible
Trailer segment • Optional bus guardian monitors network
The trailer segment contains a single access and protects against faults
field in which a 24-bit CRC checksum • Differential signal transmission
(frame CRC) is sent. The polynomial used • Collision-free transmission while opera-
to determine the frame CRC is: tion is in progress

x24 + x22 + x19 + x18 + x16 + x15 + x14 + • Redundant transmission of information
x13 + x11 + x10 + x8 + x7 + x6 + x3 + x+1 on two channels
• High transmission speed of up to
Standardization 10 MBit/s; up to 20 MBit/s with parallel
FlexRay Group transmission on two channels
FlexRay is a protocol specified by the • Startup and initialization from an inac-
FlexRay Group. The FlexRay Group was tive sleep state are directly supported
founded in 2000 by the companies, BMW, by the protocol
DaimlerChrysler, Motorola and Philips • Support for diverse fields of application
Semiconductors (www.flexray-group.com). • Event and time-controlled transmission
The members of the FlexRay Group are of information possible

organized in a hierarchy. They also have • Published standardization by the
different levels of influence on the further FlexRay Group

136 Bus systems | Diagnosis interfaces
Diagnosis interfaces ple, individual actuators can be targeted

for activation to check them for correct
Diagnosis legislation (e.g. CARB, California operation. At the end of the line (EOL),
Air Resources Board) demands constant the diagnosis interface is used to test the
monitoring of emission-related compo- electronic control unit and make changes
nents in various electronic systems (OBD, to configurations (e.g. immobilizer, trans-
On-Board Diagnosis). Faults (e.g. electrical mission type). It is even possible to pro-
short-circuits of sensors, implausible op- gram the entire flash EPROM using end-
erating states) are stored in a fault memory of-line programming.
in the control unit. These faults can be
read using a scan tool (official testing sta- Diagnostics requires an interface for con-
tion) or workshop tester. For this purpose, necting the tester to the electronic control
there is a socket in the footwell, dashboard units. For communication, there are basi-
or center console of the vehicle to which cally two options available at the present
the tester is connected by means of a stan- time:
dard connector (ISO 15031-3) (Fig. 1). • K line and
Using a workshop tester, the after-sales • CAN bus
service is able to read the entire fault
memory: emission-related faults and The K-line network is not a bus in the true
vehicle-specific faults (e.g. from Motronic, sense, which means that collisions may
ABS). With the information stored in the occur. Since most electronic control units
fault memory, it is possible to diagnose a have a CAN interface for the exchange of
fault and repair it efficiently. The tester data between electronic systems, this op-
can also be used to clear the fault memory. tion for communication with the tester is
Furthermore, measured values and con- gaining ever more in importance.
trol-unit identification values can be evalu- Various protocols are implemented on
ated. Using the workshop tester, it is also the interfaces. These tend to be the CARB
possible to control special diagnostic func- or official protocol and the manufacturer-
tions. With actuator diagnosis, for exam- specific communication, which can be
closely related in their properties and
functions. It is even possible to use both
1 Diagnosis connector (diagnosis interface) interfaces: CARB-relevant functions run on
the CAN, manufacturer-specific functions
on the K line.
1 2 3 4 5 6 7 8
Figure 2 provides an overview of the
9 10 11 12 13 14 15 16
standards and the various layers of the
OSI reference model.
Pin 2: Bus line (high) from SAE J1850
Pin 10: Bus line (low) from SAE J1850
Pin 7: K line in accordance with ISO 9141-2
and ISO 14230-4
Pin 15: K line in accordance with ISO 9141-2
and ISO 14230-4
Pin 6: CAN_H from ISO 15765-4
Pin 14: CAN_L from ISO 15765-4
Pin 1, 3, 8, 9, 11, 12, 13
Not used by OBD
UWT0106-1E
Pin 4: Vehicle ground

Pin 5: Signal ground
Pin 16: Battery positive
Bus systems | Diagnosis interfaces 137
Diagnostic protocol CARB

The communication method in diagnostics The Californian environmental authority
is characterized by the fact that the tester (CARB) introduced guidelines to promote
(client) addresses one or several electronic support for a standardized interface.
control units (servers) and requests data The CARB protocol is built on existing
output (e.g. from the fault memory) or an standards and specifies the data flow and
action (e.g. actuator diagnosis). These ser- timing (transport layer). This is established
vices and the transmission agent are de- in ISO 14230-4/ISO 9141-2 for the K line
fined in the various protocols. and in ISO 15765-4 for the CAN. The actual
communication messages (services) in the
KWP 71 application layer are precisely defined in
This protocol was developed in collabora- ISO 15031-5/SAE J1979 and permit no
tion with Bosch as the first standardized manufacturer-specific deviations.
serial interface and is based on a 5-baud
initialization and byte handshake. For ad- In California, every newly registered vehi-
dressing, the protocol provides for the use cle must support at least one of these pro-
of a unidirectional line (L line). The line tocols with all emission-related electronic
for data exchange (K line) is bidirectional. control units. The European counterpart
Alternatively, if the L line is not used at is the EOBD standard.
all in the system architecture, the K line
can also be used for addressing because
addressing and data exchange are two
chronologically sequential states. The way
in which communication is established is
predefined, the services are standardized.
2 Standards for diagnosis communication
KWP 2000 (ISO 14230: 1-3) a

This diagnostic protocol connects a tester
Layer
CARB
to the electronic control units using a K line CAN
K line. In the KWP 2000 standard, there 7 ISO 15031-5 ISO 15031-5 ISO 15031-5 ISO 15031-5
are various possibilities for the way in 6

which communication is established (fast 5 ISO 15765-4
initialization and 5-baud initialization).
4
The communication services are specified Fig. 2
ISO 15765-2
3
but there is some freedom for customer- ISO 15765-4 a CARB
ISO 14230-2 ISO 11898
specific adaptations. 2 ISO 9141-2
ISO 14230-4
SAE J1850
ISO 15765-4 communication
1 ISO 9141-2
ISO 14230-1
SAE J1850
ISO 11898 b Customer-specific
ISO 14230-4 ISO 15765-4
communication
ISO 15765: 1-3 (CAN) b
This standard reproduces the existing Manufacturer-specific
Layer
Layers of the
protocols, ISO 14230: 1-4, on the CAN. K line: KWP 2000 CAN / UDS
ISO 15765-3 OSI reference model
It defines a way to leave untouched the 7 ISO 14230-3
ISO 14229-1 7 Application
fundamental format of messages ex- 6 6 Presentation
changed between tester and electronic 5 ISO 15765-3
5 Session
control unit and to transmit them on the 4 Transport

4 3 Network
CAN bus.
3 ISO 15765-2 2 Data link
SVA0018E
1 Physical
2 ISO 14230-2 ISO 11898-1
1 ISO 14230-1 ISO 11898 UDS Unified Diagnostic

Services
Application protocols Communication on the K line

The increasing complexity of control-unit Physical layout
functions places high demands on the The diagnostic tester connects to one or
tuning and optimization of these systems several electronic control units using the
to suit the types of engine and vehicle con- K-line system (Fig. 3). All electronic con-
cerned. This process is known as applica- trol units have equal rights. Communica-
tion engineering. For this purpose, an tion flows bidirectionally between the tes-
application system is connected to the ter and the electronic control units. This
electronic control unit. The application means that data can be received and sent
protocol for the K line is the McMess pro- by all subscribers. However, only one sub-
tocol, and the CCP protocol for the CAN. scriber is ever able to send at a given time.
The communication is an asynchronous
McMess process. There is no additional line for the
With the McMess protocol, parameters in the transmission of a clock signal. The baud
electronic control unit can be modified and rate on the K line is between 1,200 and
operands recorded. In a special application 10,400 baud in accordance with ISO stan-
mode, the protocol describes the method dards. In special applications (e.g. manu-
and contents of communication between ap- facturer-specific flash programming),
plication system and electronic control unit it could even be as high as 250 kBd.
for the K line. With McMess, for example, the Data transfer on the K line is realized as
measuring device is able to read the contents follows: the resistor in the tester increases
of the variable memory (e.g. sensor values, the potential of the line – provided no elec-
measured values) quickly. tronic control unit is sending on it – to bat-
tery-voltage level (Fig. 4). If an electronic
CCP (CAN Calibration Protocol) control unit is activated, it connects its
The CCP protocol describes the contents K-line connection to the shared ground
of communication between an application and therefore switches the potential of the
system and an electronic control unit if K line to ground. Each of the connected
these communicate on the CAN bus. electronic control units is able to detect
CCP allows electronic control units in that one of the bus subscribers is actively
the network to function at higher payload accessing the K line.
rates than on the K line.
3 System architecture of bus systems 4 Circuit diagram of an electronic-control-unit network
CAN bus
ECU 1 ECU 2 ECU 3 ECU 1 ECU 2 ECU n
K line
UBat
Diagnosis interface
Tester
SVA0017E
SVA0019E
Tester
Message format Checksum

The message format of the KWP 2000 The 1-byte checksum is always positioned
protocol comprises three parts (Fig. 5): at the end of a message. It contains the
• Header value of the modulo 256 sum across all
• Data bytes and bytes in the message – apart from the
• Checksum checksum itself.
Header Initialization
The header is made up of no more than To be able to send messages, each elec-
four bytes. tronic control unit must support a subset
• The format byte (Fmt) contains informa- of the possible message formats. At the
tion about the form and composition of beginning of communication, the format
the message. Two bits (A0 and A1) pro- supported by the electronic control unit
vide address information, while six bits is communicated to the tester by means
(L0 to L5) provide the quantity of data of key bytes as part of initialization.
bytes. With this coding, there can be up The ISO standards specify which of the
to 63 data bytes. If L0 to L5 are set to possible initialization procedures are cus-
zero, the number of payload bytes must tomary for ECU diagnostics. The following
be coded using the length byte (Len). initialization procedures are supported:
• The target address (Tgt) indicates the • 5-baud initialization and
communication partner for which the • Fast initialization
information is destined.
• The source address (Src) indicates the
communication partner that sent the
information.
• The length byte (Len) specifies the quan-
tity of payload bytes (max. 255 bytes).
Up to a value of 63, the coding can take
place in Fmt. For more than 63, the 5 Message format of the KWP 2000 protocol
length must be specified using this byte.

Check-
Header Data bytes sum
The existence of a target and source ad-
dress as well as the length byte is depen- Fmt Tgt Src Len SId
dent on the parameters from the format
byte. max. 4 bytes max. 255 bytes 1 byte
Checksum calculation
Data bytes
Up to 255 bytes of payload can be sent per
message. The first byte of the payload is
always a service-identification byte (SId). A1 A0 L5 L4 L3 L2 L1 L0
The subsequent data bytes are data that
differ depending on the service con-
cerned. 0 0 No address information
0 1 Exception mode (CARB)
1 0 With address information,
physical addressing
SVA0020E
1 1 With address information,

functional addressing
These different initialization methods are recognition pattern. From the bit sequence
used either for CARB diagnostics or for 01010101 (binary format), the tester de-
after-sales diagnostics. Furthermore, a termines the baud rate of the electronic
distinction can be made between physical control unit, which may be between 1,200
initialization and functional initialization, and 10,400 baud. The electronic control
which introduces the topic of “point-to- unit then transmits the two key bytes,
point” or “point-to-multi” communication. which communicate the header format
With functional initialization, a group of and timing method that the electronic
electronic control units is addressed and control unit supports. As acknowledgment
initialized. A function initialization can that communication has been successfully
only be successful if all the electronic con- established, the tester sends the comple-
trol units of this group support the same ment of the second key byte back to the
baud rate, the same transmission timing electronic control unit and, in return, re-
and the same protocol. By contrast, only ceives from the electronic control unit the
one electronic control unit is initialized in complement of the address. Initialization
the case of physical initialization. is then complete and the regular data
transfer may begin.
5-baud initialization
Communication is initiated by means of Fast initialization
a 5-baud address sent by the tester. It is Fast initialization establishes communica-
thus possible for each electronic control tion in a shorter time than 5-baud initial-
unit to be addressed individually and ad- ization. With fast initialization, the tester
dressing is clearly separate from the data sends a wakeup pattern (WuP) comprising
transfer. a low and high phase, each lasting 25 ms
The tester outputs the address onto the (Fig. 7). At the end of the WuP, the tester
K line at a transfer rate of 5 baud (Fig. 6). sends the StartCommunication Service at a
10 bits (eight data bits, one start and one rate of 10,400 baud. The electronic control
stop bit) take approx. 2 seconds. unit returns a positive reply containing the
On receiving the address, the electronic key bytes. Initialization is then complete
control unit responds with the baud-rate and regular communication may begin.
6 5-baud initialization 7 Fast initialization
Tester Control unit Tester Control unit
5-baud initialization Fast initialization

Address (2s) TWuP
5 baud 0x55 KB1 KB2 KB2 Address
W5 W1 W2 W3 W4 W4 Tidle
1,200 to 10,400 baud TiniL 10,400 baud

Synchronization
W1 60 to 300ms Baud-rate recognition
W2 5 to 20ms TiniL 25±1ms
SVA0021E
SVA0022E
W3 0 to 20ms TWuP 50±1ms

W4 25 to 50ms Tidle 300ms
W5 300ms
CARB initialization If a falling edge was detected, the elec-

CARB initialization is a special case of tronic control unit delays sending its
5-baud initialization, although CARB may own data until the bus has returned to
also be activated by fast initialization. idle state and the P2min time period has
A functional initialization of all connected again elapsed. Only then may it attempt
electronic control units is achieved with to send its data again.
the fixed address 0x33. The baud rate is
permanently set at 10,400 Bd. The require- Collision detection
ments are described in more detail in A collision cannot be prevented if two
ISO 14230-4 and ISO 9141-2. electronic control units send at the same
One of the outcomes of these standards time. In this event, the collision still has
is that all emission-related electronic con- to be detected. This is possible by monitor-
trol units fitted in a vehicle may only sup- ing the K line. To this end, the electronic
port one protocol: ISO 14230-4, ISO 9141-2 control units reread their own transmitted
or SAE J1850. A mixture of protocols in the byte through the serial interface.
same vehicle is not permitted. In Figure 8, two electronic control units
(ECU 1 and ECU 2) send a start bit at the
Key bytes same time. ECU 2 then sends a 0 (dominant
The electronic control unit uses key bytes bit) and ECU 1 a 1 (recessive bit). The 0
to inform the tester of which header bytes, appears on the K line. With bit 2, the exact
length bytes and timing method are sup- opposite is true. Here, ECU 1 persists with
ported. The key-byte decoding process the dominant level. When the complete
is defined in ISO 9141 and ISO 14230-2. byte has been transferred, the electronic
control units compare the byte sent on the
Arbitration K line with the byte they have read. In the
Arbitration is a means of preventing or de- example, a recessive level was overwritten
tecting collisions of data sent by different by a dominant level for both electronic
electronic control units at the same time. control units. From this, both electronic
Arbitration is necessary if a correct response control units deem the transfer to have
to the tester request is to be received in been faulty and retreat from sending.
the case of functional communication
(e.g. reading the fault memory of all emis-
sion-related electronic control units using
a scan tool). Collisions can be expected 8 Sequence of unsuccessful communication
because more than one electronic control

unit is permitted to send following initial- Recessive
ECU 1 level
ization by the tester.
Dominant
Arbitration is only relevant to data Start 1 2 3 4 level
transfer on the K line. bit
Collision prevention ECU 2

To prevent collisions following a tester re-
Start 1 2 3 4
quest, each electronic control unit checks
bit
the K line for a falling edge after a time of
P2min. A falling edge may be interference, Bus
or another electronic control unit has
SVA0023E
already been sending data. Start 1 2 3 4

bit
To prevent cyclical collisions, the elec- Different addressing methods exist:

tronic control unit reports after a time of normal and extended addressing, whereby
P2min, which is calculated in accordance normal addressing is the regular address-
with a specific algorithm. ing method and is analyzed below. Longer
Let us assume that ECU 2 in the previous messages that do not fit into a CAN frame
example sent a 1 as its first data bit. ECU 1 (seven bytes with normal addressing, six
would then receive nothing from the data with extended addressing) are segmented
transfer of ECU 2 and would continue to into several CAN frames and recomposed
establish communication. ECU 2, however, by the recipient.
detects the collision in bit 2 and retreats.
The collision in this event is non-destruc- Unsegmented messages
tive. If the quantity of data bytes to be transmit-
ted is no greater than seven bytes (for nor-
Communication on the CAN mal addressing), a single frame is trans-
Physical layout ferred. The first byte to be sent is the PCI
A CAN interface containing the necessary (Protocol Control Information). It identi-
hardware already exists in many electronic fies the frame in the high nibble (top four
control units. This bus system can also be bits of a byte) as being a single frame and
used for diagnostics. It is possible to com- indicates the data length (DL) in the low
municate with the electronic control units nibble (Fig. 9).
using a diagnostic tester connected to the The data is sent by the transmitter to the
bus by a diagnosis connector. receiver in a single message (Fig. 10a).
Baud rates Segmented messages

The typical baud rates associated with If more than seven data bytes (for normal
CAN are 500 kBd or 1 MBd. addressing) are to be delivered, the diag-
nosis tester (client) sends a first frame
Addressing and message types on the CAN data bus (Fig. 10b) first of all.
Tester communication on the CAN bus The electronic control unit (server) ac-
is defined by ISO 15765. The communica- knowledges with a flow-control frame.
tion services of ISO 15765-3 or 14 229-1 The other data is subsequently sent in
are defined in a similar way to those of consecutive frames.
ISO 14230-3 (Fig. 2).
The fundamental difference between
these protocols is in the format of mes- 9 Format of an unsegmented message
sages and how they are transmitted. While

up to 255 data bytes can be transmitted in
0 1 2 3 4 5 6 7
a message on the K line, only eight data
PCI Data
bytes are possible with the CAN. To repro-
duce the service messages on the CAN, the
contents (data bytes) of the header and the
checksum are separated and embedded in
a similar, new message format. The ad-
dress of the electronic control unit is now 7 6 5 4 3 2 1 0
the CAN identifier itself. With different 0 0 0 0 DL
Single frame
CAN identifiers, it is now possible to sup-
SVA0024E
port functional or physical addressing.

The first frame contains the PCI, an ad- (Separation Time) specifies the interval
ditional length byte (DL, Data Length) and to be maintained between consecutive
the first six data bytes (Fig. 11). In the frames.
high nibble, the PCI contains the informa- The flow-control frame is followed by
tion for identifying the frames as the first the consecutive frames. Again, the PCI
frame. The extended data length (XDL) in the high nibble contains the identifier,
is stored in the low nibble. Together with while the low nibble contains the sequence
the DL, a 12-bit data word is formed with number (SN). With the first consecutive
which it is possible to express values from frame to be sent, the SN is set to 1. The SN
0 to 4,095. is incremented with each subsequent con-
The transmitter sends the first frame secutive frame; after 15, the SN restarts
and waits for an acknowledgment of re- from 0. By evaluating the sequence num-
ceipt from the receiver in the form of a ber, the receiver is able to detect whether
flow-control frame. This message again or not all the frames have arrived.
contains the PCI, the high nibble of which
identifies the frame as the flow-control
frame. The low nibble contains the flow
status (FS), which can be used to authorize
or delay the sending of further frames
(consecutive frames).
11 Format of a segmented message
BlockSize (BS) indicates how many consec-

utive frames can be received without the
need for a further flow-control frame to a 0 1 2 3 4 5 6 7
be sent. The values expressed in STmin PCI DL Data
7 6 5 4 3 2 1 0
10 Diagram of message transfer
0 0 0 1 XDL
First frame
a Transmitter Receiver
3 2 1 0 7 6 5 4 3 2 1 0
XDL DL
Single frame
b 0 1 2 3 4 5 6 7
PCI BS STmin Data
b Transmitter Receiver
First frame 7 6 5 4 3 2 1 0
0 0 1 1 FS
Flow control Flow control
Consecutive
frame Clear to send
Fig. 10
STmin Consecutive 0 0 0 0 (CTS)
frame Block a Message transfer
0 0 0 1 Wait (WT),
size (BS) receiver in a single frame
Consecutive
frame not ready (unsegmented
c 0 1 2 3 4 5 6 7 message)
Flow control PCI Data b Segmented
message transfer
Consecutive
frame
Last block 7 6 5 4 3 2 1 0
SVA0026E
SVA0025E
Consecutive Fig. 11
frame ( ≤BS) 0 0 1 0 SN a First frame
Consecutive frame
b Flow control
c Consecutive frame
144 Automotive sensors | Basics and overview
Automotive sensors
The term sensor has become common, Basics and overview

as in the past 20 to 40 years measuring
gages have also come into use in con- Sensor, term/definition
sumer applications (e.g. motor vehicle The electrical sensor outputs are not only
and domestic appliance technology). provided in the form of current and volt-
Sensors – another term for measuring age, but are also available as current or
detectors or measuring sensors – convert voltage amplitudes, frequency, phases,
a physical or chemical (generally non- pulse widths, and cycles or periods of an
electrical) variable F into an electrical electrical oscillation, or as the electrical
variable E; this process often also takes parameters, resistance, capacitance, and
place over further, non-electrical inter- inductance. A sensor can be defined using
mediate stages. the following equation:
Table 1 summarizes and compares the (1) E = f (F, Y1, Y2…)

various fields of application for sensors. Sensor output signal
Figure 1 provides an overview of the (2) F = g (E, Y1, Y2…)
abundance of electronic vehicle systems Required measured variable
already on the market. Undoubtedly,
this number will increase immensely If the functions f or g are known, then they
in the years to come. represent a sensor model with the help of
which the measured variable sought may
be mathematically calculated from output
1 Automotive systems and their sensors

Co onic) and p rake) ic
Ra l loc enso tionin sor ing)

lic ron
il)
so n Ra
i n n
(M tio drau lect
n
Hi ic) c-pr otro nic) )
on ri (M tro nic
or
ntr e s nd se itio
At ck me ontro or )
itio
b
nd
c
thr eler gnos nso onic
Pr and per (air n)

y (e
o
Kn mas esel sen troni
ns
otr he r o tro
(ce ssur ir-co ure ond

(tra ed s en s omm
g)
os
) u
oh r
ng em so ati
se
(M sp nso r (M , Mo
ng so
tr
ctr nso
r
su )
o
mo sor ol)
Ac d di e s Mo
a at -c
es nic
ati ty/t en vig

Ai nic sure l, M
,C
ori tra
r
m)
jec sor
re
-up en ontr
c s
l
ele se
en
(he idi ity s r (na

le or tics)
oa su rol,
ox tion
sh or ste
nit , ul
ro
mb ct-in en
r
ol, al
c
a e
otr n
n
ec re nt
n-b es nt
ind ns sy
os ion r
k r
La dire re s
ntr d
Bo iss enso
(ba anc iper

Hu ua enso
(el t-p co
(O k-pr n co
ai
co -pe
mo se te
ra rt
(w n se ing
Ta iss or
yg
or ota
w
line su
Pr etr
fo
m s
tro s
n io
ns e s
di
Di ield-
Ai te s
ck e s
ns en
ns f-r
so es
ott at
o s
m
da
m l
(tra ssur
se le-o
r
(ga h-pr
iv
a
w-
r-q
Dr
st
e
c
r-
e
i
An
Sp
Ya
g)
Til tion enso ag)
)
(ai g)
SP or
P)
)
y ing
rs )
an sor or P)
)
sh
)
te d an (saf ABS
s
ing
BS
sin
ce rat or ba
eri or
ll-o stem
(E ens
ES
en ens r (ES
cra
rb
ft s rd
)
r ( ety
ng
ste ns
aim
Ac aw- ens r (air
en
)
(A
ns -the gua
r(
(
pre
s
ed se
or
so
le
mp
ti e
n so
af
ve
UAE0816-1E
ns
ng
C,
ist e
en
ss qu
tio sen
dla
se
AC
S
l-a
s
es
a s
r-a or
(ro
a
ler e
ee
d
pa n
he
ur
we T
ee
cu tio
da
wh
or
ss
r(
Sp
oc era
ra
ts
so
re
g-
se
-p
rin
Y
at- el
ing
en
Se Acc
gh
(po
ee
ts
ng
Hi
ra
St
Til
Ra
w-
Ya

Automotive sensors | Basics and overview 145
1 Areas in which sensors are used
Typical features Primary Precision Industrial Consumer

standards measurement measurement technology
Accuracy 10–11 to 10–7 2 to 5 · 10–4 2 to 5 · 10–3 2 to 5 · 10–2
€ 100 k to
Costs € thousands € hundreds € 1 to 10
€1m
Units/year single figures approximately 10 100 to 1 k 10 k to 10 m
Use – Research – Calibration – Process instru- – Auto
– Testing of mentation electronics
secondary – In-process – Building
normals measurement services Table 1
signal E and the influencing variables Yi These model parameters are generally
practically without error (intelligent or stored in a programmable, nonvolatile
smart sensors). memory component (PROM) in the digital
processes for conditioning sensor signals,
Adjustment which have come to dominate. By contrast
In the real world, the sensor model always with the conventional analog compensa-
includes some free parameters, with which tion of influencing variables, it is not only
the model can be adapted to the actual roughly linearly acting influences that may
properties of the individual sensor type be corrected, but also strongly nonlinear
in a kind of calibration process (Fig. 4a). processes that can be corrected success-
fully. Another significant advantage of this
type of calibration, which takes place over
2 Sensor symbol a purely electrical connection, is that each
sensor may easily be maintained under
operating conditions during the calibra-
tion phase.
Φ
▶ Concept of the smart sensor
E
In a somewhat more general form,
UAE0815Y
intelligent sensors (smart sensors)

may be defined as follows:
Intelligent sensors, sometimes also
known as integrated sensors or sensors
with specific local electronics, allow the
3 Basic sensor function (static or dynamic) accuracy inherent
in a sensor to be exploited by means of
(generally digital) microelectronics to
a far greater degree than conventional
sensors. The sensor information, in par-
Physical/ Electrical ticularly the complex information from
Sensor multisensor structures, can be com-
chemical output pressed here by local further processing,
quantity º signal E
that is to say, it is brought to a higher
(non-electrical)
level (than the sensor alone is capable
Disturbances Yi (temperature, of providing), without a multiplicity of
LAE0913E
power-supply fluctuations ...) external connections being required

for this.
146 Automotive sensors | Basics and overview
There is no clear determination as to Saving the correction parameters

whether sensors can already incorporate The host computer calculates the model
a part of the signal processing or not; it is, parameters for the specific sensor from
however, recommended, for example, that the data previously obtained, e.g. for a lin-
no distinction be made between elemen- ear characteristic curve course and saves
tary sensor, sensor cell, or the like and these to the smart sensor’s PROM. In a
an integrated sensor. check process, these parameters may also
first be emulated in the host computer’s
Calibration process RAM before they are finally burnt to the
A smart sensor is generally programmed nonvolatile memory of the smart sensor.
or calibrated, in a procedure correspond- If characteristic curves are approximated
ing to the calibration of conventional ana- with higher order polynomials, program
log sensors, in three steps with the help of maps (look-up tables) may also be stored
an external computer (host) (Fig. 4): in the smart sensor to avoid protracted
computing processes. The solution of sav-
Recording the actual value ing a coarse-grid program map in conjunc-
The host computer systematically varies tion with a simple linear interpolation
both the measured variable xe and the between the data points (example shown
influencing variable(s) y, thus setting a in Fig. 5) has also proved very useful.
specific number of relevant and represen-
tative operating points. The smart sensor Operating phase
here passes the, still uncorrected, raw sig- The smart sensor is now disconnected
nals xa* to it. The host, however, is simulta- from the host computer and is in a position
neously also receiving the “true” values xe itself to calculate the measured variable xe
and y from substantially more accurate with very low error thanks to the stored
reference sensors. The host computes the model data. It is able to pass the value to a
correction values necessary by comparing connected control unit in digital, bit-serial
these two values, and interpolates these or even analog form (e.g. pulse-width
corrections over the full measuring range. modulated), for instance. The measured
variable can also be distributed digitally
to further control units via a bus interface.
4 Trimming/calibration of a smart sensor
1.) Actual xe Smart

value Measured variable xa*
y sensor
recording Influencing variables External
PROM
calibration
Reference xe computer
sensors y
Systematic variation
of measurement and
influencing variables
2.) Saving Smart External
correction sensor calibration
parameters PROM computer
xe Smart
SAE1014-1E
3.) Operating Operation xe Control

sensor
phase values unit
y PROM
Automotive sensors | "VUPNPUJWFBQQMJDBUJPOT 147
This process can, unlike conventional laser Automotive applications

calibration, in principle also be repeated if
an erasable PROM is used. This is a partic- With increasing demands on all functions
ular advantage in the development phase in the vehicle, the control and regulation
of sensors. functions that were previously imple-
mented by mechanical means have, in the
last 40 years, successively been replaced
by electronic devices (ECU, Electronic
5 Interpolation of measured values using a
data point program map Control Unit). This has necessarily caused
high demand for sensors and actuators
Interpolation with which these electronic control units
Measuring travel s
m = 0, …, 32 value s (T,q)
q can, on the one hand determine the rele-
ure b vant states of the vehicle, and on the other
e ratq s2 s4
mp q m+1
hand, actually influence these states.
Te
qm
dq s1 Over this period, the vehicle industry has
a s3
become one of the, previously unprece-
Dq
Pe dented, drivers of the development of sen-
ri od
du sors that could be manufactured in large
rat
ion DT numbers.
UAE1044E
T dT
Tn Whereas at the outset they still were of
n = 0, …, 64 T electromechanical or macromechanical
Tn+1
form of some sort, the trend beginning
Example: two-dimensional data point- in the 80s was for miniaturization, with
program map s (Tn, qm) of a smart sen-
sor for the measurement of a travel s:
6 Milestones in the development of sensors for
automotive applications
For a highly precise evaluation of the
sensor, which operates using variable
1950 Lambda oxygen sensor
inductance, its natural characteristic
curve and its temperature sensitivity 1960 Electromechanical pressure sensor
are each approximated using 5th order Piezoelectric knock sensor
polynomials. As a frequency-determin-
ing element of a very simple oscillator 1970 First integrated Hall sensor
circuit, it outputs the period duration T Strain-gage acceleration sensor
as an uncorrected output signal. As a for airbag
sensor model for the measuring travel s, First pressure sensor on silicone base
an outline program map embracing only
32 x 64 = 2,048 example-specific values 1980 Hot-wire air mass meter
sn,m is stored (in the PROM) along with Thick-film air mass meter
a simple interpolation algorithm (in the Integrated pressure sensor
ROM), instead of a total of 36 poly-
nomial coefficents and a protracted 1990 Micromechanical acceleration
polynomial evaluation. Should a signal T sensor for airbag
occur between these data points Tn and Piezoelectric yaw-rate sensor
Tn+1 in conjunction with a temperature q for ESP
between data points qm and qm+1, Micromechanical air-mass meter
the sensor interpolates in two dimen-
Micromechanical yaw-rate sensor
sions in accordance with the mapping
between the benchmark figures stored 2000 Yaw-rate sensor for roll-over sensing
UAE1045E
“error-free” s1,...s4 and thus the mea-

sured value s(T, q) sought is determined
as the result of interpolation.
148 Automotive sensors | Automotive applications
semiconductor methods (batch process- gallium arsenide (GaAs) have a substan-

ing) for the manufacture of high use sen- tially greater operating temperature range
sors. For a while, sensors originating in than silicon, which can be very advanta-
hybrid technology using thick-film tech- geous, particularly at some points in the
niques also played a not insubstantial role. vehicle. Thin metallic layers are very
These are still used in isolated applications well suited to the manufacture of precise
today, e.g. in the wafer-shaped Lambda strain-gage resistors, accurate tempera-
oxygen sensors and high temperature ture sensors and magnetic field-dependent
sensors for the exhaust line. resistors.
With silicon it is possible also to inte-
Where temperature and magnetic field grate the electronics monolithically with
sensors could initially still be designed in the sensor. This technique has lost its im-
circuit-like structures and be produced in portance – with a few exceptions (e.g. Hall-
batches, this trend increased as it became effect ICs) – because of the generally very
possible also to micromachine silicon in different number and type of process steps
many various ways in two or three dimen- and also because of the inflexibility associ-
sions, and also to connect them, even in ated with this. Hybrid integration technol-
multiple layers, very soundly by very ogies in the tightest of spaces as a rule lead
efficient methods. to substantially more cost effective, but
Where the technologies of the electronic functionally equivalent solutions (Fig. 7).
semiconductor circuits were practically
exclusively based on silicon as the base While the development of sensors was ini-
material, quite different materials and tially concentrated almost exclusively on
technologies played a not insignificant systems inside the detail in the drivetrain,
role in sensors. Thus, for instance, quartz the suspension and the body and driving
can also be micromachined using anisotro- safety, the direction of sensing of newer
pic etching techniques, and unlike silicon developments is increasingly named to the
also has very advantageous piezoelectric outside and to the area close to and further
properties. III-V semiconductors such as from the vehicle:
7 Hybrid integration of sensor and electronics: surface micromechanical acceleration sensor

on a microhybrid circuit
Bonding wire Sensor chip Suspension spring Seismic mass with

movable electrodes
Evaluation
circuit
UAE0787E
200 m Fixed electrodes

Automotive sensors | Automotive applications 149
• Ultrasound sensors detect obstacles on 8 Sensors in the multilayered process that is the
parking and will even allow automatic motor vehicle
parking in the foreseeable future, per-

haps in combination with other sensors. SE/ECU SA
• Near-range radar scans the area around
the vehicle to detect objects that proba- ECU
Fig. 8
bly could cause a collision, to gain time 1 Measuring sensors
and to prime safety systems before a E
2 Adapter circuit
collision occurs (precrash sensors). 2 Yi AZ 3 3 Driver circuit
• Imaging sensors can not only detect 4 Actuators
Y1 4
traffic signs and send them to the driv- 1 AK Actuator
SE AK AZ Display
er’s display, but also detect the edge
SA Operating switch
of the carriageway, warn the driver of
Φ SE Sensors
UAE0288-1Y
any hazardous deviation and, where ECU Control unit
required, in the long term also permit F Physical variable
automatic driving. In combination with E Electrical variable
infrared beams and a screen in the driv- Yi Disturbances
er’s field of vision, IR-sensitive imaging

sensors could permit long distance These customer-specific adapter circuits
observation of the carriageway, even are tailor-made for specific sensors and
at night (night vision) or in foggy con- are available in integrated design and in
ditions. a wide variety of versions. They represent
• Long-range radar sensors observe the a quite substantial and very valuable addi-
carriageway for 150 m in front of the tion to the sensors described here, without
vehicle, even in poor visibility, to adapt which their use would not be possible, and
the driving speed to vehicles ahead and, the measuring accuracy of which is, prop-
in the longer term, also to support auto- erly said, only defined in conjunction with
matic driving. these.
The vehicle can be regarded as a highly
As part of the vehicle’s peripheral equip- complex process, or control loop, which
ment, the sensors and actuators form the can be influenced by the sensor informa-
vehicle’s interface to its complex drive, tion from other processing units (control
braking, chassis, and bodywork functions, units), as well as from the driver using his/
as well as to the vehicle guidance and navi- her controls. Display units keep the driver
gation functions and the (usually digital) informed about the status and the process
ECUs which operate as the processing as a whole.
units (Fig. 8). An adapter circuit is gener-
ally used to convert the sensor’s signals
into the standardized form (measuring
chain, measured-data registration system)
required by the control unit.
150 Automotive sensors | Details of the sensor market
Details of the sensor market Micromechanical processes for the pro-

duction of sensors are continually being
The proportion of value added in the vehi- further developed in tandem with the
cle by the electrical and electronics sys- semiconductor processes. These compa-
tems is currently around 26 %. Nowadays, nies do not, however, have any specific
practically every second sensor is fitted know-how in the area of specification,
into a vehicle, with annual growth rates testing and packaging appropriately
that are, in some cases, still in double for the vehicle.
digits. Since the start of the 90s, microme- • Special, generally medium-sized manu-
chanical and microsystem-technical sen- facturers of sensors, who do not produce
sors have taken a rapidly increasing share, semiconductor circuits, but have gener-
which amounted to about a third in 2005. ally chosen just a few sensor types as
Unlike in the general sensor market, their product to supply the sensor mar-
Europe has a current market share of ket overall or just preferred branches
41 % of the vehicle sensors sector and such as the vehicle market.
Bosch is the world market leader; this has • Large vehicle suppliers and system man-
clearly surpassed America with a share ufacturers (e.g. Bosch) or large subsid-
of just 34 %. The total market for sensors iaries of vehicle manufacturer who have
for vehicle applications is expected specialized in the needs and support of
to grow from US$ 8.88 billion in 2005 to their mother groups. Here too, since the
around US$ 11.35 billion in 2010, that is introduction of electronics in the vehi-
by a total of 28 % (Fig. 9). cle, experience has been gained with the
production of semiconductor and hybrid
There are typically three groups of compa- circuits in close collaboration with semi-
nies offering sensors for use in the vehicle: conductor manufacturers (process de-
• The semiconductor industry: sensors velopment, licensing). Thanks to system
from semiconductor manufacturing knowledge, it has been possible to build
originated here with the application up comprehensive know-how in the
of a few process steps. They serve the field of specification, testing and pack-
entire sensor market including the auto- aging techniques appropriate for the
motive industry and have an efficiently- vehicle.
operating sales and distribution system.
9 Sensor market for automotive applications
Billion
US $
12 11.0511.35
10.67
10.24
10 9.66
8.88
8.22
8 7.51
7.21
6.71
6
UAE1046E
4
2001 2005 2010
Fig. 9
Source: Bosch
Automotive sensors | Features of vehicle sensors 151
Features of vehicle sensors If no sensor principles are known for the

requirements set, research will be started
While general sensors are being developed into new procedures and methods for the
for as broad a user group as possible and measurement of the variables required.
in multilevel measuring ranges, often with- This phase is one of fundamental research
out the manufacturer being aware of the which ultimately also supplies new initial
application, vehicle sensors are generally technology prototypes. This procedure
specified and optimized for a special ap- can repeat itself recursively until a prom-
plication. They are a part of a system and ising solution which can then be sent to
are often not freely available in the trade. product development is found. It is not
Their development generally takes longer unusual for this development loop to be
than that of commercial sensors, not just repeated again in full or for a different
because of the enhanced demands. It is measured variable to be selected.
more associated with the development
of the system and generally lasts as long Five distinct phases are generally identi-
as this does, as the sensor specification fied in the development of a completely
can still change until system development new sensor, just as with other electronic
is concluded. products (Table 2). While prototypes and
A-samples mostly still come from advance
The high thrust for innovation in systems development or research, B- and C-samples
for the automobile sector very often com-
pels the development of new sensor tech-
nologies or significant extensions of their 10 Genesis of a functional Bosch sensor
specifications. Figure 10 shows the typical

development phases that vehicle sensors
pass through at the supplier’s works. Implementation phase Design phase
The development process begins natu-
Technology issues
rally with the system idea originating with Application issues
the vehicle manufacturer or the supplier.

Technical Technical
The initial task is to make a selection of the assistance clarification
measured variables required, whether or Research
not this is feasible. At this stage, the sen- (Technology pool) Process engineering
sors’ function is simulated as part of the New methods
Key problems
systems engineers’ usual system simula-
Advance development
tion and a first specification is drawn up.
Prototype, A-sample Sensor principle
If it is possible to use a sensor technology Laboratory test/ Sensor technology
that has already been introduced, the measuring technique Packaging
wishes in respect of the sensors are passed Product development
directly to product development or the Vehicle test, trial System definition
B-sample/C-sample Measurement selection
producing area. If no technology is imme- Sensor target
Production model
diately available, sensor and technology specifications
experts are gradually increasingly in- Product specification Vehicle parameters
Customer
cluded in research and advanced develop-
Customer specification
ment. Here, the first laboratory prototypes Product:
can be created using known technologies Vehicle system/ Customer requirements:
Components Vehicle system
and often with the help of external part-
SAE1012-1E
ners. These prototypes can then be passed

to product development for initial tests.
152 Automotive sensors | Features of vehicle sensors
are products of product development. Sensor classification

If recursions are necessary in difficult
developments, there can easily also be a Sensors can be classified and grouped
number of B- or C-sample phases (B1, B2, according to very different points of view.
C1, C2). They can be classified with regard to their
use in the vehicle as follows:
Assignment and application

• Functional sensors (pressure, air-mass
flow) mainly used for open and closed-
loop control assignments
2 Sensor prototyping phases leading to series • Sensors for safety (passenger protec-
production
tion: airbag, ESP) and security (theft-
Prototype Function/ Manufacture deterrence feature)
phase specifications
• Sensors for vehicle monitoring
Prototype restricted Prototype con- (on-board diagnosis (OBD), fuel-con-
struction without
sumption and wear parameters) and
tools
for driver/passenger information
A restricted Prototype con-
struction without
Characteristic curve type
tools
• Continuous linear curves (Fig. 11a) are
B (possibly full Prototype con-
B1, B2) struction without
used mainly for control assignments
tools (construc- covering a wide measuring range.
tion identical to C) Linear curves are also distinguished by
C (possibly full Prototype con- uncomplicated testing and calibration.
C1, C2) struction with • Continuous nonlinear curves (Fig. 11b)
series production are often used for the closed-loop
tools control of a measured variable across
D full Pilot series, par- a very restricted measuring range
tially manual (e.g. exhaust-gas control to l = 1, vehicle
Series produc- full Automated spring-deflection level). When, for in-
Table 2
tion stance, a constant permissible deviation
relative to the measured value is de-
11 Characteristic curve types manded throughout the complete mea-
suring range (e.g. HFM air-mass meter),
curves which feature both pronounced
a b non-linearity and a special shape
S S (e.g. logarithmic) are at an advantage.
Fig. 11
• Such two-step curves (possibly even
S Output signal
X Measured variable featuring hysteresis, Fig. 11d) are used
for limit-value monitoring in such cases
X X
a Continuous, c d where remedial measures are easy to
linear apply when the limits are reached.
b Continuous,
S S
If remedial measures are more difficult,
nonlinear
then multi-step curves (Fig. 11c) can be
c Discontinuous,
used for an earlier warning.
UAE0286-1Y
multi-step
d Discontinuous,
X X
two-step (with
hysteresis)
Automotive sensors | Sensor classification 153
Type of output signal For instance, the signal is bound to be

Sensors may also be distinguished by the discontinuous if it is digital and issued
type of output signals (Fig. 13): in bit-serial form.
Analog signals
• Current/voltage, or a corresponding
amplitude
• Frequency/period duration
• Pulse duration/pulse duty factor 13 Signal shapes
Discrete output signal

• Two-step (binary coded). a f
• Multi-step, with irregular steps (analog U
coded). U
• Multi-step equidistant (analog or digital
f
coded).
• A further distinction must be made, as is Fig. 13
t a Output signal U,
represented in Figure 12 in a systematic
information
overview of the determined, i.e. nonran- b Tp
U parameter:
dom (stochastic) signal – as to whether frequency f
UAE0287-1Y
the signal is permanently available at U b Output signal U,
the sensor output (continuously) or only Tp
information
at discrete intervals (discontinuously). t parameter:
pulse duration TP
12 Classification of the determined signals according to the information parameter (IP) with examples
Infinite Finite
Deterministic signal
range range
discrete
analog
Multiple signals Digital signals
Two-point signals Three-point signals Remaining Constant

(binary signals) (ternary signals) multiple signals quantific-
ation level
conti- disconti- conti- disconti- conti- disconti- conti- disconti- disconti-

nuous nuous nuous nuous nuous nuous nuous nuous nuous
1 2 3 4 5 6 7 8
1 IP 2 IP 3 IP 4 IP
L L
0 0
Time t Time t Time t Time t
5 IP 6 IP 7 IP 8 IP Clock1 Clock2 Clock3

L
0
UAE1047E
–L
Time t Time t Time t Time t
154 Automotive sensors | Error types and tolerance requirements
Error types and ing), related to the measuring-range final

tolerance requirements value it becomes the percentage error of
the range.
The deviation of the actual characteristic Assuming a generally desirable linear
curve of a sensor from its reference char- characteristic, the absolute deviation eabs
acteristic curve is designated its error e. can be classified into three categories
It is best stated in relation to the input (Fig. 15):
range y (measured variable) and not to • Zero offset (offset error) ezero
the output range x (output signal): • Gradient deviation (gain error) egain
(3) e = yindicated - ytrue • Linearity deviation elin
yindicated = indicated value for the mea-
sured variable The causes of these errors are predomi-
ytrue = ideal value, setpoint value for nantly found in
the measured variable (is determined • The production scatter of the character-
with a measuring sensor that is more istic curve
accurate than the sensor being exam-
ined by at least 1 class)
15 Subdivision of the total error
The amount of the deviation represents,

as shown in Figure 14, the absolute a y
error eabs (unit as measured variable).
Related to the true measured value ytrue,
yi ezero
this becomes the relative error (% of read-
Dys
14 Characteristic curves and error graph for a sensor ys
Dx
a y eabs x0 x
b y
yi egain (x)
ic
ist
Fig. 14 er
ct
ra
a Actual and yi
c ha
reference ys
al
tic
tu
characteristic
ris Dys
Ac
Dys
b Error graph
racte
a ys
ch
y Measured variable nce
e Dx
fer
x Output signal Re Dx
Dx Measuring range
x0 x
e Error (deviation) x0 x c y
elin (x)
b eabs
eabs = yi – ys
Fig. 15 yi
a Zero error eabs
b Gradient error
raph Dys
c Linearity error rg
rro
E ys
y Measured variable
UAE1048E
UAE1049E
Dx
x Output signal
Dx Measuring range x0 x x0 x
e Error
Automotive sensors | 3FMJBCJMJUZ 155
• The temperature sensitivity of the char- Reliability

acteristic curve and
• The production scatter of the tempera- Failure rate
ture sensitivity The operational reliability of a sensor is
a purely statistical quantity and, as with
The deviations referred to are exclusively any component, is characterized by its
systematic or terministic errors which, by failure rate l, which is stated in 1/h, %/h
contrast with random (stochastic) errors, or ppm/h. Here l is determined with a
are well defined, foreseeable and, as a mat- large number of parts. If it is wished to
ter of principle, correctable, and are also determine approximately the failure rate
for the most part more or less precisely with a number N (< 40) of sensors that is
corrected. not too great, it is possible to observe the
The non-correctable, stochastic errors failure behavior of this random sample
include, for example, under operating conditions until, after a
• Drift (low and higher frequency back- finite time, all the parts have failed. If this
ground noise) and observation is begun at time t0 and if the
• The effects of aging inventory of parts remaining at a later
point ti is identified as B(ti), then a good
In the specification of a sensor, the total approximation for the failure rate l is the
error in the new state and after aging is failure quota q:
prescribed in the product specifications
B(ti) – B(ti+1)
by means of a tolerance graph (Fig. 16). (6) q(Dti, ti) = –––––––––––––––––– Failure quota
Dti · B(ti)
Sometimes, however the permissible indi-
vidual error shares such as the offset, where Dti = ti+1 - ti
lead and linear deviations are additionally
specified. ti are the times at which individual parts
or several parts fail (Fig. 17). The ratio of
Strict metrology teaching states that, the current inventory of parts to the initial
in the case of systematic errors, the sum inventory of parts is known as the relative
of the amounts of the individual errors inventory BR:
must be assumed as the total error (these
may add up in the worst case). In the case
of stochastic errors, statistical addition,
which calculates the total error as the root 16 Tolerance graph for a sensor
of the sum of squares of the individual er-

rors, is permitted. As statistical addition K
leads to a smaller total error, it is, however, 4
often also applied to the systematic errors 3
Class B
in a less strict configuration: 2
n 1
(4) etotal = o|e | i Sum of n systematic 0
Class A
DT
1 errors –1
!w
n
oe
–2
(5) etotal = i
2 Sum of n stochastic
–3
1 errors
–4 Fig. 16
(statistical addition) Tolerance graph
UAE1050E
–5
–200 0 650 850 C illustrated on the
Temperature T example of a resistance
temperature sensor
156 Automotive sensors | Reliability
Furthermore, it is absolutely necessary to

B(ti)
(7) BR(ti, t0) = –––––––– specify under what operating conditions
B(t0)
the failure rate defined in this way is to
With a very large number of parts (N → ∞), be understood. Particularly in the case
the survival probability R(t) at the, here of electrical components such as sensors,
now continuously variable, time t corre- for example, it is necessary to distinguish
sponds to this. The failure rate is calcu- between true, active operating time
lated here for a large number (in practice (switched on state) and service life in the
N ≈ 2,000) of sensors as a percentage varia- sense of pure storage time. Any statement
tion on the survival probability R per unit of a failure rate without this additional in-
of time dt as: formation is worthless.
Failure rates are normally determined
1 dR
(8) l(t) = — –––––– · –––––– Failure rate using time-acceleration methods. The ac-
R(t) dt
celeration factors are achieved here by ex-
posing the sensors to enhanced operating
Reliability is the inverse value of the conditions. A high degree of experience is
failure rate: required in order to apply time accelera-
tion methods that truly reflect reality.
1
(9) z = –––––– Reliability
l
The concept of the mean service life TM is
also used to characterize the reliability of
The definition of the failure rate l a sensor. A good approximation of this can
demands a failure criterion: be calculated from a random sample from
• Complete failure the sum of the individual service lives Ti:
• Partial failure N
• Sudden failure (sudden change in a
feature)
N
1
(10) TM ≈ ––––– · oT1
i or
• Drift failure (gradual change in a ∞

feature) (11) TM = eR(t) dt
0
for a very large
number of parts
17 Inventory and relative inventory of a random 18 Distribution of the failure rate l(t) over time
sample
20 DBi = Bi – Bi+1 Ai = 1– BR (ti) 100 Early Random Ageing

18 % failures failures and
Bi Cumulative
16 Bi+1 failure 80 wear
frequency effects
14
Relative inventory
BR (ti) = B(ti)/ B(t0)
12 60
Inventory B(t)
10
Failure rate
8 40
6 Increasing
20 reliability
4
Fig. 17
2
0 0
Observation for a
0 2,000 4,000 6,000 8,000 10,000 h
random sample of
UAE1051E
UAE1052E
Dti Service life

N = 20 sensors;
ti ti+1
mean service Time
life TM = 4,965 h.
Automotive sensors | Reliability 157
The failure rate of a product against time chemical and electrical ambient influences
shows a typical bathtub pattern (Fig. 18). must be provided (packaging, passivation).
At the start, the failure rate is somewhat On the other hand, it is costly to create op-
high because of early failures, it then erational reliability by testing alone, i.e. to
passes through a longer, relatively low, exclude early failures through premature
horizontal range, then to rise again drasti- aging.
cally towards the end of the service life It is sensible to provide monitoring and
(aging and wear effects). In the case of sen- diagnostic capabilities for sensors in com-
sors for which high operational reliability plex systems (e.g. signal range check, etc.),
is demanded, an attempt is made to avoid so that any failures occurring may be de-
the elevated failure rate at the start by tected in good time. In an emergency, a
sorting out the early failures by pre-aging, sensor’s function can here be temporarily
e.g. by conditioning at higher tempera- replaced by other measured variables or
tures (“burn in”). Early failures are basi- sensible fixed values (limp-home mode,
cally nothing other than manufacturing back up). A sensor-less, often purely me-
defects that have not been recognized. chanical, limp-home mode can be pro-
Table 3 gives some examples of permis- vided. For instance, should the accelera-
sible failure rates l applicable for the tor-pedal sensor failure in a diesel vehicle,
vehicle. The ppm values given relate to the vehicle could also be driven slowly
a period of 10 years, or alternatively to back for repair under regulated idle speed
150,000 km covered, if this is completed alone (limp home).
in a shorter time. Where the sensors are Where reliability must be guaranteed
given a flat rate of <10 ppm, it means that with a probability bordering on certainty
in 10 years, only fewer than 10 of 1 million (e.g. for sensors in electronic braking and
sensors may fail. This value is, however, steering systems), redundancy, i.e. provi-
substantially lower for sensors in passen- sion of multiple parts, is generally the
ger protection systems. means used. Dual redundancy of a sensor
of the same type here only permits failure
Measures for increasing reliability detection for the case of widely different
The best method for ensuring high quality indications, whereas triple redundancy
is to engineer and design in reliability. with 2 from 3 analysis additionally still
This means that as soon as the sensor is supplies a correct measured value. Here,
designed, correspondingly durable mate- however, care must be taken to provide not
rials must be selected and solid protective only redundancy of the sensors, but other
measures against anticipated mechanical, essential parts such as the power supply,
signal evaluation and transmission means
must also be correspondingly redundant,
3 Reliability requirements on vehicle systems as otherwise the probability of simultane-
Warranty target: 150,000 km/10 years ous failure increases. It is often also advis-
→ ECU failure rate (field) < 50 ppm able to provide sensor redundancy in dif-
→ ECU failure rate (0 km) < 15 ppm ferent technologies.
→ Failure rate of modules < 10 ppm
and sensors
→ ASIC failure rate < 3 ppm
→ IC failure rate << 1 ppm
→ Failure rate of discrete < 0.5 ppm
components
→ For comparison ~ 5,000 ppm
Mobile phone Table 3
158 Automotive sensors | Main requirements, trends
Main requirements, trends sensors simultaneously. For example,

semiconductor sensors are manufactured
The vehicle sensors customized for the re- using “batch processing” in which 100 to
quirements of special electronic systems 1,000 sensors are manufactured simulta-
in the vehicle, unlike the universal sensors neously on a single silicon wafer. On the
conventionally found in the market, are other hand, such manufacturing equip-
subject to five strict requirements (Fig. 19) ment is only an economic proposition
which must be fulfilled by development when correspondingly large numbers of
and to which the most important develop- sensors are produced. These quantities
ment trends also correspond. sometimes exceed an automotive-industry
supplier’s own in-house requirements,
Low manufacturing costs and can typically often be between 1 and
Electronic systems in modern vehicles eas- 10 million per year. Here, the high num-
ily contain up to 150 sensors. Compared bers of sensors needed by the automotive
to other sectors of sensor application, industry has played an unprecedented and
this abundance forces a radical reduction revolutionary role, and set completely new
in costs. The target costs – typically in the standards.
range of € 1 to 30 – are here often less than
one-hundredth of those for conventional High reliability
sensors of the same capacity. Naturally, In accordance with their assignments,
particularly with the introduction of a new vehicle sensors are subdivided into the
technique or technology, the costs are gen- following reliability classes, given in
erally on a falling learning curve, starting descending order of severity:
from a higher level. • Steering, brake, passenger protection
• Engine/drivetrain, chassis/tires
Development trends • Comfort and convenience, diagnosis,
Automated manufacturing processes information, and theft deterrence

(Fig. 20) are largely used, working to a high
yield. This means that each process step
is always executed for a large number of
19 Principal requirements on vehicle sensors 20 Relationship of costs, markets, technologies
Vehicle sensor Requirement

1,000
Rational mass
Low costs
Sensor costs per unit
production 100
Robust, proven High reliability 10
technology
1
Appropriate pack- Extremely severe
aging technology operating conditions 0.1
Production 10 100 1,000 100,000 10 million
Appropriate (units) 10,000 1 million 100 million
Low space per year
miniaturization
technologies
requirement Manufacturing manual semi-auto- fully automated mono-
technologies work mated discrete hybrid lithic
SAE1013-1E
On-site error
UAE1053E
High-level accuracy Market Traditional markets New markets

compensation classification
Automotive sensors | Main requirements, trends 159
The requirements on the highest class Severe operating conditions

here easily match the high reliability val- Like practically no other breed, vehicle
ues familiar from aerospace applications. sensors are exposed to extreme stresses
To some extent, they demand similar mea- because of where they are fitted, and must
sures, such as the use of the best materials, withstand all kinds of aggression:
redundant componentry, self-monitoring, • Mechanical (vibration, shock)
(short-term) backup power supply, multi- • Climatic (temperature, moisture)
ple programming of critical decision-mak- • Chemical (e.g. spray, salt mist, fuel,
ing algorithms. engine oil, electrolyte)

• Electromagnetic (incident radiation,
Development trends line-conducted interference, excess

Appropriate design measures guarantee voltages, polarity reversal)
built-in reliability. For instance, this neces-
sitates the use of reliable, top-quality com- It is this trend of using sensors immedi-
ponents and materials, coupled with rug- ately at the site where the measurement
ged and well-proven techniques and engi- is taken in order to exploit the benefits
neering. And furthermore, an effort is associated with this which itself leads
made to achieve a consistent integration to the requirements being considerably
of the systems to avoid connection points cranked up.
which may be broken and are at risk of Figure 21 illustrates this problem on
failure. This is possible, for instance, a sensor which might be a temperature
with radio-scanned sensors based on the sensor, speed sensor, flow-rate sensor
antenna-coupled SAW (Surface Acoustic or concentration sensor, for instance.
Wave) elements which do without wiring The sensor cannot in the slightest always
completely. If necessary, redundant sensor be surrounded by a hermetic protective
systems are fitted. sleeve. Although this sleeve could offer
rough protection for the measurement of
temperature, flow rates and concentration,
it must permit more or less direct contact
between the sensor and the generally very
aggressive monitored medium (an excep-
tion to this is the inertia sensor). Some-
times, thin, but very resistant, passivation
21 Technologies accompanying sensors layers are permissible on the sensor.
The form of the plug-in sensor demands
Monitored External environment a permanently tight seat of the mount in
medium
5 6 7 the associated wall (inner packaging, outer
packaging). The sensor may be connected
Fig. 21
to the electronic control units via a fixed
1 Sensor
housing plug or also, as illustrated, via a 2 Protective sleeve
1 8 plug flexibly fastened with a length of ca- (coating)
ble (e.g. ABS wheel-speed sensor). Here, 3 Seal
too, there are three critical electrical con- 4 Mount
5 Seal, mounting
nection points that absolutely must be pro-
6 Data point
tected against conductive fluids and corro-
7 Seal, strain relief
sion: the inner connection on the sensor 8 Insulation (flexible)
UAE1054E
element, the connection of the cable tail 9 Contacts

2 3 4 9 10 11
and finally the connection of the partially 10 Plug housing
shielded cable to the external plug connec- 11 Plug contact
tion. If the seal is inadequate, it is only a Fiber-optic sensors in which the light
question of time until corrosive liquids guided in the optical fibers (glass, plastic)
penetrate from the external connector can be modified as a function of the mea-
into the internal sensor terminal. sured variable are particularly immune to
The plug-in connection itself must be electromagnetic interference. This is true
sufficiently tight overall, that no shunts up to the point at which the optical signals
form at the sensor output. The cable itself are converted back into electrical signals.
must retain its flexibility and tightness If these are to be used in the future, there
after many years of operation under the will have to be some development work
most adverse conditions. the provision of low-priced measuring
Separable plug-in connections in the elements and the accompanying technolo-
vehicle unfortunately still represent one gies. There are, for example, some very
of the most frequent causes of failure. interesting applications for this in the
Wireless signal connections (e.g. infrared field of force measurement, e.g. in provid-
light or radio) could break down this prob- ing finger protection for electric power
lem in the longer view, especially if the windows and sliding sunroofs (Fig. 22).
sensors were even to have a wireless Sensors of this type have also been tested
energy supply (autonomous sensors). very successfully as very early responding,
The costs of these indispensable tech- distributed sensors in the door and frontal
nologies complementary to the sensors area of the vehicle for triggering passenger
often exceed those of the actual sensor and pedestrian protective systems.
element many times over. They constitute
the actual value of an vehicle sensor, not
only in terms of cost, but also in terms of
function. 22 Microbending effect
Development trends Microbending effect

Protective measures against the stresses F=0 F>0
Damping ≈ 0 Damping > 0
referred to demand a very high degree of
specific know-how in the field of sensor
packaging.
• Passivation and connecting techniques
• Sealing and joining techniques
• EMC protective measures Door frame
• Low-vibration installation
Sensor
• Service life test and simulation methods
SAE1015-1E
• The use of resistant materials, etc.

Sealing rubber
Furthermore, detailed knowledge of the
loading to which the sensor will be sub- Example of development: fiber-optic fin-
jected at the particular installation point ger protection in power windows based
on the microbending effect:
is required. It is often forgotten that the
The light in an optical fiber is attenu-
quality of a sensor stands and falls with ated proportionally to the corrugated
total competence with these protective distortion on the application of a force F
measures. transverse to the fiber (large measur-
ing effect), independently of where the
force is applied or whether it is acting
at a point or is distributed (distributed
sensors).
Low-volume design Micromechanical manufacturing refers

On the one side, the number of electronic on the one hand to dimensions in the µm-
systems in the vehicle continues to climb range and tolerances in the sub-µ-range
steadily. On the other, today’s vehicles are which cannot be achieved with conven-
becoming more and more compact. These tional machining methods. On the other
facts, together with the need to retain the hand, sensors which may have dimensions
high level of passenger-compartment com- in the mm-range, but which are manufac-
fort forces development to concentrate on tured with micromechanical methods, are
an extreme space-saving designs. Further- also considered micromechanical (Fig. 24).
more, the increasing demand for further We will discuss only the most well-known
improvements in fuel economy mean that and important method here, anisotropic
minimization of the vehicle’s weight is of etching of silicon. This is the most impor-
prime importance. tant method, because silicon is manufac-
tured to high precision and in large num-
Development trends bers at low cost, and is the most thor-
Widespread use is made of the familiar oughly researched and best known
technologies applied in circuit engineering material. In addition, it offers the capabil-
for the miniaturization of electronic com- ity of the monolithic integration of sensor
ponents: and evaluation electronics. The etching
• Film and hybrid technologies (deforma- speeds, which can be very different de-
tion-dependent resistors, thermistors, pending on the crystal axis (1:100), with
and magnetoresistors); nanotechology which suitable etching fluids such as
• Semiconductor technology (Hall-effect KOH attack are exploited in this method
and temperature sensors) (Fig. 25); some crystal faces, for instance,
• Surface and bulk micromechanical tech- remain practically untouched, while other
niques (silicon pressure, acceleration faces are rapidly etched to some depth.
and yaw-rate sensors) In conjunction with appropriate etching
• Microsystem technology (combination stop methods (doping, depletion layer)
of two and more microtechnologies practically three-dimensional structures
such as microelectronics and micro- of the smallest dimensions may be pro-
mechanics, Fig. 23) duced to the highest precision in this way
23 Microsystem technology 24 Micromechanics and mechanical structures
Packaging
Basic Advanced Micro- Nano-
mechanics mechanics mechanics mechanics
Micro-
mechanics: 1,000mm 100mm 10mm 1mm 0.1mm
– Sensor micro-
technology Traditional material machining
Micro-
– Actuator mi-
Microsys- electronics (turning, milling, etc.)
crotech-
nology tem tech-
nology Special machining methods
(e.g. extruder, laser, etc.)
Semiconductor technology methods
Micro-optics (lithography +
UAE1055E
etching/layer separation)
LAE0915E
Nanolithography
25 Test structure for the determination of the (Figs. 26, 27). The problem of simultane-
etching rate in Si monocrystal wafers ously integrating the circuit can be consid-
ered largely resolved, even if currently the
associated circuit for most sensors is still
made separately for reasons of a better
yield and higher flexibility.
By contrast with bulk micromechanics,

anisotropic, often wafer-depth etching has
no role to play in surface micromechanics
(SMM). SMM sensor structures are mostly
built up additively on the surface of a sili-
UAE1056Y
con substrate (Figs. 28 and 29). Even if
the dimensions of bulk silicon sensors
are still mostly in the mm-range, those of
SMM structures are an order of magnitude
26 Anisotropic etching of (100) silicon smaller (typically 100 µm), as a rule.
a Frequently, mechanically indispensable

Screen <111>
parts are used simultaneously for enclos-
ing and even cooling (associated) sensors
A A’ and electronics (e.g. microhybrid control
unit as an add-on control unit to the VP44
<110> diesel distributor-type injection pump).
<100> This fusion of mechanical and electronic
components known as mechatronics is
Screen coming more and more to the forefront
<100>
<111> in the search for cost and space savings.
b
Fig. 26 54.74 In the foreseeable future, practically all
UAE1057E
a Top view systems will operate on this basis.

b Cross-section
at AA'
27 First attempts at three-dimensional forming of silicon

UAE1058E
Fig. 27
Source:
Prof. Heuberger,
Fhg Berlin
Another example of a mechatronic sensor High accuracy

system is the Hall-effect speed sensor built By comparison with the probes and sen-
into the wheel bearing. The addition of sors used for instance in the processing
magnetic particles to the indispensable industry, with only a few exceptions (e.g.
shaft seal in the bearing can even give it the air-mass meter), the accuracy demands
the function of the rotor or a pole wheel. made on vehicle sensors are relatively
The sensor benefits from excellent protec- modest. Generally, the permissible devia-
tion and encapsulation along with the high tions are > 1 % of the measuring-range fi-
precision of the roller bearing and no lon- nal value. This applies in particular when
ger needs any bias magnets. Finally, the considering the unavoidable effects of
Hall-effect technology allows installation aging. The permissible deviations are nor-
in the very tight space within the bearing. mally achieved by the application of com-
plex techniques to compensate for manu-
facturing tolerances, and to balance the
effective compensation measures used
against interference. Particularly since
28 Surface micromechanical pressure sensor with the above-mentioned requirements have
piezoresistive pick-off
for the most part become achievable, con-
tinually more demanding and sophisti-
cated systems are imposing higher and
higher demands in this sector.
Development trends
Initially, a tightening up of the tolerances
in manufacture, and refinement of the cali-
bration and compensation techniques help
in this regard. An important step forward
here is the hybrid or monolithic integra- Fig. 28
tion of the sensor and signal electronics Source:
UAE1059Y
directly at the measuring point, even em- Prof. Guckel,

University of Michigan
bracing complex digital circuits such as
in Madison, USA
analog/digital converters and microcom-
puters (Fig. 30).
29 Surface micromechanical acceleration sensor with Such microsystems are also known as
capacitive pick-off
“intelligent sensors”. They take full advan-
tage of the sensor’s inherent accuracy,
and offer the following features:
• Reduction of the load on the control unit
• Uniform, flexible, bus-compatible inter-
face
Sensor element
• Multiple application of sensors
• Use of smaller measuring effects and
of radio frequency measuring effects
(local amplification and demodulation)
• Correction of sensor deviations at the
measuring point, and mutual calibration
UAE1060E
and compensation of sensor and elec-

tronics, is simplified and improved by
storing the individual correction infor- namic characteristics (evaluation of the

mation in the PROM differential equation which defines the
dynamic performance).
While simultaneously detecting and digi-
tizing interfering factors, intelligent sen- Local electronic circuitry necessitates the
sors can calculate the required measured use of multi-sensor structures which use a
variable practically without error by ap- number of identical sensors, or a number
plying the mathematical sensor model of different sensors, to register a variety
(Fig. 31). Here, the item-specific model of highly complex facts and locally reduce
parameters are defined in a preceding these to their basic information content,
process which is equivalent to the calibra- as necessary. This applies in particular to
tion as previously performed, and stored imaging sensors which in future will play
in a PROM integrated within the sensor. an ever increasing role in detecting the
In this manner, it is possible to consider- situation inside and outside the vehicle.
ably improve the sensor’s static and dy-
30 Sensor integration levels
Sensor(s) Transmission path Control unit

Susceptible to
SE interference
SA A
Without electronics (analog) D ECU
Resistant to
Multiple interference A
1st Integration level SE SA ECU
tap-off (analog) D
Fig. 30 Immune to
Bus-com-
SE Sensors 2nd Integration level SE SA A BS interference
BS ECU
D patible (digital)
SA Signal conditioning
A/D Analog-Digital Immune to
Bus-com-
SA A interference
UAE0037-3E
converter SE MC BS ECU
3rd Integration level D patible
ECU Control unit (digital)
MC Microcontroller
BS Bus interface
31 Correction model for a smart sensor
Faulty measure-
ment signal xa* C Fault-free measure-
A Digital xa par. xea ment signal
D correction ser. digital
Influencing y computer
variables signal
UAE0818-1E
D xea
Model parameter PROM
A analog
Automotive sensors | Overview of the physical effects for sensors 165
With a number of pressure sensors inte- Overview of the physical

grated into the smallest of spaces, it is pos- effects for sensors
sible to not only increase the reliability
of measurement, but also to reduce the, The sensors are not classified here into a
generally irregular, aging drift by applying systematic structure by measuring effect,
mean-value generation. Individual failures but rather by measured variable. For this
(outliers) can also be detected and elimi- reason, the most important physical ef-
nated. Sensors of this type are also known fects or measuring concepts exploitable
as soft sensors. If the individual sensor for measuring purposes will here only be
elements are designed for differing mea- reproduced in the form of a rough over-
suring ranges – and at the same time fea- view, which cannot be complete. Overlaps
ture high overload capabilities (e.g. capaci- between the various categories are un-
tive) – such a sensor can be used to con- avoidable. The table specifies principally
siderably extend the high-acccuracy the actual electrical effects; mechanical or
measuring range. Sensor structures of this fluid prior stages, such as strain springs
type were developed and tested years ago, (force), membranes (pressure), spring-
but have not been converted into mass systems (acceleration), vibration sys-
marketable products. tems (tuning forks) or even fixed and rotat-
ing turbine blades (flow) are not included.
4 Physical effects for sensors
Physical effect Example As at: Series

production1/
Development2
Resistive effects
(dependency of electrical resistance):
Influence of temperature on metallic and NTC and thin-film resistors for the measure- S
semiconductor materials ment of air and engine temperature
Length or angular proportionality of resistors Accelerator-pedal and throttle-valve sensor, S
(potentiometer sensors) fuel tank level
(In plane) tension or pressure dependency High-pressure sensors (e.g. common rail, ABS):
(piezoresistive): strain-gage resistors metal diaphragm, S
low pressure sensors (silicon diaphragm), S
Vertical pressure dependency (out of plane) force sensor D
Magnetic field dependency (magnetoresistive): Speed or delivery angle measurement in S/D Table 4
semiconductor (magnetoresistors), AMR3) thin diesel distributor injection pumps 1) Series production,
metal layers (e.g. NiFe, also in barber’s pole form),
at RB or
GMR4) sensors (nanolayers)
competitors,
Light dependency: semiconductor photoresistors Rain sensor, dirt sensor for headlight S
cleaning, automatic headlights, possible also
automatic low-beams already phased out
2) Development
Inductive effects (effects of Faraday’s law)
Induction voltage sensors (alternator): Wheel, camshaft, engine speed, S possibly also
movement in the magnetic field needle lift (injection nozzle) concluded and
Wiegand effect Speed of rotation D in reserve
Variation in inductance as a result of the Solenoid armature sensor D 3) AMR = Anisotrop
positional change of a ferromagnetic coil core Magneto Resistive
Variation in inductance as a result of field limiting Semi-differential short-circuiting ring sensor S 4) GMR = Giant
conductive elements (eddy current) (diesel pump load sensor) Magneto Resistive
166 Automotive sensors | Overview of the physical effects for sensors
Physical effect Example As at: Series

production1/
Development2
Variation of the degree of transformer coupling Full-differential short-circuiting ring sensor D
(by electrical or magnetic conductive elements)
Variation in inductance or of the degree of Load bolt (Hitchtronik), S
transformer coupling (by magneto-elastic braking force D
conductive elements)
Saturation core sensors (e.g. Foerster probe) Compass sensor S
Capacitive effects (influence)
Capacitance change as a result in change in plate Micromechanical acceleration sensors, S
distance and degree of cover e.g. for airbag, ESP,
MM2 yaw-rate sensor, S
pressure sensor D
Capacitance change as a result in a change of the Oil grade, S
relative dielectric constant humidity sensors S
Capacitance change as a result of change in the Fuel tank level D
electrolyte level with a dielectric medium
Charge generating effects
Piezoelectric effect (quartz, piezoceramics) Knock sensor, airbag sensor, S
DRS1 yaw-rate sensor S
Pyroelectric effect IR sensor (dynamic) D
Photoelectric charge generation CCD and CMOS imaging sensor D
(also IR range)
Voltage generating, galvanic effects
Hall effect (out-of-plane sensitivity, Hall switch (ignition), wheel and engine S
semiconductor material) speed, acceleration sensor (ABS, 2g),
front passenger weight detection (iBolt™),
ARS1,2 (accelerator pedal etc.)
Pseudo-Hall effect LWS2 and LSW4 steering-angle sensors S
(in-plane sensitivity, metal thin film)
Electrolytic diffusion probes Lambda oxygen sensors S
(doped Zr oxide ceramic)
Thermocouple, thermopile IR sensor (Bolometer) D
Photoelectric and fiber-optic effects
Photoelectric cells, photodiodes, phototransistors Rain sensor, dirt sensor for headlight S
(also in IR range) cleaning, automatic headlights,
automatic low beams
Medium-dependent absorption Soot particles, humidity D
Extrinsic and intrinsic fiber-optic effects: influence Finger protection (windows, sliding sunroof), D
on intensity, interference (influencing phases), pedal force,
influencing polarization; e.g. microbending effect collision
Thermal effects
Resistor cooling as a function of the flow velocity, HFM air-mass sensors, S
1) Series production, of the medium, of the density or fill level of a analysis, –
medium concentration, –
at RB or
fill level (fuel tank) D
competitors,
Wave propagation effects
possible also
already phased out Sound waves: propagation time effects (echo Parking-aid assistant, S
sounding), superposition with medium speed, volume flow rate, D
2) Development
Doppler effect (moving source/receiver) speed over the ground S
possibly also
Light waves: total reflection on boundary surfaces, Rain sensor, S
concluded and
optical resonators (color analysis) fiber-optic finger protection, D
in reserve fill level (analog, limit value), D
3) AMR = Anisotrop Sagnac effect yaw rate: fiber and laser gyroscopes, D
Magneto Resistive Propagation time lidar (light wave radar) D
4) GMR = Giant Electromagnetic radiation: Doppler, FMCW, ACC distance sensor S
Magneto Resistive propagation time radar
Automotive sensors | Overview and selection of sensor technologies 167
Overview and selection • Semiconductor technology (mono-

of sensor technologies crystalline or polycrystalline resistors,
depletion layers, charge-storage cells,
The various sensor technologies for ex- etc.), electronics for signal conditioning:
ploiting the measuring effects described Si (bipolar, CMOS, BICMOS, EEPROM),
are naturally closely bound up with the GaAs
measuring principle. They will first be • Micromechanics (silicon and other
listed here in a rough overview: materials, e.g. quartz, metal (LIGA tech-
• Wound or photolithographically pro- nology) etc.)
duced induction coils (with electrical • Piezoceramics
or magnetic conductors), eddy current • Piezofoil
and short-circuiting ring sensors • Isolating ceramics as a spring material
• Flux gate probes (Metglas etc.) for mag- (e.g. as a pressure-sensor diaphragm)
netic field measurement • Ceramic solid electrolytes (e.g. as
• Pulse jump sensors (Wiegand) Lambda oxygen sensor)
• Wire-wound (inductance-free) resistors • Quartz and other piezoelectric crystals
• Foil resistors (laminated onto plastic • Glass or plastic optical fibers or plates
substrates)
• Sintered ceramic resistors The matrix shown in Figure 32 indicates
• Thin- and thick-film techniques (espe- which sensor technologies should be taken
cially resistors and capacitances) for various requirements. Thus, for in-
stance, should the requirements be for
small size, high reliability and high accu-
racy, it is appropriate to take a thin-film
32 Selection of suitable sensor technologies technology, as long as this is available for
the measured variable desired.
Measurement sensitivity
Interference resistance
advantageous
of limited benefit
Environment
Reliability
Accuracy
Costs
Size
Semiconductors
m-Mechanics
Thick film
Thin film
FOS (fiber optic)
Piezoelectric
Eddy current/
short-circuiting ring
Flux gate
UAE1087E
Pulse jump (Wiegand)
Ceramic (spring material)

168 Sensor measuring principles | Position sensors
Sensor measuring principles
There is a great number of sensors at Position sensors

work in motor vehicles. They act as the
sensory organs of the vehicle and convert Characteristics
input variables into electrical signals. Position sensors record the widest variety
These signals are used in control and of one or multi-dimensional travel and an-
regulation functions by the control units gular positions (translational and rotary
in the engine-management, safety and variables) in a wide variety of areas. This
comfort and convenience systems. Vari- includes gaps, distances, displacements,
ous measuring concepts are applied, fill levels and even minor expansions –
depending on the task. anything that can be measured in meters
and angles.
1 Travel/angular positions as direct measured In this sector of applications, activities
variables have long since been directed at changing
Measured variable Measuring range over to proximity or non-contacting sen-
Throttle-valve position in the 90° sor principles. Such sensors are wear-free
gasoline engine and thus have a longer service life as well
Accelerator/brake-pedal position 30°
as being more reliable. The costs involved
Seat, headlamp, rear-view mirror
position
though often force vehicle manufacturers
Control-rack travel and position for 21 mm to retain contacting sensor principles
diesel in-line fuel-injection pump which still perform sufficiently well for
Angular setting of the injected 60°
many measuring tasks.
fuel-quantity actuator on the diesel
distributor-type injection pump Position variables are “extensive mea-
Fill level in the fuel tank 20 to 50 cm sured variables”, the measured variable
Travel of clutch servo unit 50 mm or the measuring range of which has
Distance: vehicle – vehicle or 150 m
a substantial influence on the size of the
vehicle – obstacle
Steering (wheel) angle ±2 · 360° sensor (by comparison, pressure and tem-
(± 2 revolutions) perature, for instance are “intensive mea-
Tilt angle 15° sured variables”). Sensors for extensive
Angle of direction of travel 360°
variables of medium and larger ranges
GPS (Global Positioning System) 360° geographical
latitude/longitude, (e.g. including flow and force sensors) are
geographical altitude thus initially less suited to miniaturization
Near-range distance 1.5 m on the one hand and hence for cost-effec-
(US parking-aid assistant)
Near-range radar (precrash) 10 m
tive mass production. As this is much less
External video 40 m true for angular-position sensors than for
Long-range and near-range 100 m travel sensors (the angle value does not
Table 1 infrared viewer depend on the radius or leg length), the
former are clearly preferred in the vehicle.
2 Travel/angular positions as indirect measured In the case of extensive variables with
variables
a large measuring range, however, the
Measured variable Measuring range extensive variable can often be first con-
Spring compression travel 25 cm verted into an intensive variable, which
(headlamp range, vehicle tilt)
can then be measured using microsensors.
Torsion angle (torque) 1 to 4°
Deflection of a sensor plate Thus flow measurement can be converted
30 to 90°
(volume flow) to flow density measurement (partial flow
Deflection of a spring-mass system 1 to 500 µm measurement, e.g. 1:100) where the flow
(acceleration)
tube becomes an essential component part
Deflection of pressure sensor 1 to 20 µm
diaphragm of the sensor without which the sensor ul-
Deflection of force measuring 10 to 500 µm timately cannot be tested and calibrated.
Table 2 spring (passenger weight)

Sensor measuring principles | Position sensors 169
In the case of travel and distance sensors tion by a reference mark detectable by the
covering large measuring ranges (approxi- sensor, in case a fault should have occurred.
mately 0.1 to 150 m), the measured vari- This also takes account of the fact that the
able is converted into a pulse or phase position of the crankshaft when the engine
runtime by means of wave-propagation is started is not known; it must first rotate
sensors (sound and electromagnetic almost one revolution under the starter so
waves), which can then be easily measured that the reference mark is passed at least
electronically with relatively small trans- once. The crankshaft’s motion is also very
ceiver devices. In the case of sound waves, monotonous, i.e. it is always pretty much
however, the propagation medium (e.g. air) smooth (without sudden changes) and is
and the time standard required (e.g. oscil- always in one direction (that is forwards,
lating crystal) then become an essential and never backwards). You can thus be
component part of the measuring device. sure that no more than one revolution will
ever be incorrectly detected.
Measured variables: overview The disadvantages of an incremental
There are a large number of applications angle measurement cannot be absorbed
in which position represents the actual if the engine starts immediately without
measured variable. This is shown in over- being cranked, as is also the case with the
view by Table 1. In other cases, the mea- steering angle. An absolute angle sensor
sured position or angle represents a differ- for a 360° range is required here.
ent measured variable (Table 2).
In practice, “incremental sensor systems” Potentiometer sensors
are also often referred to as angular-position The wiper potentiometer, usually made as
sensors, even when they are used for mea- an angular-position sensor (Fig. 1), uses
suring rotational speed. These sensors are in the correspondence between the length of
reality not angular-position sensors. Since a wire or film resistor (Cermet or conduc-
the increments (steps with which a given tive plastic) and its resistance value for
variable increases) which have to be mea- measuring purposes. At present, this is
sured with these sensors in order to measure still the lowest-priced travel/angular-posi-
the deflection angle must be counted with tion sensor. The shape of the characteristic
the correct preceding sign (in other words, curve can be influenced by shaping the
added), these sensors are in reality not angu- measuring track (or just sections of it).
lar-position sensors. Due to the danger of the
counter being falsified as a consequence of 1 Principle of the wiper potentiometer
spurious pulses, such angular-position mea-

suring systems are only in limited use. Fixed,
U0 UA
directly locatable reference marks only pro-
vide very little help in this dilemma. Another R V1 RS
2 Fig. 1
disadvantage of such angular-position mea- ϕmax IA
1 Wiper
suring systems is the fact that the absolute 3 2 Resistance track
position is lost when the operating voltage is 3 Contact conductor
1
switched off. It is no use storing the final po- R0 track
sition in a non-volatile memory here, since IA Wiper current
U0 Supply voltage
most angular positions can also change me-
UA Measurement
chanically when power has been removed ϕ
Rϕ voltage
(e.g. steering angle).
UAE0289-1Y
R Resistance
Although it is true that the crankshaft R V2 wmax Maximum angle
angle is measured incrementally, this mea- of rotation
surement is recalibrated after each revolu- w Measured angle
The standard wiper connection is fur- • Measuring errors due to abraded parti-
nished by a second contact track consist- cles
ing of the same material mounted on a low- • Problematic when operated in fluids
resistance substrate. The supply voltage is • Variable contact resistance between
usually applied to the measurement track wiper and measurement track
through low-resistance series resistors RV • Strong acceleration or vibration can
(these can also be used for calibration of result in wiper lift-off
zero point and progression-rate adjust- • Testing is costly
ment) to protect against overloading. • Limited possibilities for miniaturization
Wear and falsification of measured val- • Background noise
ues can be kept to a minimum by keeping
the electrical loading of the pick-off as low Applications
as possible (IA < 1 mA) and by dust-proof Examples of potentiometer-type sensors:
and waterproof encapsulation. One of the • Accelerator-pedal sensor or accelerator-
prerequisites for low wear is the optimal pedal module for capturing the torque
friction pairing between the wiper and the request for the engine-management
conductor track. To this end, wipers can system
be of “spoon” or “scraper-shape” design, • Fuel-level sensor (Fig. 2)
and one or more can be mounted. Brush- • Sensor plate potentiometer (KE-Jetronic
shaped wipers are also in use. and L-Jetronic) for detecting the air
A whole range of clear advantages are mass taken in by the engine
faced by a considerable number of serious • Throttle-valve position sensor for cap-
disadvantages: turing the position of the throttle valve

in a gasoline engine (Figs. 3 to 5)
Advantages of the potentiometer-type
sensors 2 Potentiometric fuel-level sensor (design)
• Low costs
• Simple design 1
• Very broad measuring effect
(measuring range ≈ supply voltage)
• No electronic circuitry required
• High level of interference immunity
• Broad temperature range (up to 250 °C) 5
2
• High precision (better than 1 % of mea-
3
6
suring range) 4
• Wide measuring range (almost 360° is
possible)
• No difficulty with redundant design
• Calibration possible (laser ablation etc.)
• Flexible characteristic curve (variable 7
Fig. 2
1 Electrical conductor track width)
connections • Flexible assembly (on curved as well as
2 Wiper spring on flat surfaces)
3 Contact rivet • Numerous manufacturers
4 Resistor board
• Rapid prototyping
5 Bearing pin
9 8
UMK1790Y
6 Twin contact
7 Float lever Disadvantages of potentiometer-type
8 Float sensors
9 Fuel-tank floor • Mechanical wear from abrasion
3 Throttle-valve sensor (design) Magnetically inductive-type sensors

Of all the sensors using proximity and
non-contacting principles for position
1 2 3 4 measurement, the magnetic sensors have
proved to be the most rugged and most in-
sensitive to interference. This applies in
2 cm particular to those principles relying on al- Fig. 3
ternating current, in other words magneti- 1 Throttle-valve shaft
cally inductive principles. Compared with 2 Resistance track 1
a micromechanical sensor though, the coil 3 Resistance track 2
configuration needed here requires far 4 Wiper arm with
wipers
more space. This means therefore that
5 Electrical
there is no favorable possibility of redun-
UMK1306Y
connection
dant (parallel measurement) design. Fur-
5 thermore, coil contacting is less favorable Fig. 4
from the costs and reliability viewpoint. A Internal stop
Although there are a multitude of differ- 1 Characteristic
4 curve for high
Throttle valve sensor with two characteristic curves ent principles in use for this form of sen-
resolution in the
sor, only two have come to the forefront
angular range
1.00 for motor vehicle applications. Regarding 0° to 23°
their operating concepts, these are very 2 Characteristic
0.80
similar to each other. However, they are curve for angular
UV
1 2
UA
0.60 currently no longer in use in their macro- range 15° to 88°

Voltage ratio
mechanical version by Bosch for motor UA Measurement

voltage
0.40 vehicle sensors in new developments.
UV Operating voltage
aW Effective measured
0.20 Eddy-current sensors angle
0.05 When an electrically conductive flat or
0
0 15 30 60 90 curved disk (for instance in aluminum or
UMK1787-1E
Fig. 5
aw
A A copper) approaches a coil (usually nonfer- 1 Throttle valve
Angle of rotation w rous) (Fig. 6) to which high-frequency al- 2 Throttle-valve
sensor
ternating current has been applied, it has
UA Measurement
voltages
5 Throttle-valve sensor (circuit) 6 HF damping and eddy-current principle
UV Operating voltage
R1, R2 Resistance
tracks 1 and 2
R1 R3
R3, R2 Calibration
1 R5 resistors
R5, R6 Protective
R2 R4 1 2 3 4 5 resistors
Fig. 6
R6 1 Spoiler
2 A(s)
2 Eddy currents
UA(s) 3 Air-core inductor
4 Variable-damping
UV
oscillator
UMK1307-2Y
UA1 s
5 Demodulator
UAE0845Y
s Measured travel
UA2
A(s) Oscillator voltage
UA(s) Output voltage
an effect upon the coil’s equivalent resis- There are many ways to adapt the eddy-
tance and its inductance. This is the result current principle to the measuring assign-
of the eddy currents generated in the disk ment. It is suited both to the determination
(otherwise known as a spoiler) due to the of large travel distances or angles (there
increasing magnetic coupling. The disk’s has already been a practically production-
position represents the measured travel s. ready development, for the throttle valve
The generally high operating frequency and accelerator pedal, for instance) and
(MHz range, a low-frequency feed would also for very small variables (e.g. torque
take up too much current because of the sensors). Because of the self-compensating
low coil inductance) requires a direct as- properties, this principle generally only
signment of the electronics to the sensor shows a very small temperature sensitivity
or a shielded supply lead, which is gener- over a wide temperature range. As eddy-
ally not supportable. In order to convert current sensors can also, in principle, be
the measuring effect into an electrical out- manufactured micromechanically, their fu-
put voltage, either the damping effect ture use cannot be completely written off
(equivalent resistance) or field-displace- given their very advantageous properties.
ment (inductance) principles can be ap- This sensor principle is, however, widely
plied. In the first (damping effect) case, used in manufacturing and quality mea-
a variable-amplitude oscillator can be suring technology, for the precise detec-
used, and in the second a variable-fre- tion of the smallest displacements/dis-
quency oscillator or a constant frequency tances or for the measurement of layer
supplied inductive voltage divider (differ- thicknesses in the µm range.
ential configuration).
Short-circuiting ring sensors
7 Measuring principle of the short-circuiting In contrast to eddy-current sensors, the
ring sensor coil of a short-circuiting ring sensor has a
straight or curved U or E-shaped soft mag-
x1 netic, (usually) laminated iron core (Fig. 7).
I The coil, or short-circuiting ring (“spoiler”),
is of highly conductive material (copper or
aluminum), and is located around one (or
L(x1) 1
all) of the core limbs. Because of their iron
IW core, sensors of this type have a far greater
inductance than eddy-current sensors and
also a far higher measuring effect because
2 of the guidance of the magnetic flux being
3 F x1 very concentrated by the iron core. This
x2 means that they operate well at low fre-
Fig. 7
I quencies, and do not necessarily require
Imaging for two
different measured their signal electronics locally, that is di-
travels rectly on the sensor itself. The iron core
1 Short-circuiting L(x2) also protects the measuring space between
ring the core limbs very well against external in-
2 Soft-magnetic core
terference fields.
3 Coil
The alternating field generated in and
I Current
UAE0021-2Y
IW Eddy current
around the iron core by the coil current I
L(s) Inductance and cannot pass through the short-circuiting
F(s) Magnetic flux in
F (x2) ring, since the eddy currents in the short-
measured travel s circuiting ring reduce it practically to zero.
Thus the eddy currents in the short-cir- The cores are laminated (e.g. NiFe plate,
cuiting ring, acting like a “magnetic insula- 0.2 mm thick) because of the alternating
tor”, limit the extent of the magnetic flux F current operation. To achieve the necessary
to the space between the coil and the mechanical stability to keep the plates sta-
short-circuiting ring (hence its other name ble under acceleration, they are not only
of shading ring) and thus make it depen- bonded in the normal fashion, but also in-
dent on the position x of the short-circuit- terlocked. To this end, the laminations are
ing ring F = F(x). given a knobbly pattern in selected places
Inductance is defined: which can be used to connect the lamina-
(1) L = F/I tions together before soft annealing (just
like LEGO bricks), without interfering too
This is why the position x of the short-cir- much in the electrical insulation from lami-
cuiting ring also directly influences the in- nation to lamination required.
ductance L of the field coil. The relation- The half-differential short-circuiting
ship L = L(x) shows a good linear fit over ring sensor with moving measuring ring
a wide domain. Virtually the entire length and fixed reference short-circuiting ring is
of the sensor can be utilized for measure- very accurate; it is evaluated as an induc-
ment purposes. A mechanically tight guid- tive voltage divider (analysis of the induc-
ance of the short-circuiting ring is not nec- tances L1/L2, or (L1 – L2)/(L1 + L2) or even as
essary here. a frequency determining element of an os-
The mass of the short-circuiting ring to cillating circuit to generate a frequency
be moved is very low. Shaping the gap be- analog signal (very interference-proof,
tween coil and limb has an effect upon the easy to digitize).
shape of the characteristic curve: reducing
the gap towards the end of the measuring Applications
Fig. 8
range further improves the already good Examples for short-circuiting ring sensors:
Design of the rack-
linearity. Operation is generally in the 5 to • Rack-travel sensors for measuring the
travel sensor (RWG) for
50 kHz range, depending on material and control-rack setting on in-line diesel diesel in-line injection
structural shape. This sensor can also be injection pumps (Fig. 8) pumps
used in very severe operating conditions, • Angular-position sensors in the in- 1 Soft-magnetic core
for instance on diesel fuel-injection pumps jected-fuel quantity actuators of diesel 2 Reference coil (L2)
3 Reference short-
(Figs. 8, 9). distributor-type injection pumps
circuiting ring
4 Control rack
8 Half-differential short-circuiting ring sensor 9 Half-differential short-circuiting ring angular-
5 Measuring coil (L1)
position sensor
6 Measuring short-
circuiting ring
s Control-rack travel
3 1 1 2 60 3 4
ax
wm Fig. 9
URef 0 1 Measuring coil
2
w
2 Measuring short-
UA UA circuiting ring
3 Soft-magnetic core
4 Control-collar shaft
6 5
5 Reference coil
4 URef
6 Reference short-
s
circuiting ring
UAE0746-1Y
UAE0290-1Y
w Measured angle
5 6 wmax Adjustment-angle
range for the
control-collar shaft
Figure 9 shows the structure of a half-dif- or even:

ferential short-circuiting ring sensor for (3) u1 = U?sinw
diesel distributor-type injection pumps. (4) u1 = U?sin(w - 120°)?sinvt
A sensor coil and a reference coil are at- (5) u1 = U?sin(w - 240°)?sinvt
tached to each leg of the core. As the alter-
nating current transmitted by the control After rectification, the angle of rotation a
unit permeates the coils, it creates alter- can be calculated with great precision
nating magnetic fields. The reference from these voltages. The sensors prefera-
short-circuiting ring is fixed in position, bly made as angle sensors identified in
while the measuring short-circuiting ring classical measurement technology as syn-
is attached to the control-collar shaft (ad- chro, resolver or even inductosyn methods
justment angle w). As the measuring short- function in this way.
circuiting ring moves along, the magnetic In a simple bipolar arrangement, the
flux changes and the voltage across the electrical phase angle also corresponds
coil changes, since the control unit main- directly to the mechanical angle of rota-
tains the current constant (load-indepen- tion a. In an arrangement of n pole pairs,
dent current). An evaluation circuit forms the mechanical full angle w = 2p is con-
the ratio of the output voltage UA to the verted into a phase displacement of
measurement inductance L1 to the refer- a = n?2p, so that the phase displacement a
ence voltage URef to the reference induc- corresponds only to an angle of rotation
tance L2. It is proportional to the excursion of only w/n, which considerably enhances
of the measuring short-circuiting ring and the resolution of the measuring system.
can be evaluated by the control unit. However, the unambiguity of the measure-
ment signal must be restored by additional
Sensors with rotatable alternating fields means, such as a simple angle sensor,
With alternating current of the angular fre- in larger measuring ranges.
quency v fed coils or coil-like structures
(such as meandering conductor track Hella sensor
structures) two or more pole alternating The angle sensor developed by Hella
fields structures may be arranged either does not strictly correspond to any of the
in a circle or even linearly. These pole methods described above, but is most simi-
structures with fixed pole pitch can be lar to the inductosyn method. Figure 10
displaced in relation to a generally fixed shows the sensor with a 6-count pole struc-
set of receiver coils having the same pole ture (n = 6), for instance, which electrically
pitch, by the movement of the system to be seen converts an angle of rotation of w = 60°
measured, either rotary or translatorial. into a phase displacement of the signal am-
Here the amplitudes of the receiver signals plitudes of a = 360°. All the conductor track
are changing with the movement in a sinu- structures required are applied to multi-
soidal pattern. If the receiver coils are dis- layer PCBs, at least in the case of the fixed
placed in relation to one another by a spe- part (stator). The rotor may also be made
cific portion of the pole pitch T (e.g. T/4 or as a stamped part, if necessary, either free-
T/3), the sinus pattern is phase-shifted by standing or mounted on a plastic carrier
a corresponding angle (e.g. by 90° or 120°). (hot stamping).
The voltages arise as follows, for instance: A circular conductor track loop is made
(1) u1 = U?sinw?sinvt on the stator. This induces an eddy current
(2) u2 = U?sin(w - 90°)?sinvt at an operating frequency of 20 MHz in a
= U?cosw?sinvt closed meandering loop of approximately
the same outside diameter on the rotor.
This eddy current naturally generates a According to Figure 10c, the receiver coils
secondary magnetic field, just like the are connected in a star connection. Their
exciter loop, which superimposes itself on signals are sent to an ASIC to determine
the excitation field in the sense that it at- the electrical phase angle a or mechani-
tempts to eliminate it. If the track on the cal angle of rotation w, which applies the
rotor were only a circular conductor track necessary (sign-correct) rectification,
congruent with the stator loop instead of selection and ratio formation. A version
the meandering track, this would simply ASIC 1 receives the digital control signals
extinguish the primary field to a great ex- necessary for this from an adjacent mi-
tent. The meandering structure, however, crocontroller. A different version ASIC 2
creates a resulting multi-pole field which is, however, also capable of operating the
can be rotated with the rotor and the total sensor completely independently (on a
flux of which is also naturally almost zero. stand-alone basis). The ASICS also permit
This multi-pole alternating field is an end-of-line calibration of the mechani-
sensed by concentric receiver coils or me- cal and electrical tolerances in production.
anders of practically identical form located It is also possible, for the case of applica-
on the stator. These are offset within a pole tions with elevated safety requirements,
pitch (of e.g. 60°) by 1/3 each, i.e. electri- to construct a redundant system with two
cally in their signal amplitude by 120° each galvanically isolated signal paths and two
(Fig. 10b). The receiver coils however ex- ASICs. The sensor principle can also very
tend over all n pole pairs (series circuit) advantageously be applied in a “cut-open”
and use the sum of all pole fields. form as a travel sensor.
The principal benefits of this sensor, in ad-

dition to the very favorable production
10 Hella sensor costs, are as follows:
• Contact-free and wear-and-tear resis-
a U3
i(t) U2 tant measuring principle
U1 • Complete temperature independence
(up to 150 °C)
• High accuracy (up to approximately
±0.09° for a measuring range of 360°)

• Flexibility (may be adapted to any angu-
lar range up to 360°)

Stator Rotor • Possibility of a redundant design
b • High EMC
U1
• Use of standard materials exclusively,
60°
without the need for ferromagnetic
U2
components
• Flat design
U3
Angle α The sensor can therefore be used at very

many points on the motor vehicle.
c Rotor
Stator
SAE1016-1E
Fig. 10
a Schematic design
b Circuit
c Output signals
Magnetostatic sensors current I by angle w from their otherwise

Magnetostatic sensors measure a DC mag- straight path (Fig. 11). This means that it is
netic field. In contrast to the magnetically possible to pick-off a voltage UH transverse
inductive (coil) sensors they are far more to the current direction between two op-
suitable for miniaturization and can be posing points on the edge of the wafer pro-
manufactured at reasonable cost using portional to the field B and to the current I
microsystem technology. As DC magnetic (Hall effect).
fields can reach through housing walls of (2) UH = RH ? I ? B/d
plastic, and also those of non-ferromag- where RH = Hall coefficient,
netic metals, with no problems, magneto- d = wafer thickness
static sensors have the advantage that the
sensitive, generally fixed part, can be well At the same time, the series impedance of
encapsulated and protected against the ro- the wafer increases independently of the
tating part, generally a permanent magnet field direction following an approximately
or soft-magnetic conductive element – and parabolic characteristic curve (Gaussian
also against the environment. Above all, effect, magnetoresistor).
galvanomagnetic effects (Hall and Gauss- The coefficient determinant for the mea-
ian effects) are used, as well as anisotropic surement sensitivity of the wafer RH is only
magnetoresistive (AMR and GMR) effects. comparatively small for silicon. As the
wafer thickness d can be made extremely
Galvanomagnetic effect thin, however, using diffusion techniques,
The Hall effect is normally evaluated using the Hall voltage amounts to a technically
thin semiconductor wafers. If a current- usable magnitude. When silicon is used as
carrying wafer of this type is exposed the basic material, a signal-conditioning
perpendicularly to magnetic induction B, circuit can be integrated on the wafer. Sen-
the charge carriers are deflected by the sors using these principles are very cheap
Lorentz force normal to the field and the to produce. Silicon is by a long way not the
best semiconductor material for Hall sen-
11 Galvanomagnetic effect sors in respect of measurement sensitivity
a +B and temperature sensitivity. III-V semicon-
ductors such as gallium arsenide or indium
antimonide possess better characteristics.
I
Hall switch
IH a In the most simple case, the Hall voltage is
Fig. 11 IV applied to an electronic threshold circuit
a Circuit (Schmitt trigger) built into the sensor
b Curve of Hall
UR
which then outputs a digital signal. If the
voltage UH UH
magnetic induction B applied at the sensor
c Increase of wafer d
resistance R is below a given minimum threshold level,
(Gaussian effect) b c R the Schmitt trigger’s output value corre-
B Magnetic induction sponds to a logical “0” (release status);
I Wafer current if it exceeds a given upper threshold the
+UH
IH Hall current output value corresponds to a logical “1”
IV Supply current
-B +B -B +B (operate status). Since this behavior is
UR Longitudinal
guaranteed across the complete operating-
UAE0699-5Y
voltage
a Deviation of the temperature range and for all sensors of a
electrons due to +UH given type, the two threshold values are
the magnetic field relatively far apart (approx. 50 mT). A con-
siderable induction jump DB is thus neces- Hall sensors using the spinning-current
sary to operate the Hall switch. principle
Sensors manufactured in bipolar tech- A disadvantage of a simple silicon Hall
nology of this kind were used in Hall vane sensor is its simultaneous sensitivity to the
switches (Fig. 12), for example, as were fit- unavoidable mechanical strain (piezoelec-
ted in the housing for the camshaft-driven tric effect) resulting from the packaging,
ignition distributor. This Hall vane switch leading to unfavorable temperature sensi-
contains the sensor and a permanent mag- tivity of the offset. Applying the spinning
net and soft magnetic conductive ele- current principle (Fig. 13), associated with
ments. The magnetic circuit is U-shaped or a transition to CMOS technology, has over-
forked in such a way that a trigger wheel come this disadvantage. Although the
vane in soft magnetic material can travel piezoelectric effect still occurs here, it is
through the open end, alternately screen- compensated for by the signal being aver-
ing or exposing the magnet, thereby tog- aged over time, as it occurs with a different
gling the Hall sensor between the operate sign in the case of very rapid electronically
and release states. controlled interchange (rotation) of the
A further application is the digital steer- electrodes. If it is wished to save the ex-
ing-angle sensor LWS3 (see “Hall angular- pense of the complex electronics for
position sensor up to 360°”). switching the electrodes, it is possible also
Hall sensors of this type may be very to integrate several Hall sensors (two, four
cost-effective, but are really only good for or eight) with correspondingly different
a switch operation and are too imprecise alignment of the current paths in close
for capturing analog values. proximity and to add their signals together
in order to form an average. It is only in
this way that the Hall ICs were also suit-
able for analog sensor applications.
12 Hall vane switch 13 Hall sensor using the spinning-current principle
a Fig. 12
a Magnetic flux:
a 2
unhindered
I I b Magnetic flux:
short-circuited
1 Vane width b
S N
2 Soft-magnetic
conductive element
3 Hall IC
b 2 3 4 1 1 2 3 UH 4 Air gap
b U0 Supply voltage
I
US Sensor voltage
Fig. 13
US S N
a Rotary phase w1
b Rotary phase
w2 = w1 + 45°
1 Semiconductor
U0 b
I wafer
UAE0700-2Y
UAE0639-1Y
2 Active electrode
3 Passive electrode
UH I Supply current
UH Hall voltage
These measures though did not result in a Differential Hall sensors

reduction of the considerable effects of The double Hall sensor (differential Hall
temperature on the sensor’s measurement sensor, Fig. 14) has two complete Hall sys-
sensitivity. tems mounted at a specified distance from
Such Hall ICs are suitable above all for each other on one chip. The associated
the measurement of small travel distances, electronics evaluate the difference be-
in which they register the fluctuating field tween the two Hall voltages. The advan-
strength of a permanent magnet as it ap- tage of such sensors lies in the fact that
proaches (e.g. the iBolt force sensor de- their output signal is for the most part
tects the passenger’s weight to optimize independent of the absolute value of the
the triggering of the airbag). Similarly magnetic-field strength, and as differential
good results could, in the past, only be sensors they only register the magnetic in-
achieved by using single Hall elements, duction’s change in space, in other words
e.g. in III-V compounds with a hybrid am- the field gradient (thus the common name
plifier connected downstream (e.g. Hall of gradient sensor).
acceleration sensor). Since the polarity of their output signal
is independent of the air gap between ro-
tor and sensor, these sensors are mostly
used for rotational-speed measurement.
If a simple Hall sensor only is used to
sense a gear wheel, this will not be able to
distinguish whether the magnetic flux has
modified as a result of the gear wheel turn-
14 Differential Hall sensor ing further or by a change in distance (e.g.
vibrations, mounting tolerances). Signifi-
a cant scanning errors occur, as the signal
w
must be passed to a threshold detector.
1 The situation with differential sensors is
fundamentally different. Here only the dif-
2 ference in the signals from two Hall sen-
3 sors located at a suitable distance from
S
4 N
one another around the circumference is
evaluated. If the signal difference is posi-
b tive, for example, it is possible to freely
w
vary the distance between the gear wheel
and sensor; the difference will remain pos-
itive, even if the absolute value may be-
come smaller. The sign can only change
Fig. 14
if the rotor turns further. A threshold de-
a Design
b Field-strength tector connected afterwards thus has no
distribution (1.5 problem in distinguishing between a
times increment c change in distance and rotation.
spacing)
L = 2 mm
Signal curve
c Signal curve for

Usually, in order to achieve as high an
air-gap widths L
3 mm output signal as possible, the distance be-
1 Ring gear
UAE0862-2E
2 Differential Hall IC
tween the two Hall sensors each located
3 Homogenizing on the edge of the (elongated) chip is then
wafer (soft iron) Angle of rotation w selected to correspond to about half an
4 Permanent magnet increment (half the rotor tooth interval).
The signal maximum is very wide and cov- sensor is located slightly outside the cen-
ers a broad range of variation of the incre- ter of the circular arc, the characteristic
ment spacing. More pronounced varia- curve increasingly deviates from the sinu-
tions in the spacing necessitate a highly soidal, It now features a short measuring
complex redesign of the sensor. range of just 90°, and a longer measuring
A gradient sensor must be precisely range of a little over 180° with good linear-
aligned to the rotor’s direction of rotation.
15 Analog Hall angular-position sensor
(movable magnet)
Hall angular-position sensors in the range
up to approximately 180° a w b
Using a rotatable magnetic ring (“movable
magnet”), together with a number of fixed
soft-magnetic conductive elements, a lin-
ear output signal can be generated for a Fig. 15
larger angular range without conversion Linear characteristic
being necessary (Fig. 15). In this case, the curve for angles up
bipolar field of the magnet ring is passed to 180°
through a Hall sensor arranged between 1 2 3 4 5 a Position a

b Position b
semicircular conductive elements. The ef-
c c Output signal
fective magnetic flux through the Hall sen- 1 Magnetic yoke
sor is dependent on the angle of rotation w. 2 Stator (1, 2 soft
This principle is used on accelerator- Pos. a Pos. b iron
w
pedal sensors. 3 Rotor (permanent
UAE0702-3Y
90 0 + 90 180
magnet)
4 Air gap
The Type ARS1 Hall angular-position
5 Hall sensor
sensor with a measuring range of ap-
w Angle of rotation
prox. 90° (Fig. 16) is derived from the basic
“movable magnet” principle. The magnetic 16 ASR1 Hall angular-position sensor
flux from a practically semicircular perma- (movable magnet)
nent-magnet disk is returned to the mag- a

nets through a pole shoe, two additional 4
conductive elements each of which con- 5
w
tains a Hall sensor in its magnetic path, 1
and the shaft which is also ferromagnetic.
Depending upon the angular setting, the 2 3 Fig. 16
flux is led to a greater or lesser degree Linear characteristic
through the two conductive elements, in curve for angles up
4 6
to approximately 90°
the magnetic path of which a Hall sensor
a Design
is also situated. Using this principle, it is b b Characteristic
possible to achieve a practically linear curve with
mT A
characteristic in the measuring range. 150 working range A
1 Rotor disk
Flux density B
The Type ARS2 is a simplified version 100 (permanent-

magnetic)
which does without soft magnetic conduc-
50 2 Pole shoe
tive elements (Fig. 17). In this version, the
3 Conductive element
UAE0770-4E
magnet moves around the Hall sensor in a 0 4 Air gap

45 90 135 180 225 270 315
circular arc. Only a relatively small section 5 Hall sensor
-50
of the resulting sinusoidal characteristic Angle of rotation w 6 Shaft (soft
curve features good linearity. If the Hall magnetic)
ity. A great disadvantage though is the low (3) UH1 = Ux = B?sinw

level of shielding against external fields, as (4) UH2 = Uy = B?cosw
well as the remaining dependence on the
geometric tolerances of the magnetic cir- The angle w can be calculated using the
cuit, and the intensity fluctuations of the trigonometric relationship w = arctan (UH1/
magnetic flux in the permanent magnet as UH2) in an associated commercially avail-
a function of temperature and age. able evaluator chip, through which the
sensor signal is generally digitized.
Angular-position sensors in the range up to
approximately 360° A Hall sensor arrangement of this nature
Figure 18 shows an analog Hall angular- can also, in principle, be integrated verti-
position sensor with a measuring range cally with VHD (Vertical Hall Devices) in
of up to 360°. As shown, a permanent mag- the manner illustrated so that the plane of
net rotates in front of two Hall sensors ar- the sensor chip is arranged perpendicular
ranged at right angles to each other. The to the axis of rotation and the sensor has
magnet should be adequate in size in order an in-plane sensitivity, unlike a normal,
not to be too dependent on the tolerance planar Hall sensor (Fig. 19). Monolithic in-
of the magnet positioning. The two Hall tegration also guarantees the high level of
sensors, fitted as closely together as possi- precision as required for the 90° arrange-
ble, should here be positioned in the ho- ment and for the requisite compact form of
mogeneously directed portion of the per-
manent-magnet’s stray field, which also 18 Analog Hall sensor for 360° angles
represents the angular location w of the a

Fig. 18 permanent magnet. They are perpendicu- I
Bx
a Constructed from
lar to one another and parallel with the
discrete Hall ICs
permanent magnet’s rotary axis. When the
b Constructed from Ux
planar integrated field-strength vector B rotates past the
Hall ICs sensors, therefore, they register its x and
w
1 Signal electronics y components:
B
2 Camshaft By
3 Control magnet
Uy
B Induction
I Current 2
U Voltage 17 ASR2 Hall angular-position sensor I
(movable magnet) 3
UA Output voltage
w Angle of rotation
a b
Fig. 17 90 mT b
y 60
Flux density B
Linear characteristic
w
curve for angles over 30
180 1 0
180° x 0
90 180 270
a Principle of NS -30
operation 3 -60 w
b Characteristic
270 Angle of rotation w B
curve 90 mT
Flux density B
1 Hall IC positioned y 60
in the mid-point of 30 2
180 2 0
the circular path 0
UAE0863-3E
UAE0701-3Y
x 90 180 270
2 Hall IC located NS -30
outside the mid- 3 -60
point (linearization) 270 Angle of rotation w
UA 1 3 2
3 Magnet
the two Hall systems. Sentron (Melexis), of Because of its high relative permeability,
Switzerland, is working on sensors of this this disk acts as a flow concentrator (IMC,
type and will be launching them on the Integrated Magnetic Concentrator) and
market in the near future. compels all field lines to pass into its sur-
face vertically. This compels the field lines
Angular-position sensors of this type were running horizontally (in-plane) without
considered for the implementation of the concentrator into a vertical direction
starter motor-less instant start systems, (out-of-plane) at the position of the Hall
for instance, where the absolute rotary sensors and thus they can modulate the
position of the camshaft over a 360° range Hall elements. As diagonally opposing
must be measured. elements each see fields in opposite direc-
tions, the difference between the two Hall
There is, however, another trick which can voltages running in opposite directions is
be exploited to (additionally) give a nor- evaluated to capture a field component. At
mal, planar Hall sensor an in-plane sensi- the same time, any vertical components Bz
tivity. In the case of the method used by present have their effect eliminated.
Melexis-Sentron (e.g. its 2SA-10 model) a
total of four Hall sensors are mounted on The additional evaluation of the summary
the same chip in a tight circle offset at 90° signals of diagonally opposing elements
to one another to measure the field com- can also be used to capture the vertical
ponents Bx and By. To be accurate, in this field component Bz; since these are not in
case, each sensor consists of a Hall ele- the slightest influenced by the ferromag- Fig. 19
ment pair rotated at 90° for its part for the netic disk and the elements have also not The currents directed
reasons described above. Once the chip lost their normal out-of-plane sensitivity into the interior of the
chip I are deflected
has been manufactured, a circular disk
by the Lorentz forces
of soft magnetic material approximately 20 Set of 4 x 2 planar Hall sensors with IMC
because of the flux
200 µm in diameter is fixed on its surface density aligned in
as shown in Figure 20 in such a way that parallel with the chip
the Hall sensors are located exactly be- surface B in such a way
neath the edge of the disk. that a Hall voltage UH
can be picked off
between the surface
electrodes A and A'.
19 Vertical hall device (VHD) in cross-section

Fig. 20
Bx By
IMC Integrated Magnetic
Concentrator
The Hall elements are

mounted on the chip in
UH
pairs at an angle of 90°
I/2 A I A` I/2 B_I B_I to one another directly
BII BII below the edge of a
small ferromagnetic
disk. Thanks to their
B presence, the flux
Lorentz force
1 2 3 density parallel to the
chip surface (induction)
SAE1019-1E
SAE1020-1Y
B|| is converted into a

flux density B that can
be sensed by the Hall
sensors.
by it being fitted. In this way, an arrange- The maximum intrinsic deviation of the
ment of this type can be used to capture sensor is stated by the manufacturer as
the induction vector B acting at the sensor + 2° for a measuring range of 360°. This
location in all its three components. The can, however, be significantly improved
electronic circuitry required to evaluate by the user on calibration of the sensor.
the signals including a µC-based digital The accuracy of the digital output is then
signal processor (DSP) used to calculate 10 bits at a resolution of 12 bits. Because of
the arctan function and the means re- the processing time required by the signal
quired for an end-of-line calibration of processor, the smallest sampling rate for
the sensor (EEPROM) can cost-effectively a timing frequency of 20 MHz is 200 µs.
also be integrated on the same chip as the A pulse width modulated signal output
sensor. can be operated at a maximum of 1 kHz.
The sensor needs 15 ms for initialization.
In order to measure angles up to 360°, it is A typical value for the measuring field
sufficient to rotate a permanent magnet, strength is approximately 40 mT. The con-
preferably round in form, magnetized in centrator plate becomes saturated at flux
parallel with the chip over the sensor chip densities in excess of 0.7 T.
with its integrated signal processing The sensor may also be programmed for
(Fig. 21). As the angle of rotation is formed any measurement range < 360° to good ef-
using the arctan function from the ratio of fect, with additional functional security
the two sine wave and cosine wave sensor being generated by out-of-range monitor-
signals, the strength of the magnetic field, ing (application as an angular-position
and thus also the aging of the magnets, its sensor in the accelerator-pedal module
temperature sensitivity and its distance FPM2.3, for instance).
from the sensor surface are largely irrele-
vant. Only the rotary position of its magne- Angles up to 360° may also be captured
tization is detected. with the simplest Hall ICs (Hall switches),
as are also used for the measurement of
speeds of rotation, (application in LWS1
type steering-angle sensors). Here, in or-
der to obtain an n-bit resolution, n Hall
switches are arranged on a single track in
21 Angular position measurement over 360° with a circle at equal distances from each other
quadruple Hall sensor (Fig. 22). Depending upon its position, a
rotatable soft-magnetic code disk blocks
a a b the magnetic field of the individual perma-
nent magnets located above each Hall
switch, or opens it when it rotates further
Fig. 21
2 S N so that one after another the Hall switches
Angle measurement
Hall voltage
at the end of a generate n different code words (serial

rotating shaft with code). It is useful to apply the Gray code
quadruple Hall sensor (cyclic binary code) to prevent errors in
manufactured by transitional states of the code disk. The
Melexis containing
Gray code is designed such that the code
an integrated flux 1
0° 180° 360° 540° words for two adjacent positions – unlike
concentrator (IMC)
the case with a natural binary code, for in-
UAE0997-2E
to deflect the field. Angle of rotation a

a Measuring system sin a stance – differ in only 1 bit, and so any in-
cos a correct indication on the transition from
b Primary output
signals
one position to the next can never amount Magnetoresistive sensors

to more than one angular step. In addition to the transversally directed
For the practical implementation of a Hall effect, there is also a longitudinal re-
steering-wheel-angle sensor, the code disk sistance effect, also known as the Gaussian
is connected to the steering shaft for in- effect, that occurs on semiconductor wa-
stance, and the sensor’s non-moving parts fers. Elements that exploit this effect are
to the vehicle body. One difficulty with this known as magnetoresistors which are
sensor is that the soft magnetic code disk manufactured from a III-V-semiconductor
must be mounted floating on the steering (crystalline indium antimonide (InSb)).
column for reasons associated with assem- In contrast with the Hall sensors, the opti-
bly tolerances, which leads to a complex mum wafer shape for a magnetoresistor
and space-hungry mounting there using a tends to be shorter and squatter, and rep-
flexible plastic sleeve. Because of the air resents a very low electrical resistance.
gap tolerance between the upper disk and In order to arrive at technically applicable
the lower disk, the code disk is mounted values in the kV range, it is therefore nec-
in a wiping guide. A corresponding opto- essary to connect a large number of these
electronic solution using photoelectric wafers in series. This problem is solved
light barriers is less satisfactory, as it is elegantly by adding microscopically fine,
not possible to construct an adequate highly conductive nickel-antimonide nee-
encapsulation against the susceptibility dles to the semiconductor crystal. These
of the sensor elements to dirt. are located transverse to the current direc-
In order to measure a number of com- tion. A further measure is to apply mean- Fig. 22
plete rotations, an additional 3-bit configu- der techniques to the semiconductor resis- Angular position
ration can be used in which the code disk tor (Figs. 23 and 24). measurement up to
360° with a circular,
is rotated by a reduction gear. The resolu- The dependence of the resistance on the
equidistant arrange-
tion of such configurations is mostly no magnetic induction B follows a square-law
ment of simple Hall
better than 2.5°. function up to inductances of approx. switches
0.3 T, and above this point it is increasingly 1 Housing with
linear. There is no upper limit to the con- permanent magnets
trol range, and dynamic response can be 2 Code disk
regarded as practically free from lag, as is 3 Printed-circuit

board with Hall
the case for the Hall sensor.
switches
22
Digital Hall angular-position sensor with 23 Magnetoresistive differential sensor
Fig. 23
n Hall switches
Magnetic activation
of a magnetoresistive
differential sensor for
gear wheel scanning
ϕ
(incremental angular
1 4 position measurement,
1
2 speed of rotation
S sensing)
3
2 N 1 Magnetoresistors
R1, R2
2 Soft-magnetic
substrate
3
1 3 Permanent magnet
UAE0022Y-1
UFL0029-2Y
R2 R1 4 Gear wheel
U0 U0 Supply voltage
UA( ϕ) UA Output voltage at
angle of rotation w
Since their temperature sensitivity has a tive measures which would otherwise be
pronounced effect upon their resistance needed. Furthermore, in their role as pas-
value (approximately 50 % reduction over sive, resistive components they are highly
100 K), they are usually supplied only in insensitive to electromagnetic interference
the dual-configuration form in voltage-di- and, as a result of their high bias voltage,
vider circuits (differential magnetoresis- practically immune against external mag-
tors). For the particular application, each netic fields.
of the two resistor sections must then be Because of its pronounced temperature
magnetically triggered (as far as possible sensitivity, the magnetoresistor is used al-
with opposite polarities). Notwithstanding most exclusively in incremental angular-
the high temperature coefficients of the in- position and rev counters, or in binary
dividual resistors, the voltage-divider cir- limit-value sensors (with switching charac-
cuit guarantees good stability of the work- teristic).
ing-point (that point at which both resistor
sections have the same value). Anisotropic magnetoresistive sensors
In order to achieve good measurement (AMR)
sensitivity, it is best to operate the magne- Thin, only around 30 to 50 nm thick NiFe
toresistors at a magnetic working point be- alloy coatings show an electromagnetically
tween 0.1 to 0.3 T. Generally, the required anisotropic behavior. That means that
magnetic initial-tension is supplied by a their electrical resistance changes under
small permanent magnet, the effects of the influence of a magnetic field. Resis-
which can be increased by using a small tance structures of this type are thus
magnetic return plate. known as anisotropic magnetoresistive
The magnetoresistor’s main advantage is (abbreviated to AMR) elements or sensors.
its high signal level which is usually in the The metal alloy generally used is also
volts range. This means that amplification known as permalloy.
is unnecessary, as well as the local elec-
tronic circuitry and the associated protec- Technology and designs
Even without an external control fields,
24 Microscopic sections of a magnetoresistor a small spontaneous magnetization MS
a will arise in the longitudinal direction of
a conductor track on a long slim resistor
strip, as shown in Figure 25a, (shape an-
isotropy). In order to give this magnetiza-
tion a clearly defined direction – theoreti-
cally, it could be in the other direction –
AMR sensors are often provided with weak
bias magnets, as shown. In this state, the
longitudinal resistance achieves its great-
b
est value Rll. If the magnetization vector is
rotated under the effect of an additional
external field Hy through angle q, the lon-
gitudinal resistance declines gradually un-
til it reaches its minimum value R at q = 0.
Fig. 24
Here, the resistance depends only on the
a Microscopic
UAE0864-1Y
section (cut-out)
angle q which is enclosed by the resulting
b On ferrite substrate magnetization MS and the current I, and it
in Super-8 film has an approximately cosine wave pattern
carrier as a function of q:
(5) R = R0?(1 + b?cos2 q) one of the resistances increases while the

where Rll = R0?(1 + b) ; R = R0 other drops.
Being thin-film sensors, anisotropic
The coefficient b here identifies the maxi- magnetoresistive sensors also have the
mum possible variation in resistance. It advantage that they can be calibrated to
amounts to around 3 %. If the external a setpoint value (e.g. zero point) by laser
field is much stronger than the spontane- trimming, for instance. In principle, the
ous generated magnetization, and this is oxidized silicon wafers which serve as the
usually the case when permanent magnets substrate material can also incorporate the
are used, the effective angle q is almost electronic circuitry for signal conditioning.
completely a function of the direction of At present, cost considerations dictate that
the external field. The field strength is now sensor chips and electronic chips are for
irrelevant, and in other words the sensor is the most part manufactured separately,
now operating in a “saturated state”. and then mounted and packaged in a com-
Highly-conductive short-circuiting mon leadframe, for instance. The control-
strips (for instance of gold) on the AMR ling magnetic field B is usually generated
film force the current to flow at an angle of by magnets moving side-to-side or rotat-
below 45° to the spontaneous magnetiza- ing over the sensor.
tion (longitudinal direction) without the
application of an external field. As a result In addition to the simple, two-pole AMR
of this “trick”, the sensor curve shifts by elements, there are also pseudo Hall sen-
45° compared to that of the simple resistor. sors in practically square NiFe thin layer
This results in the so-called “Barber Pole” structures, that have four connections, just
sensor. This means, therefore, that even like normal Hall sensors. Two for the cur-
with the external field strength Hy = 0, the rent path and two transverse to this for
curve is at the point of maximum measure- pick-off a (pseudo) Hall voltage (Fig. 26).
ment sensitivity (reversal point). Unlike a normal Hall sensor, the pseudo
The striping of two resistors in opposite Hall sensor, however, has its sensitivity to
directions causes them to change their re- magnetic fields in the plane of the layer,
sistances in opposite directions under the and not perpendicular to it. In addition, it
effects of the same field. In other words, does not have a proportional characteris-
tic curve, but a sinusoidal one with very a
25 AMR basic principle, barber pole structure
a 1
I
b 3
I
Ms q Ms q
I I
Hy S y Hy
N
2
z x
R(q)/R0 R(q)/R0
1+b 1+b
1 1
≈ ≈
SAE1021-1Y
q q
-90° 0 90° -90° 0 90°
Hy Hy
0 0
high match to a sinus wave, which does not Simple AMR angular-position sensor for
depend in the least on the strength of the ranges < 30°
control field and the temperature. The out- Barber pole AMR angular-position sensors
put voltage vanishes for a field parallel to with limited precision and a restricted
the current path, to describe a sinus half- measuring range (max. ±15°) rely on the
period on rotation up to an angle w = 90°. unbalance of a magnetoresistive voltage
The sinus voltage obtained in this way divider (differential sensor) comprising
combined with the amplitude results in: an elongated (possibly meander-shaped)
(6) UH = uH?sin2w permalloy resistor with highly conductive
gold cross strips (Fig. 27). Although in
If the external control field is rotated once the case of such sensors the zero point
through w = 360°, the output voltage follows is largely independent of the distance be-
two full sinus periods. The amplitude uH is, tween the magnet and the sensor, the gra-
however, very dependent on the tempera- dient of the characteristic curve, which for
ture and the air gap between the sensor its part also depends on the temperature
and the control magnet; it declines with (TK around -3?10-3/K), on the other hand,
increasing temperature and widening is not. A sensor of this kind has already
air gap. been considered as a cost-effective alter-
The measurement sensitivity of the native for the measurement of pedal posi-
pseudo Hall element can be considerably tions.
increased (without excessively falsifying
the sinusoidal shape) by “hollowing out”
the element from the inside so that only
the “frame” remains (Fig. 26b). This modi-
Fig. 27
fication converts the pseudo Hall sensor to 27 Magnetoresistive angular-position sensor
Measuring range
a full bridge consisting of four AMR resis- (barber pole)
up to 615°
a Measuring principle tors from its geometrical form (Fig. 26c).
b Characteristic Even when the bridge resistors are mean-
a ϕ
curve der-shaped, provided a given minimum
1 Permalloy resistors conductor width is not fallen below, this
2 Rotatable
still has negligible effect of the signal’s U0
permanent magnet B
sinusoidal shape.
with control
induction B
3 Lower operating 26 Pseudo Hall sensor
UA 2
temperature
4 Higher operating
a
temperature
B
a Linear measuring 1
range I I
UH
b Effective measuring
range w b 100
3
UA Output voltage mV
UA – U0 / 2
U0 Supply voltage (5V) 4

w Angle of rotation b c
0
Fig. 26 RH1 RH2 a

a Full area basic form UH I I
b
UAE0641-1E
RH3 RH4
SAE1022-1Y
b Modified form with –100

hollowed surface Angle of rotation ϕ
c Electrical circuit
diagram for form b
Simple AMR travel sensor for the mm range position sensor for scanning a gravity pen-
Similarly, the simplest travel sensors may dulum tuned to a low frequency, which
be built for a measuring range of typically may be used to determine the orientation
just a few millimeters using individual of the vehicle (tilt sensor).
differential barber pole sensors (Fig. 28).
However, to achieve good accuracy a con- Angular-position sensor for 180°
stant distance (air gap) is required be- Pseudo Hall type magnetoresistive angu-
tween sensor and moving magnets, which lar-position sensors utilize the practically
embodies the position s to be measured. 100 % sinusoidal signal that is picked off
The temperature sensitivity of the charac- from the output terminals of the four-pole,
teristic curve gradient can be very easily sensor structure, whereby two complete
compensated for by means of an additional periods of the electrical output signal cor-
metallic thin-film sensor on the same car- respond to a mechanical rotation of the
rier substrate, which will have around the magnets through 360°. Using a second ele-
same temperature coefficient, but with the ment, which has been turned through 45°,
inverse sign (e.g. Pt, Ti, Ni). a cosine signal is also generated (Fig. 30).
By applying the arc-tangent function for
High-precision multisensor (POMUX) for instance, the relationship between the
travel ranges > 10 mm two signal voltages can be applied to
There is an intelligent multisensor for po- determine the measured angle using a
sition measurement built up with simple microcontroller or an ASIC. This applies
barber poles also known under the com- throughout a range of 180°, and is for the
mercial name of POMUX (position multi- most part independent of temperature
plexer). It exploits the fact that the symme- changes and magnetic-field fluctuations
try point or zero point of the characteristic (distance, aging).
curve is absolutely independent of the air A further prerequisite for the high accu-
gap and the temperature. Its use in a motor racy of this sensor principle is that the
vehicle has often been considered (e.g. field at both bridges is at least in the same
shock absorber stroke measurement), as it direction (above a given magnitude, field
has excellent characteristics, in particular, strength is irrelevant). This can only be
extreme accuracy. It has also already been guaranteed when both bridges are directly
illustrated in a circular form as an angular- above one another. A design was drawn up
28 Displacement measurement with single differential barber pole
500 μm
5 μm
Ua
MS Ua
MS
I
I U0/2
s 1 to 2 % U0
U0 H
N
SAE1062E
0
S
H 0
Measured travel
in which the two bridges, which are at 45° Angular-position sensor for 360°
to each other, were interwoven so that they One serious drawback with AMR angular-
can be regarded as being at the same point position sensors is their natural limitation
and quasi “on top of one another” (Fig. 29). to an unambiguously detectable angular
range of 0° to 180°. This restriction may,
however, be overcome by modulation of
29 Nested AMR bridges the direction of the magnetic flux by
means of an alternating auxiliary magnetic
field (Fig. 31). By contrast with the COS
and SIN sensor signals which have a 180°
periodicity, these modulation signals have
a 360° periodicity. In this way it is possible,
simply by evaluating the sign of the modu-
lation signals, to distinguish between
31 AMR360 angular-position sensor with

Fig. 29 field direction modulation
“Nested” design of
two AMR bridges offset a UDD
NAE0866Y
from each other by 45°
Fig. 30
BH
BH
Measuring range
sin
U
up to 180° Ucos
BH
a Measuring principle
30 Magnetoresistive angular-position sensor
BH Bext
b Output signals (pseudo Hall)
1 Thin NiFe film w
(AMR sensor)
2 Rotatable
USS
permanent magnet b
a 1 2
with control B
Usin
Voltage
induction B
IV Supply current
UH1,UH2 Measurement Ucos
UH1 UH2
voltages
w Angle of rotation IV c
IV DUsin
45
Voltage
Fig. 31
a Bridge circuit w
b Output signals from DUcos
the Wheatstone
bridges UH1 = U0 . sin2w; UH2 = U0 . cos2w d
c Influence of the
Signal curve
b UH1
auxiliary field on
the signals 2w
UH2 45
d Output signal from p
2 0 UH2
the evaluation UH1 90
p 0 90 180 270 360
circuit 180 Angle of rotation w
BH Control induction
45 Usin
UDD, USS Supply 135 a = 1 · arctan
UAE0642-3E
2 Ucos
UAE1069Y
voltages
Angle of rotation w a0 – 180 : DUcos < 0
USIN, UCOS Measurement
a180 – 360 : DUcos > 0
voltages
w Angle of rotation
ranges and thus to double the unambigu- The magnetic auxiliary field is generated
ous range to 0° to 360°. Because the auxil- by a planar coil integrated into the AMR
iary field for the two sensor bridges points sensor chip (Fig. 32). The auxiliary field
in different directions, the two modulation in the COS and SIN full bridge points in
signals are phase-shifted. It is sufficient, directions 45° apart in accordance with
to distinguish ranges, merely to determine the wiring arrangement.
the sign of the largest modulation signal in
respect of absolute value. This means that The sensor can only be mounted on the
distinguishing the ranges is not critical and end of a rotating shaft. Thus, for instance,
is still easily possible, even in the case of it may be mounted on the end of the steer-
modulation signals suffering heavy inter- ing spindle for measuring the steering-
ference. wheel angle in systems that do not require
determination of multiple rotations of the
steering spindle (e.g. steering-angle sensor
LWS4).
32 Layout of the AMR360 angular-position sensor
Angular-position sensor in the range over
360° (multiple revolutions)
A dual-configuration of pseudo-Hall angu-
lar-position sensors (180°) can be used to
measure multiple rotations of a rotating
component, a steering spindle, for in-
stance. The two associated permanent
magnets are rotated via a step-up gear
train (Fig. 33). As the two driving smaller
gear wheels, which carry the control mag-
nets, differ by one tooth (number of teeth
m and m+1), their mutual phase difference
UAE1074Y
(difference of the angle of rotation: C - U)

is an unambiguous measure of the abso-
lute angle position w of the steering spin-
dle. The system is designed such that this
33 Arrangement for angular measurement in excess phase difference does not exceed 360° in
of 360° a total of four revolutions of the steering
spindle and so the unambiguity of the
measurement is ensured. Each individual Fig. 33
sensor also offers an indeterminate fine 1 Steering axle
2 Ring gear
2 resolution of the angle of rotation. Using
w 3 Gear wheel with
such a configuration, it is possible for in-
m teeth
stance to determine the complete steering- 4 Gear wheel with
1
wheel range with a resolution of better m + 1 teeth
than 1° (this principle is applied in the 5 Magnets
U LWS3 steering-angle sensor). 6 AMR sensor
c
elements
w Angle of rotation
Giant magnetoresistive sensors (GMR)
of the steering
UAE0704-2Y
GMR sensor technology was developed a column

few years ago and is now finding its first C,U Angle of rotation
5 6 3 4 6 5 applications in sensing angles and speeds of the small gear
of rotation in the automotive sector. wheels
The essential advantages of GMR sensors GMR multilayers

over AMR sensors are the natural 360° The relative change in resistance (GMR ef-
range of unambiguity in angle sensing and fect) is 20 to 30 % in GMR multilayers and
the higher sensitivity to magnetic fields in thus is ten times greater than the AMR ef-
speed of rotation sensing. fect. A hard magnetic layer, which will pro-
Unlike AMR sensors, GMR sensors con- vide a suitable bias field, is added to the
sist not of one magnetic function layer, but GMR multilayer stack to place the working
rather of a complex system of layers. A dis- point in a sensitive region of the character-
tinction is essentially made between two istic curve. The sensitivity of a sensor
systems: on the one hand the GMR multi- structure of this nature is about 0.8 %/mT
layers with typically some 20 sequences of which makes it greater than that of AMR
alternating soft magnetic and nonmagnetic barber pole structures.
individual layers (Fig. 34a). On the other
hand there are the GMR spin valves, com- GMR spin valves
prising antiferromagnetic, ferromagnetic Unlike the GMR multilayers, GMR spin
and nonmagnetic functional layers. The in- valves (Fig. 35) are also suitable for angle
dividual layer thickness are in the 1 to 5 nm sensing. In the latter, the reference magne-
range in both systems, and so embrace just tization necessary for angle sensing is gen-
a few atomic layers. erated, fixing (pinning) the direction of
The electrical resistance of a GMR layer magnetization of one of the ferromagnetic
stack depends on the angle between the layers (FM1) through interaction with an
magnetizations of adjacent ferromagnetic adjacent antiferromagnetic (AFM) layer.
individual layers. It is at its maximum This is hence also known as the “pinned
when they are aligned antiparallel and at layer”. On the other hand, the magnetiza-
their minimum when in parallel alignment tion of the second ferromagnetic layer
(Fig. 34b). (FM2), which is largely magnetically de-
Fig. 34
a Design
b Change in
resistance as a 34 GMR multilayer stack with 35 GMR spin valve layer stack with
function of the characteristic curve characteristic curve
magnetic field
a a
FM Soft magnetic
M1
FM M2
single layer NM
NM Nonmagnetic FM FM2
single layer NM NM
FM magnetic FM1
NM nonmagnetic
FM AFM
Fig. 35 magnetization
a Design b b
% 3
b Change in
%
resistance as a 20
Resistance change
Resistance change
function of the
15
magnetization
angle 10 0
5
FM1 Antiferromagnetic
single layer 0
UAE1070E
UAE1071E
FM2 Ferromagnetic
–5 –3
single layer –80 –40 0 40 mT 0 90 180 270 360
NM Nonmagnetic single Magnetic field Magnetization angle U
layer
coupled by a nonmagnetic interlayer, is Like an AMR angular-position sensor,

freely rotatable using the external mag- a GMR angular-position sensor consists
netic field. This is accordingly known as of two full bridges, of which one supplies
the “free layer”. a cosine-wave signal and the other a sine
The resistance varies with a cosine-wave signal depending on the external field di-
dependency on the angle w between the rection (Fig. 37). The different magnetic
external field direction and the reference reference directions MR required to
direction. The stability of the reference achieve this are generated by local heating
magnetization against the action of the ex- of the individual bridge resistors and cool-
ternal field is decisive for the accuracy of ing in the presence of a magnetic field of
the angular measurement. This stability is suitable orientation. The field direction
significantly enhanced by the use of an ad- can be unambiguously determined over
ditional synthetic antiferromagnet (SAF) the full angular range from 0° to 360° by
(Fig. 36). This consists of two ferromag- an arctan function.
netic layers which are strongly antiferro-
magnetically coupled through a very thin
nonmagnetic interlayer and, because of
their vanishing magnetic net moment, no
longer offer a face for the external field to
attack. The magnetization of one of these
two layers is pinned by the adjacent natu-
ral antiferromagnet (AFM). Both the inter-
action of the pinned layer (PL) with the
natural antiferromagnet (AFM) and the
mutual interaction of the ferromagnetic
layers (PL, RL, FL) are determinant for the 37 GMR angular-position sensor with two full bridges
magnetic behavior of the layer system.

Furthermore, the behavior is also formed a
by the shape and crystal anisotropy of the
individual magnetic layers, as is the case
with the AMR sensor structures.
36 GMR spin valve with synthetic antiferromagnets
FL
200 mm
NML Fig. 37
b Vcc1 Vcc2
a Layout
RL b Circuit
I/3 MR MR MR MR Magnetic reference
SAF NML MR I/1 II/1 II/3
magnetization
+ V1 − + V1 −
PL
MR I/2 I/4 MR MR Fig. 36
AF MR II/2 II/4
FL Free layer
0
UAE1073Y
UAE1072Y
NML Interlayer
GND1 B GND2
a RL Reference layer
PL Pinned layer
AF Antiferromagnet
Wave-propagation sensors In practice, the geometrical distance a be-

In recent years, the development of sen- tween an obstacle and the front of the ve-
sors has concentrated increasingly on sen- hicle is determined using the triangulation
sors that acquire information from the im- method from the results measured (dis-
mediate and more distant surroundings of tance b and c) from two ultrasound heads,
the vehicle. This refers to the distance to mounted at a distance d from one another
other vehicles or traffic and to obstacles (Fig. 40):
(all-round view from the vehicle, Fig. 38),
either for safety reasons or to support ve- Long-range radar sensors (76 GHz)
hicle handling. The vehicle-speed controller fitted is
Ultrasound propagation time methods equipped with a long-range radar sensor
are suitable for vehicle-to-vehicle distance (adaptive cruise control, ACC) for the auto-
monitoring in the near range (0.5 to 5 m), matic detection of vehicles preceding in
with optical propagation time or triangula- the same lane or even of obstacles in the
tion methods using short-wave infrared path of the vehicle which could eventually
light (lidar) for the medium range up to necessitate application of the brakes. The
50 m, and electromagnetic radar for the working frequency is 76 GHz (wavelength
long range (up to 150 m). l ≈ 3.8 mm) permits the relatively compact
construction as needed for automotive ap-
Ultrasound sensors plications. A Gunn oscillator (Gunn diode
According to the echo-sounding process in the cavity resonator) feeds in parallel
(Fig. 39), the sensors transmit ultrasound four adjacently arranged patch antennas,
pulses with a frequency of around 40 kHz, for example, which also serve to receive
and record the time taken for the echo the reflected signals. Referred to the vehi-
pulses to arrive having been reflected back cle axis, a plastic Fresnel lens in front con-
from an obstacle. The distance l from the centrates the transmit and receive beam
send/receive head to the closest obstacle horizontally and vertically.
is determined from the propagation time t The signal reflected on the vehicle in
of the first echo pulse to arrive and the front is delayed in accordance with the
velocity of sound c thus: propagation time. The distance a to the
(8) l = 0.5?t?c vehicle in front and its relative speed Dv
Fig. 38
1 Long-range radar 38 Near-range and long-range distance measurement in the motor vehicle
(77 GHz)
2 Long-range/near-
range infrared
viewer (night 6 4 6
3
vision)
3 Video coverage 2
of the vehicle’s
immediate vicinity 3
(night vision, mid-
1 5
range < 40 m)
4 Near-range radar
24 GHz (near
range < 10m)
UAE0869Y
5 Passenger-
compartment video
6 Ultrasound (near
range < 1.5 m)
can be determined from the signal propa- being off-set from the center. The angle
gation time and/or the phase displacement of the vehicle in front can be determined
between the signal transmitted and the from this (see “Adaptive cruise control”
echo received. The antenna transmit/re- chapter).
ceive characteristics are aligned in four
different directions due to the antennas GPS position and distance traveled
measurement
39 Principle of distance measurement using GPS is the global positioning system using
ultrasound satellites, originally created by the USA for
military purposes, but later increasingly re-
a Medium leased for civil uses. The system was com-
(velocity of sound c)
US transmit- pleted in 1993 with a total of 24 satellites
ter/receiver Propagation time distance s, t (21 operating, three in reserve) orbiting at
Measuring distance l an altitude of 20,183 km. More than 24 sat-
ellites have since become available. They
b are distributed over six circular orbital
Transmitter Receiver
pulse pulse planes taking 12 hours to complete an orbit
and so at least four (but generally up to
eight) satellites are always visible above the
horizon from any point on the earth’s sur-
UFS0043-1E
face. They transmit continuous (digital) sig- Fig. 40

nals on a carrier frequency of 1.57542 GHz. a Distance between
Propagation time t The signals contain the following message, the fender and the
in simple terms: obstacle
b Distance sensor 1
• Satellite identification code
to obstacle
40 Calculation of the distance (triangulation) • Satellite position
c Distance sensor 2
• Message transmit time (including date)
to obstacle
d Distance sensor 1
(d 2 + c 2 – b 2 ) 2
a= c2 – The satellites each have two cesium and to sensor 1
4d 2
two rubidium clocks which vary by less 1 Transmit/receive
than 20 to 30 ns on board for the high-pre- sensor

3 2 Receive sensor
3 Obstacle
41 Position determination by GPS
Fig. 41
Representation
in the plane (two-
c a b dimensional):
If the position of the
B satellites is known,
with the propagation
times t1 and t2
t1 measured, the possible
2 1
reception locations are
t2 on two circles around
A
the satellites which
UKD0085-3Y
intersect at points A
SAE1024-1Y
and B. The point A

d located on the earth’s
surface is the location
of the receiver.
cision determination of the transmission the sky. Since complete release for civil
time. The signal propagation times to be use (May 2000), the accuracies obtainable
evaluated lie typically in the region of ap- are about 3 to 5 m in a plane and around
proximately 70 ms. The satellites identify 10 to 20 m in the determination of altitude.
themselves with a pseudo random code With the refinement of GPS to DGPS (Dif-
(PRN number), which is 1,023 bits long ferential GPS, mobile stations have contact
and is continuously repeated after every with a fixed reference station via long
1 ms. It is imposed on the carrier as a wave signals) this variance can be signifi-
phase modulation. cantly further reduced. This makes resolu-
A user can take this information and de- tions down to less than 1 m possible today.
termine his own geographical position in
three dimensions with a GPS receiver, con- These accuracies are the more astonishing
sisting of a GPS antenna, signal receiver, when the manifold sources of error in this
precision clock and microprocessor. Theo- very complex measurement is taken into
retically, the signals from three satellites consideration. For example, simply the
are sufficient for this; since, however, the fact that at the very high velocity of ap-
less complex clocks in mobile receivers proximately 12,000 km/h at which the
(oscillating crystals) can differ from the satellites move, relativistic effects already
satellite clocks somewhat, a fourth satellite play a not insignificant role. Time passes
must be used in order to eliminate the more slowly at high velocities than it does
errors generated by the receiver clock. on earth. Time passes more quickly in
The distance to the visible satellites is de- weaker gravitational fields as are found in
rived from the signal propagation times the orbital planes. The high velocity causes
measured. a dilation of time of 7.2 µs per day, the
The position is determined – as shown lower gravitation even has an effect that
in Figure 40 in only two dimensions of is around six times greater, although
space – then using a trilateration method, opposing.
according to which there is precisely one If the position data determined in a
point in space which satisfies the three di- close sequence is stored in a mobile GPS
mensional conditions. In addition, there is receiver, such as in a motor vehicle, it is
also the knowledge that the motor vehicle, easy to derive the orbit and the speed of
unlike an airplane, must be located on the the vehicle. If the data is sent via a radio
earth’s surface. Measurement is thus the module to a fixed central location, the lo-
more accurate the more satellites can be cation of the vehicle at any time will then
received and the better, i.e. the more uni- be known.
formly the satellites are distributed across
Sensor measuring principles | Speed and rpm sensors 195
Speed and rpm sensors Relative rpm and speed measurement

Examples of relative rotational speed are
Measured variables as follows:
Speed and rpm sensors measure the num- • Crankshaft and camshaft rotational speed
ber of revolutions or the distance traveled • Wheel speeds (for ABS/TCS/ESP)
per unit of time. When automotive applica- • Rotational speed of the diesel fuel-injec-
tions are concerned, these are in both tion pump
cases measured variables which occur
between two components or with respect Here, measurement usually makes use of
to the road surface or another vehicle. an incremental pick-up system comprised
However, it is also necessary to measure of a gear wheel and a speed sensor.
the absolute rotational speed in space or In the detection of relative yaw rate, de-
about the vehicle axes (yaw rate). For in- pending upon the number and size of the
stance, for the electronic stability program, scanned rotational rotor markings, a dis-
the yaw rate about the vehicle’s vertical (or tinction is made between the following
yaw) axis must be determined by sensing. types of sensor (Fig. 1):
• Increment sensor with closely spaced
Measuring principles rotational markings. Up to a certain
Conventional sensors used for rotational- point, this form of sensor permits in-
speed measurement are based on pro- stantaneous speed to be measured at
nounced measuring effects (e.g. inductive). points on the circumference, or the reg-
They are therefore for the most part elec- istration of very fine angular divisions.
trically passive. That is, they are usually • Segment sensor, with only a small num-
not provided with any form of local/on-site ber of scanned rotational segments
electronics. With the newer sensors, how- (for instance, equivalent to the number
ever, measurement is based on less-proof engine cylinders).
nounced measuring effects (for instance, • Simple speed sensor, with only a single
the Hall effect), and these sensors thus scanned marking per revolution, so that
need local, integrated electronics for signal only the average rotational speed can be
conditioning. They belong in the broadest registered.
sense to the category of “intelligent sen-
sors” (which are also often referred to as
“active” sensors).
In fact, the sensors used for measuring
absolute rotating speed (yaw rate) need
highly complex electronic circuitry di-
rectly at the sensor since the measuring
effects used here are not only particularly
small, but also require complex signal con- 1 Determination of the rotational speed (rotor forms)
ditioning.
Incremental rotational-speed measure-
ment takes advantage of a wide variety of
a b c
different physical effects (some of which
can be applied in sensors at a very reason-
ably price). Optical and capacitive sensors,
SAE0780-2Y
though, are highly unsuitable for the rough Fig. 1

operating conditions encountered in the a Incremental sensor
motor vehicle. Here, magnetic-effect sen- b Segment sensor
sors are used almost exclusively. c Speed sensor
196 Sensor measuring principles | Speed and rpm sensors
Sensor shapes In many respects, the traditional inductive-

The following shapes of sensor are presently type sensors are highly unsatisfactory.
in use (Fig. 2) for speed and rpm sensors: Their output signal features an rpm-depen-
• Rod sensor dent amplitude, and for this reason they are
• Fork-shaped sensor unsuitable for measuring low rotational
• (Internal and external) toroidal sensor speeds. In addition, they only permit com-
paratively low-level air-gap tolerances and
The preferred form is the rod sensor, in are usually unable to differentiate between
which the rotor teeth approach the sensor air-gap fluctuations and genuine rotational-
and which is also the easiest to install. The speed pulses. And at least the sensor tip
fork-shaped sensor is permissible in some must resist high temperatures when in-
cases, and is also in service in the field. stalled near hot components (e.g. the
This sensor must be roughly aligned to brakes). These disadvantages are behind
the rotor when installed. The sensor type the additional features aimed at with new,
in which the sensor surrounds the rotor innovative sensors:
shaft in the form of a ring is practically no • Static detection (that is, at zero speed,
longer used. From time to time, an inner- or at extremely low cranking or wheel
ring sensor is used which is introduced speeds)
into the end of a hollow shaft featuring an • Larger air gap (non-aligned mounting
incremental internal structure. with air gaps > 0)

Unfortunately, although it is the most • Small size
widely used sensor type, the rod-shaped • Independence from air-gap fluctuations
sensor features the lowest measurement • Temperature stability (≤ 200 °C)
sensitivity and is problematical when air • Detection of the sense of direction
gaps become excessive. The fork-shaped (optional for navigation)

sensor on the other hand is practically im- • Reference-mark identification (ignition)
pervious to axial and radial play. Regard-

ing toroidal sensors, the most widely used Magnetostatic sensors (Hall, magnetore-
form combines a large measurement signal sistors, AMR) are highly suitable for com-
with pronounced insensitivity to geometric plying with the first two demands. And, as
tolerances. a rule, they also permit compliance with
the second and third stipulations.
Figure 3 shows three fundamentally suit-

2 Sensor shapes able sensor shapes, which generally are in-
sensitive to air-gap fluctuations. Here, one
must differentiate between sensors which
a b sense radially and those which sense tan-
gentially. This means that independent of
the air gap, magnetostatic sensors are always
dL
able to differentiate between the north and

south poles of a magnetically active pole
wheel or rotor ring.
Fig. 2 In the case of magnetically passive ro-
a Fork shape
tors, the sign of the output signal is then
(vane principle)
SAE0778-2Y
b Rod shape
no longer independent of the air gap when
(proximity they register the tangential-field strength
principle) (here though, the fact that the air gap is
dL Air gap
often enlarged due to the sensor is a disad- As a rule, presuming the same increment
vantage). width and output signal, the pole wheel’s
Radially measuring differential-field or intrinsic magnetism (a pole wheel is de-
gradient sensors are often used. These al- fined as a magnetically active rotor) per-
ways register only the gradients of the ra- mits a considerably larger air gap.
dial-field components, the signs of which
do not change with the air gap but only Normally, passive rotors are in the form of
with the angle of rotation. gear wheels. In many cases, these are al-
ready present on the engine (for instance
Rotors the starter ring gear). Otherwise they must
The rotor is of decisive importance when be installed in a specific position in order
measuring rotational speed. It is usually to generate the required signal (as needed
provided by the vehicle manufacturer, for instance for ABS). In the latter case,
while the sensor itself comes from a com- both planar-toothed and axial-scan ver-
ponent supplier. Previously, magnetically sions are in use.
passive rotors have been used almost ex- An easily identifiable reference mark is
clusively. These are made from magneti- required for each revolution when pick-off
cally soft materials (usually iron), and are the crankshaft speed or position (for in-
less expensive than magnetically hard ro- stance, at the starter ring gear). This refer-
tors (also known as pole wheels). Apart ence mark must allow a fault-free and
from that, since they are not magnetized optimum timing of ignition and injection.
they are easier to handle, and there is no The reference mark can be in the form of
danger of mutual demagnetization (for in- a completely (or partially) removed tooth.
stance during storage). Unfortunately, this Due to the fact that a tooth space “takes
form of rotor is the most difficult to scan, more time”, the reference mark is immedi-
particularly in combination with rod sensors. ately identifiable, particularly since the
engine speed can only vary gradually and
3 Sensors insensitive to air-gap fluctuations never abruptly.
In addition to gear wheels, stamped per-
forated disks or wave-shaped metal rings
have been introduced in the meantime as
S N
N S
low-priced rotors (ABS).

a
N
N
S
S
N
S N N It was the integration of ABS sensors in the
S N S
vehicle’s wheel bearing which led to the in-
troduction of pole wheels some of which
also assume the role of a sealing ring (plas-
b
tic-bound magnetic powder). Small, for the
most part encapsulated, tachometer sen-
sors connected through a short flexible
N
S shaft with one of the vehicle’s wheels, also
use pole wheels (with only very few poles)
c for generating a speed signal. These are
usually picked off by means of integrated
Fig. 3
Hall sensors.
a Radial-field sensor
SAE0779-2Y
N with pole wheel

S
b Tangential sensor
c Differential sensor
with gear wheel
Inductive-type sensors generation of voltage pulses which are

Inductive-type sensors were available on either impossible, or at least difficult, to
the market as coil versions long before the separate from the genuine engine-speed
first suitable microstructure sensor ver- signals. Since the flux varies exponentially
sions (e.g. using the Hall principle) became along with the change in air gap, and the
available. Such inductive-type sensors use air-gap fluctuations are often of the high-
Faraday’s law when measuring rotational frequency type (e.g. brake flutter), these
speed. In other words, they generate a unwanted pulses can feature a high voltage
voltage UA at their two-pole output which amplitude.
is proportional to the change (referred to Inductive-type sensors are therefore
time) of a magnetic flux F (w = number of always of the dynamic type, and being as
turns). their output signal tends to fall to zero in
such cases, they are in principle also un-
UA = Uind = w ? dF/dt suitable for the detection of extremely low
speeds (quasi-static or static). The only ex-
The magnetic flux F is also a function of ception here are the coil sensors which are
the rotational position x and the air gap dL: powered by a carrier frequency and which
where are based on the eddy-current or damping
F = F(x, dL) and dL = const principle. These though are hardly ever
the following applies: used in automotive applications.
UA = Uind = w ? −F/−x ? dx/dt
Basically, inductive speed sensors are
where dx/dt represents the (rotational) comprised of three important magnetic
speed to be measured. component parts (Fig. 4):
• Fixed coil
The inductive-type sensors’ weak point is • Soft-iron component
underlined though by the equation: if it is • Permanent-magnet component
impossible to keep the air gap dL constant
(due to flutter or other forms of mechani- The change in flux needed for the genera-
cal play), the air-gap fluctuations induce tion of the output voltage results from the
the same change of flux as does a fluctua- motion or rotation of the hard or soft-iron
tion of speed. This effect can cause the magnetic component. The inductive-type
sensors are preferably composed of a
4 Inductive speed sensor rod magnet (1) with a soft-magnetic pole
pin (2) carrying the 2-connection induc-
tion coil (3). When a ferromagnetic gear
wheel (5), or some form of similar rotor,
rotates in front of this sensor (pick-up) the
changing magnetic flux (which varies as a
Fig. 4
1 Rod magnet function of time) induces a practically si-
S nusoidal voltage.
2 Soft-magnetic
1
pole pin For scanning very fine tooth structures,
3 Induction coil N the end of the pole pin is sometimes pointed
4 Air gap dL
2 and acts as a flux-concentrating element.
UMZ0138-4Y
5 Ferromagnetic gear-
3 In other words it is shaped like a pole “blade”
wheel (or rotor or
pulse wheel)
l 4 which usually projects through the metal
6 Rotational or
6 5 or plastic housing and is adapted to the in-
reference mark crement structure regarding shape and di-
l Tooth interval rection.
The rotor can be provided with one or detection of such a rotational or reference
more rotational markings (6). Figure 5 mark. According to Faraday’s Law, in all
shows the flux curve and the voltage in- phases the signal’s amplitude is propor-
duced by a single rotational or reference tional to rotational speed.
mark (slot, cam, or pole pin). In order to ensure adequate, interfer-
Normally, the steep passage through ence-free evaluation in the control unit,
zero which takes place at the mid-point of the spacing between the peaks of a double
maximum flux is utilized for the electronic pulse (or of a periodic voltage pulse) USS
5 Flux and voltage curve for an inductive-type sensor
Magnetically passive Magnetically active

Individual brands Individual brands
a
1 N 8
L
2 8 S
3 4
U~
3
Schmitt
Schmitt
trigger
trigger
output
output
5 6
Angle of rotation w Angle of rotation w
Gear wheel Pole wheel Fig. 5

b a Single magnetic
marking per
revolution
N
N N
L
b Periodic increment
structure (e.g. gear
S S
S
wheel, pole wheel)

7 g
3 2 c Evaluation circuit
U~
3 1 Cam
2 Slot
3 Switching point
4 Steep passage
Schmitt
trigger
output
through zero
evaluated
5 Priming edge
Angle of rotation w Angle of rotation w 6 Switching flank
7 Shifting point
8 Pole pin
Schmitt 9 Sensor
c 9
UAE0782-1E
trigger
g/2 Phase shift
L Inductance
U Voltage
F Magnetic flux
should be at least 30 mV. The major disad- The reference point is detected when the
vantage of the inductive-type sensors is distance between the passages through
the fact that at high speeds/rotational zero changes abruptly and causes a far
speeds their output voltages can reach higher signal voltage (corresponding to
levels far in excess of 100 V which are dif- an apparently larger tooth) which has a
ficult to process electronically. negative effect upon the previous and
If Zener diodes are used to clip the high upon the subsequent incremental voltage –
voltage peaks, the resulting changes in the this can under certain circumstances be
sensor’s load impedance lead to considerable undesirable.
phase-angle errors. With camshaft and
crankshaft sensors this can have highly Advantages
undesirable results with regard to the igni- • Low manufacturing costs
tion where the correctness of phase rela- • High immunity to interference: low
tionship must be better than approx. 0.2°. static internal resistance (dynamically
Normally, the prepulse generated by the higher)
magnetic return field can be ignored at low • No local electronics (electrically pas-
speeds. With some magnetically passive or sive), that need to be protected

active rotational markings though, at high • No problems with DC voltage drift
speeds the prepulse voltage can increase to (dynamic measuring principle)

such an extent that it exceeds the threshold • Broad temperature range (limited
value of the downstream threshold dis- primarily by the sealing compound

criminator and can cause an even greater characteristics)
error (Fig. 5a). For this reason, the thresh-
old values of the control unit input circuit Disadvantages
are adapted dynamically to the speed in • Conventional coil technology imposes
question. limits on size reduction
Provided that the air gap is not too nar- • The output signal is rpm-dependent,
row, a uniform tooth structure results in unsuitable for quasi-static movements

the practically sinusoidal voltage curve • Sensitive to air-gap fluctuations
shown in Figure 5b. The speed is reflected

in the periodic interval between the volt- Examples of application
age’s zero transition points, while the am- • Inductive engine-speed sensor (crank-
plitude is also proportional to the rotating shaft-speed sensor)
speed. • Inductive wheel-speed sensor
The signal amplitude is highly dependent • Inductive camshaft sensor (transistor-
(exponentially) upon air gap and tooth size. ized ignition with induction-type pulse
As is the case with all magnetic increment generator TZ-I)
processes, up to air gaps of dL, teeth can be • Needle-motion sensor (diesel-fuel
efficiently detected as from half or 1/3 of a injection)

tooth interval I.
dL ≤ l/(2…3)
The conventional ABS gear wheels and

crankshaft applications cover air gaps of
up to 0.8 and 1.5 mm respectively. The ref-
erence mark needed for the ignition and
injection results from leaving out one or
two teeth or by closing a tooth space.
Magnetostatic sensors This “Hall IC” (using bipolar technology,

Quasi-static speed measurement is best for sustained temperatures ≤ 150 °C, and
performed using magnetostatic sensors. for direct connection into the vehicle’s
Their output signal is independent of rota- onboard electrical supply), is located in
tional speed and depends solely on field a practically closed magnetic circuit com-
strength, so that at high rotational speeds prising permanent magnet and pole pieces
their low signal voltages are easier to han- (Fig. 6). A soft-magnetic trigger wheel
dle electronically. Furthermore, they not (driven, for instance, by the camshaft)
only have the advantage of imposing prac- rotates through the remaining air gap.
tically no limits with regard to miniaturiza- With a trigger-wheel vane in the air gap,
tion, but also signal amplification and/or the magnetic flux is short-circuited past
signal preprocessing can be integrated the sensor element. On the other hand
locally. The fact that they are very small though, when a trigger-wheel opening
means that multiple systems such as dif- passes through the air gap the flux is un-
ferential configurations, or arrangements hindered on its way to the sensor. This
with integrated recognition of direction, principle ensures that the sensor also per-
present no difficulties. forms perfectly when the trigger-wheel
On the other hand, such active sensors varies in how far it penetrates into the air
have a serious disadvantage, and that is gap, or in case the air-gap shifts radially,
the fact that their operating-temperature that is vertically to the direction of rotation.
range is for the most part defined by the Since Hall vane switches of this type fea-
relevant Si (silicon) evaluation electronics ture limited rotational resolution they are
which as a rule cannot withstand such high mainly used as segment sensors. If the vane
temperatures as the sensor element itself. slot is too narrow, it is practically impossible
For some time now, active sensors have for the magnetic field to pass through, with
been available with the option of a two- the result that the required induction jump
pole current output, so that the two-core is not generated.
connection can no longer be regarded as
a specific advantage of the inductive-type
coil sensor.
6 Hall vane switch
Hall vane switches

If Hall Si sensors are to be used for incre-
a 2
mental rpm measurement, pronounced
manufacturing scatter together with the
effects of temperature mean that they must
be provided with an adequate induction S N
jump of typically 40 to 50 mT in order to
ensure reliable high-speed switching. With
Fig. 6
conventional Hall sensors and acceptable b 2 3 4 1 a Magnetic flux:
air gaps, this was only possible with the unhindered
sensor in the form of a “Hall vane switch” b Magnetic flux:
(for instance when used as ignition-trig- short-circuited
S N 1 Vane width b
gering sensors in the ignition distributor US
2 Soft-magnetic
for earlier electronic ignition systems).
conductive element
UAE0639-2Y
The sensor and the electronic circuits 3 Hall IC

for supply and signal evaluation are U0 b 4 Air gap
integrated on the sensor chip. U0 Supply voltage
US Sensor voltage
Simple Hall rod sensors (B = 0) remains unchanged. Considering

In contrast to the Hall vane switch, the that the working point on new types of
working-point magnetization depends too Hall sensors is to a great extent thermally
much upon the width of the air gap, and stable, the switching points of the down-
the induction excursion is too small for stream threshold comparator can be set
reliable switching with this configuration. relatively close together. This means that
Simple rod-shaped Hall sensors provided relatively wide air gaps become possible.
with a working-point magnet are therefore Air-gap fluctuations in this configuration
unsuitable for static or quasi-static scanning cannot cause missing pulses as they do not
of a magnetically passive rotor (Fig. 7a, gear lead to pole change. The pole change only
wheel). The switching point of a downstream characterizes the continuing measurement
threshold comparator (Schmitt trigger) motion (rotation).
would have to be continually adapted to
the varying working point. Such applica- Gradient sensors
tions are only feasible when DC coupling is When it comes to scanning passive rotors,
dispensed with and therefore also static gradient sensors (Fig. 8) designed on the
signal evaluation. The coupling capacitors basis of differential Hall sensors or differ-
required for such configurations (also ential magnetoresistive sensors, are far
known as DF sensors) involve high costs better than simple Hall sensors. They are
though and lead to reduced operational provided with a permanent magnet. This
reliability. pole face of this magnet facing the gear
However, simple Hall sensors are highly wheel is homogenized by a thin ferromag-
suitable for scanning a magnetically active netic wafer (2). Two galvanomagnetic ele-
rotor (Fig. 7b, pole wheel). The sensor is ments (generic term for Hall sensors and
triggered by the pole wheel with varying magnetoresistors) are located on each ele-
polarity only in the vicinity of the magnetic ment’s sensor tip, at a distance of roughly
zero point. Although the magnetic control one half a tooth. Thus, one of the elements
excursion reduces along with increasing is always opposite a tooth space when the
air gap, the position of the working point other is opposite a tooth. The sensor mea-
Fig. 7
a Passive rotor
7 Hall rod sensors 8 Gradient sensor for gear wheel scanning
b Active rotor
1 Incremental rotor
a
2 Simple Hall sensor
3 Permanent magnet
4 Pole wheel w 1
5 Housing
w
2
Fig. 8 N 4
3 1
1 Magnetoresistors
S 2
R1, R2 or Hall S
5 3
elements H1, H2 N
2 Ferromagnetic wafer
(soft-magnetic
b w
substrate) S
3 Permanent magnet N N
S 1
4 Gear wheel R2 R1
UAE0022Y-3
UAE0783-1Y
U0 Supply voltage 4 2 U0
UA(w) Measuring UA(w)
voltage for an angle 5
of rotation w
sures the difference in field strength be- given point. Local amplification is neces-
tween two neighboring points on the cir- sary, even though their measuring effect is
cumference of the rotor. The output signal approximately 1 to 2 orders of magnitude
corresponds roughly to the field strength larger than that of the silicon Hall sensors.
derived as a function of the angle at the In the case of the crankshaft speed sen-
circumference, and its sign is therefore in- sor integrated in the bearing (composite
dependent of the air gap. Being as they do seal with sensor), the AMR thin-film sen-
not alter the gradient signal’s sign, air-gap sor is mounted together with an evaluation
fluctuations does not cause missing pulses. IC on a common leadframe.
For signal evaluation, it is a simple mat- For the purposes of space saving and
ter to connect the two magnetoresistors temperature protection, the evaluation IC
as a voltage divider which is supplied by is bent at an angle of 90° and also located
a constant voltage and whose (usually un- further away from the sensor tip.
loaded) output signal is registered by the
control unit. At room temperatures and Giant magnetoresistive (GMR) elements
with the customary air gap, this signal is in The giant magneto resistance effect (GMR
the volts range, and even at high tempera- effect) was first demonstrated on multilay-
tures it is suitable for transmission to the ers consisting alternately of thin soft mag-
control unit without any form of preampli- netic and nonmagnetic layers (thickness of
fication. a few nanometers). An antiparallel align-
Provided appropriate circuitry is used, ment of the magnetization of adjacent fer-
measuring the loaded output current of romagnetic layers leads to the maximum
the magnetoresister divider instead of its electrical resistance, while parallel align-
open-circuit voltage permits the sensor’s ment shows a significantly lower resistance.
pronounced temperature measurement The relative change in resistance for appli-
sensitivity to be compensated for to a great cation-relevant layer structures thus var-
extent. ies in the range 20 to 30 %. This represents
In the case of a gradient sensor based on an increase over the AMR effect by a factor
the Hall effect, the current paths of both of around 10.
Hall elements can be connected in parallel,
and their opposite-polarity output voltages
in series, so that their differential voltage 9 AMR speed sensor in the form of a
can be picked off directly and inputted to tangential-field sensor
the downstream amplification and evalua- Fig. 9
tion stages. 1 Gear wheel (Fe)
w 2 Permanent magnet
1
Tangential sensors 3 Sensor
B Control field
Tangential sensors differ from gradient
strength with
sensors by their reaction to variations in
tangential
polarity and intensity in the components B Br B B’
components Bt and
of a magnetic field located tangentially to radial component
the periphery of the rotor. Using AMR S Br (B' off position,
2
thin-film techniques, tangential sensors Bt = 0),
R1, R2 permalloy
are available as barber-pole or permalloy 3
thin-film resistors
resistor types in full-bridge or half-bridge R1 R2
(AMR)
UAE0643-2Y
circuits (Fig. 9). In contrast to gradient Bt w Angle of rotation

U0 UA
sensors, they need not be matched to the U0 Supply voltage
particular tooth pitch of the rotor and can N UA Measurement
in fact be designed to sense practically at a voltage
The simplest layer system that shows a A bridge circuit of four GMR resistance
GMR effect is a spin valve (see also “Posi- elements, designed as a gradiometer is
tion sensors” section). It consists of two suited for the determination of rotational
soft magnetic layers, separated by a non- speed (Fig. 12). The resistors are intercon-
magnetic layer. While the magnetization nected here in such a way that only one
of the one soft magnetic layer is fixed by magnetic field difference at the two bridge
an antiferromagnet, the magnetization of locations results in a signal. The effects
the second soft magnetic layer can ideally of homogenous interference fields can be
follow the external field without interfer- suppressed in this way. As the reference
ence (free layer). The resistance is then direction for all resistors is identical, local
minimal if the magnetizations of the two heating in the magnetic field, as is neces-
layers are aligned in parallel and is in- sary to write differing reference magne-
creased by approximately 5 % if the mag- tization directions in an angular-position
netizations are aligned antiparallel. The sensor, is not required.
interaction of the reference magnetization
with the external field may be minimized
with the help of a synthetic antiferromag- 11 Characteristic curve for a GMR spin valve
net (SAF), comprising two ferromagnetic

layers, that are strongly antiferromag- %
netically coupled by means of very thin
nonmagnetic interlayer (Fig. 10). This sig- 5
nificantly increases the robustness against
Resistance change
interference fields. 4
3
Figure 11 shows a typical characteristic
curve for a spin valve. The two charac- 2
teristic states for the two opposing direc-
tions of the external magnetic field can 1
clearly be distinguished. The spin valve is
UAE1065Y
sensitive to field strength changes in the 0
–120 – 80 – 40 0 40 80 mT
transitional range between high and low m0H
resistance.
10 Layer stack for a GMR spin valve 12 Bridge circuit for the GMR sensor
a Uv
1
R1 R3
Fig. 10
1 Free layer 2 UBr
2 Interlayer
3 Reference layer
3 SAF R2 R4
4 Pinned layer 2
5 Antiferromagnet 4
b
SAF Synthetic
antiferromagnet
5
UAE1064Y
UAE1066Y
Fig. 12
a Electrical wiring
b Layout of a
gradiometer sensor
The maximum signal amplitude for a gradi- are out of the question for automotive ap-
ometer sensor then results when the direc- plications due to the high costs involved.
tion of the field at the two bridge locations On the other hand, the slightly less-severe
is precisely 180° phase-shifted. This is the requirements on precision for applications
case on a multipole wheel, for instance, in the motor vehicle can be complied with
if the distance between the resistance ele- by gyroscopes manufactured using fine-
ments corresponds to the distance between mechanical and micromechanical pro-
poles in the multipole ring (Fig. 13). cesses. Instead of a rotational movement,
these units use an equivalent, elastic, os-
Examples of application cillatory movement to generate the mea-
• Hall sensor (transistorized ignition TZ-H) suring effect. In principle, these sensors
• Hall phase sensor (camshaft) are similar to mechanical gyroscopes.
• Transmission Hall sensor (RS50, RS51) They exploit the Coriolis accelerations
• Active Hall speed sensor which are generated when rotary motions
• Active AMR speed sensor are coupled with an oscillatory component
• Magnetoresistive sensor (for diesel (velocity v) (Fig. 14). These sensors are
radial-piston distributor pumps) known as oscillation gyroscopes and are
• Active GMR speed sensor used in the electronic stability program
(ESP) for the detection of the vehicle
Absolute rotating-speed measurement breaking away, in navigation systems for
Measuring principle of the oscillation the detection of the direction of travel and
gyroscope in cameras to stabilize the image against
Basic principles wobbling. They also comply adequately with
Mechanical gyroscopes (gyros for short) all automotive stipulations such as freedom
utilize inertial forces in precisely measur- from maintenance, long service life, switch-
ing angular movements in space indepen- on time constant etc., not to forget the ques-
dent of the reference system. Notwith- tion of costs.
standing their pronounced measuring ef- Oscillation gyroscopes measure the
fect, rotating gyroscopes as well as optical yaw rate around their “sensitive” axis.
sensors based on the interferometric The yaw rate Vz around the vehicle’s Fig. 13
Sagnac effect (laser and fiber gyroscopes), vertical axis (yaw axis) is measured in Field lines in the space
surrounding a multipole
ring
The two squares
13 Speed of rotation measurement with GMR sensor 14 Relationships in the rotary motion
symbolize the two
bridge locations
The arrows indicate the
direction of the field
z
lines at these points
Fig. 14
Ωz
Interrelationship
between yaw rate,
y oscillation, and Coriolis
acceleration at a
N yy
S lumped mass m.
S m
Vz Yaw rate
acoriolis vy Speed of the
SAE0781-3Y
x
UAE1067Y
oscillation
aCoriolis Coriolis
acceleration
m Lumped mass
the electronic stability program (ESP) and The Coriolis force acts in accordance with
in the navigation system. Systems for trig- a known vector law perpendicular to the
gering rollover protection systems need rotary motion and speed of the mass in
data on the yaw rates Vx and Vy around the motion (Fig. 14). This gives the Coriolis
vehicle’s pitch and roll axes. acceleration as:
(1) aCoriolis = ax = 2?vy?x?Vz
Measurement of the Coriolis acceleration
Oscillation gyroscopes measure the Whereby, the velocity vy changes sinusoi-
Coriolis acceleration in the following dally in accordance with the oscillatory
way: as the distance between a mass and movement:
the axis of rotation increases, the mass (2) vy = vy?sinvt
must cover a greater distance in the time
of one revolution because of the greater This means that, at a constant yaw rate Vz,
radius. Circumference increases with the a sinusoidal Coriolis acceleration aCoriolis of
radius. The mass must, therefore, acquire the same frequency is measured. The am-
a higher tangential speed and must be plitude is then:
accelerated to achieve this. This acceler- (3) aCoriolis = 2?vy?Vz
ation (Coriolis acceleration) is measured
by suspending the mass on springs and The acceleration ay which is also applied
measuring the bend in the suspension. to the mass m in the oscillatory direction is
If the distance of the mass, now rotating usually several orders of magnitude higher
at a higher tangential speed, from the than the useful Coriolis acceleration.
axis of rotation is now reduced, the mass (4) ay = dvy/dt = v?vy?cosvt
must now become slower, as it needs
only cover a shorter distance for each As the Coriolis acceleration has the same
revolution. The braking deceleration re- frequency as the exciting function, the use-
quired is also measured via the bend of ful signal (the yaw rate) can be obtained by
the suspension. multiplication of the exciting signal and
In the oscillation gyroscopes, the dis- the Coriolis signal with subsequent averag-
tance of one (or more) masses from the ing (phase-correct rectification). Interfer-
axis of rotation is periodically increased ence signals of other frequencies are fil-
and then reduced again by an oscillation tered out (principle of the lock-in ampli-
excitation (Fig. 14, movement in the y di- fier). Averaging (with a low-pass filter)
rection). This means that the mass must be frees the output signal from the excitation
accelerated and then braked again in the frequency. An output voltage is generated
same cycle (movement in the x direction). which is proportional to the yaw rate:
The forces necessary for this depend on (5) UA = const ?aCoriolis = const'?V
the amplitude of the exciting oscillation
and the current yaw rate. If the exciting Examples of application
oscillation is kept constant, the yaw rate V • Piezoelectric yaw-rate sensors
can be determined from the acceleration • Micromechanical yaw-rate sensors
forces.
Sensor measuring principles | Acceleration sensors 207
Acceleration sensors Measuring principles

Acceleration sensors measure the force F
Measured variables executed on an inert mass m by an acceler-
Acceleration sensors are suitable ation a:
• For knock control in gasoline engines
• For triggering restraint systems (e.g. air- (1) F = m?a
bag and seat-belt pretensioners)
• For detecting the accelerations of the Here, just as in force measurement, sys-
vehicle for the antilock brake system tems are available to measure displace-
(ABS) or the electronic stability program ment as well as mechanical strain.
(ESP) or
• For the evaluation of body acceleration Displacement or travel-measuring systems
for use by the chassis and suspension In the case of travel-measuring systems
control systems (Fig. 1), a mass m – the seismic mass – is
elastically attached to the body, the accel-
The measured variable is acceleration a, eration a of which it is intended to mea-
often stated as a multiple of the accelera- sure.
tion due to gravity gn (1 g ≈ 9.81 m/s2) (see
Table 1 for typical values in motor vehicles). Constant acceleration
In the event of a constant acceleration, the
acceleration force is in equilibrium with
the return force applied to the spring with
1 Measuring range of acceleration sensors
spring constant c deflected by x:
Applications Measuring
range
(2) F = m?a = c?x
Knock control 40 g
Passenger protection
– Airbag, seat-belt pretensioner 35 to 100 g The system’s measurement sensitivity S is Table 1
– Side impact, front sensing 100 to 400 g therefore: g Acceleration
– Roll-over detection 3 to 7 g due to gravity
1 g ≈ 9.81 m/s2
ESP, HHC, ABS 0.8 to 1.8 g (3) S = x/a = m/c
HHC Hill hold control
Chassis and suspension control
ABS Antilock brake
– Design 1g This indicates that a large mass together system
– Axle/damper 10 to 20 g with low spring stiffness result in high ESP Electronic stability
Car alarm 1g measurement sensitivity. program
Fig. 1
1 Displacement or travel-measuring acceleration sensors Schematic:
a Excursion-
measuring
a b b Closed-loop
position controlled
a a a Measured
acceleration
m m
x x System excursion
x 0
FM Measuring force
(inertial force on
the mass m)
FM FM FK
SAE0809-2Y
UA ~ x ~ a FK Compensating
IA ~ force
IA Output current
UA Output voltage
208 Sensor measuring principles | Acceleration sensors
Variable acceleration The resulting excursion is converted into

In the dynamic case, a damping force and an electrical signal by a suitable measuring
an inertia force are to be taken into consid- technique (cf. Table 2).
eration in addition to the spring force. The
essential damping force is proportional to Position control
.
the velocity x and is described using the Travel measuring systems allow the appli-
damping coefficient p. The inertia force is cation of the compensation principle, in
..
proportional to the acceleration x . The re- which the system deflection due to accel-
sulting equation (4) describes a resonant eration is controlled out by an equivalent
system: return force (Fig. 1b). The sensor element
is now a component part of a closed con-
. ..
(4) F = m?a = c?x + p?x + m?x trol loop. The return force or the variable
generating it (e.g. current or voltage) acts
Assuming negligible damping (p ≈ 0), this as a measure of the acceleration. The sys-
system has a resonant frequency of: tem operates through closed-loop control
____ very close to the zero point of the excursion
(5) v0 =
√ mc and achieves a high degree of linearity.
Closed loop position control produces
a greater measuring range, which is only
This means that in accordance with equa- limited by the return force and a higher
tion (3), the measuring sensitivity S is di- cutoff frequency than similar, non-motion-
rectly linked to the resonant frequency controlled systems (Fig. 1a).
v0 in the following manner:
Damping
(6) v02?S = 1 To prevent an excessive resonance, which
would cause interference, pure deflection
In other words, sensitivity can be expected systems demand defined damping that is
to be reduced by factor 1/4 when the reso- independent of the temperature. If the
nant frequency is doubled. It is only damping coefficient p is related to the
clearly below their resonant frequency other parameters in equation (4), this
that such spring-mass systems display ad- results in Lehr’s damping factor D:
equate proportionality between measured
p p
variable and deflection. (7) D = ____ ?v0 = ________
2?c 2?√c?m
2 Measuring principles
Electrical Principle Technical implementation Examples of

pick-off application
Piezoresistive Stress measurement via the change Silicon volume microme- Laboratory applica-
in resistance as a result of elongation chanics, thin-layer systems tions, previously also
caused by acceleration on the spring- crash sensors
mass system
Piezoelectric Charge shifts in the crystal as a result of Ceramic material, PZT, Knock sensors, pre-
the force applied (spring-mass system) quartz, PVDF viously also crash
sensors
Capacitive Excursion of the spring-mass system is Silicon volume microme- Used throughout the
determined through capacitance mea- chanics, surface microme- motor vehicle
surement of a plate capacitor chanics
Thermal Lateral excursion of a heated gas zone Volume micromechanics Rollover sensing
and detection of the asymmetry with
Table 2 respect to the heated zone
This dimensionless variable permits a sim- chanics. Because of the capacitive evalua-
ple description and comparison of differ- tion, the measurement signal is deter-
ing resonant systems. Natural vibration mined only by the geometry and stable
and increase in resonance sharpness are material parameters and is barely influ-
largely determined by this damping factor. enced by variables such as temperature.
Values of D = 0.5 to 0.7 are preferred in The very small measuring capacities of
practice (Fig. 2). these sensors, however, requires that eval-
In the case of position-controlled sys- uation electronics are provided in the im-
tems, damping is implemented and ad- mediate vicinity.
justed in the closed control loop. The advantage of surface micromechan-
ics is the opportunity it provides to control
Physical implementation by position or to deflect the systems by ap-
The majority of the acceleration sensors plying electrostatic forces (to the measur-
currently in use in the motor vehicle are ing electrodes or to an additionally fitted
manufactured in silicon surface microme- electrode set). The latter, being a genuine
electro-mechanical self test, is an effective
means of checking the entire signal path.
2 Resonance curves for the transfer function G
Systems for measuring mechanical stress

a 10 Applications
7 Piezoelectric sensors are only used in the
5
4 motor vehicle in acceleration and yaw-rate
3
sensors.
2 D = 0.1
D = 0.3
1 Longitudinal effect principle of action
D = 0.5
0.7 Fig. 2
Value G
0.5 D = 2/2 Under the action of a force F, piezoelectric

a Amplitude
0.4 D= 1 materials generate a charge Q on their sur-
0.3 resonance curve
D = 1.5
0.2 D= 2
faces to which electrodes are attached b Phase resonance
D= 3 (Fig. 3). This charge is proportional to the curve of the
0.1
0.07 mechanical stress generated by the force F. complex transfer
0.05 function G(i?V) =
0.04 [x(i?V)]/[a(i?V)]
0.03
0.02
x(i?V)
0.01 3 Types of piezoelectric effects
Deflection
b 0 F amplitude
D = 0.1 Q a(i?V)
-20 D = 0.3
Acceleration
D = 0.5 a X
Y
-40 D = 2/2
Z
excitement
-60 D= 1 Q amplitude
D = 1.5 F V = v/v0
-80
D= 2 Q Standardized
-100 D = 3
Angle G
circular frequency
b F X
F D Damping
-120 Z
Y
-140
Q
-160 Fig. 3
F
Q a Longitudinal effect
-180
SAE0812-1Y
SAE0810-2E
Y
0.1 0.2 0.4 0.7 1 2 3 4 5 7 10 b Transverse effect
c X
Z
0.3 0.5 c Tangential force
Standardized angular frequency Q F Force

F
Q Charge
210 Sensor measuring principles | Acceleration sensors
Materials for piezoelectric components In practice, these effects occur together.

can be broadly classified into monocrys- The relationship between the charge gen-
tals, such as quartz, and ceramics. erated and the mechanical stress acting
Piezoelectrically active ceramics are pro- can thus be formulated as a tensor equation.
duced by sintering finely ground ferroelec- The transverse effect is used for exam-
tric materials. These ceramics then consist ple in a bimorph. These are composed of
of very many small crystals randomly ori- two oppositely polarized piezoceramics
ented in space. These are then polarized in used for the measurement of bending
production by the application of a high stresses (Fig. 4). When the two-layer
electric field strength. This largely aligns ceramic composite bends, one half is
the randomly oriented polarizations of the stretched (« > 0) and the other compressed
microcrystals. (« < 0). The opposed-polarity of the ce-
Depolarization, with the consequent loss ramic plates means that the resulting part
of the piezoelectric properties, can be voltages U1 and U2 now add to form a total
caused in piezoceramics by temperatures voltage U which can be picked off on the
above the Curie temperature, by intensive two outside metal layers.
mechanical stresses or by polarization with
opposite field strengths. Electrical signal evaluation
The charges generated are discharged Piezoelectric sensor elements may be
through the external resistance of the mea- modeled as a voltage supply with a capaci-
suring circuit or through the piezo sensor’s tive internal resistance. The internal resis-
internal resistance. Such sensors cannot tance is infinitely high for ideal elements.
measure statically, and are only used where Signal processing must be rated with cor-
dynamic measurement is needed. The typi- responding care.
cal limit frequencies of this high pass re- There are two options for matching the
sponse, however, lie above 1 Hz, depending impedance to subsequent systems:
on the application. • Electrometer amplifier (Fig. 5a): in this
Fig. 4 case, the voltage applied across the

a Non-active state Transverse effect and shear effect electrodes is detected and amplified
b When bent, In addition to the longitudinal piezoelec- by an amplifier with extremely high im-
stretched on top tric effect (Fig. 3a), there are also trans- pedance input. Parasitic capacitances
(« > 0)
verse and shear effects (Figs. 3b and 3c).
compressed below
(« < 0)
1 Direction of 4 Piezoelectric bimorph 5 Pick-off on piezoelectric sensors
polarization
F Measuring force F
1 a
U Total voltage F= 0 1
a
U1, U2 Part voltages Q
U1 = 0
2
U2 = 0 U U
U= 0
Fig. 5
a Voltage pick-off
b Charge pick-off CM
1 Supply lead 1 F 0
2 Piezoelectric b b F 1
specimen with
«>0 Q CP
capacitance CP U1
U 0 «<0 2 U=Q/CM
SAE0813-1Y
SAE0814-1Y
CM Measuring capacity U2
F Measuring force
Q Charge
U Voltage
however contribute to the transmission Packaging

function. A packaging which is tailor made to suit
• Charge amplifier (Fig. 5b): here the each individual application plays a deci-
charge is stored on a second capacitor. sive role for the quality of the sensor. Ac-
Parasitic capacitances have no influence. celeration sensors detect the measured
variable without a movable connection to
In both cases, resistors parallel to the sen- the outside world, it is no problem to en-
sor element must be taken into consider- capsulate them hermetically. A rigid me-
ation for the transfer function. chanical coupling to the body to be mea-
sured must, however, be achieved, as addi-
Thermal acceleration sensors tional elastic or loose elements could lead
Thermal acceleration sensors generate to falsification of the measurement. On the
a “bubble of heated gas” over a heater other hand, this fixed coupling must not
element. The tight heated gas space has lead to thermal expansion which may
a lighter density than the surrounding, occur at the structure being transferred
cooler gas. In the event of a lateral acceler- to the sensor so that measurements are
ation, the low density area of gas is dis- falsified.
placed within the surrounding, cooler gas.
The ensuing asymmetry is detected by Examples of application
thermocouples or resistors as a bridge • Piezoelectric acceleration sensors
circuit. The bridge voltage represents (bimorphous spring elements, longitudi-
the acceleration signal. nal elements such as knock sensors)
• Micromechanical acceleration sensors,
made in bulk micromechanics and
surface micromechanics
6 Principle of thermal acceleration sensors
3 2 3
a
5
Fig. 6
a Acceleration a = 0
b Acceleration a > 0
1 Heated gas zone
2 Heater element
3 Temperature sensor
SAE1078Y
4 Substrate coating
5 Displaced heated
zone
212 Sensor measuring principles | Pressure sensors
Pressure sensors • Combustion-chamber pressure (100 bar,

dynamic) for detection of misfiring and
Measured variables knock detection
Pressure measurement takes place either • Element pressure on the diesel fuel-
directly, by way of diaphragm deforma- injection pump (1,000 bar, dynamic)
tion, or using a force sensor for the follow- for electronic diesel control
ing applications in the motor vehicle • Fuel pressure on the diesel common rail
(examples): (up to 2,000 bar)
• Intake-manifold or boost pressure • Fuel pressure on the gasoline direct
(1 to 5 bar) for gasoline injection injection system (up to 200 bar)
• Brake pressure (10 bar) on electropneu-
matic brakes Measuring principles

• Air-spring pressure (16 bar) on pneu- Pressure as a measured variable is a non-
matic-suspension vehicles directional force acting in all directions
• Tire pressure (5 bar absolute) for tire- which occurs in gases and liquids. It is
pressure monitoring propagated in liquids, and also very well
• Hydraulic reservoir pressure (approxi- in gel-like substances and soft sealing
mately 200 bar) for ABS and power- compounds. There are static and dynamic
assisted steering measuring sensors for the measurement
• Shock-absorber pressure (approxi- of these pressures.
mately 200 bar) for chassis and suspen- The dynamic pressure sensors include
sion control for instance all microphones which, since
• Coolant pressure (35 bar) for air-condi- they are insensitive to static pressures, are
tioning systems used to measure pressure pulsations in
• Modulation pressure (35 bar) for auto- gaseous and/or liquid media. Since up to
matic transmissions now, practically only static pressure sen-
• Brake pressure in master cylinder and sors have been used in automotive engi-
wheel-brake cylinder (200 bar), and au- neering, these will be dealt with in more
tomatic yaw-moment compensation on detail here.
the electronically-controlled brake
• Overpressure/underpressure of the tank Direct pressure measurement
atmosphere (0.5 bar) Given that all known resistors are more or
less pressure-dependent (volumetric ef-
1 Pressure measurement fect), when very high pressures (>104 bar)
are to be measured it would suffice to sim-
ply subject an electrical resistor to the
Fig. 1
a b
pressure medium (Fig. 1a). On the other
a Direct 1
hand, they are at the same time more or
measurement,
3 less temperature-dependent, a character-
pressure-dependent
resistor (3) istic which is usually very difficult to sup-
b Measurement using p p press. Furthermore, the sealed lead-out
a force sensor (1)
c d
of their connections from the pressure
c Measurement medium presents difficulties.
through diaphragm
deformation via 2
Diaphragm-type sensors
strain gage (2) 4
The most common method used for the mea-
UAE0295-1Y
d Capacitive
measurement using surement of pressure (also in automotive
p p
the deformation of applications) uses a thin diaphragm as a
a diaphragm cell (4) mechanical intermediate stage which is
Sensor measuring principles | Pressure sensors 213
exposed on one side to the pressure to be varies under the influence of mechanical
measured and which deflects to a greater or stresses. The resistors are connected
lesser degree as a function of the pressure. together to form a Wheatstone bridge.
Within a very wide range, its diameter and Its voltage is a measure for the pressure.
thickness can be adapted to the particular Table 1 presents a systematic overview
pressure range. Low-pressure measuring of the proven pressure-measurement tech-
ranges lead to relatively large diaphragms niques which to a great extent have al-
which can easily deform in the range 1 to ready been used in automotive applica-
0.1 mm. Higher pressures though demand tions. The list is arranged according to
thicker, small-diameter diaphragms which the type of diaphragm material and the
generally only deform by a few µm. applied strain gage technology. The combi-
In case (capacitive) pick-offs for spacing nations have been marked to indicate ei-
or distance measurements are also re- ther those which are already actually in
quired, voltage-measuring methods domi- production and which will be described as
nate in the medium-pressure to high-pres- actual examples in the “sensor designs”
sure ranges. Here, practically only strain chapter (x), or those which are already
gage techniques are used. being closely considered for production
(marked in blue):
Capacitive pick-off
In contrast to their application in inertia 1 Strain gage pick-off and diaphragm material
sensors (see acceleration/yaw-rate sen- Strain gage Diaphragm material
sors), capacitive pressure sensors are still pick-off
Ceramics Metal Silicon
only rarely encountered even though they
(steel)
could possibly provide similar advantages
Foils 1)
(particularly with respect to their accu- (glued)
racy). This is more than likely the result of Thick-film
one important difference compared to the Metal x Table 1
other sensors referred to above: pressure thin-film 1) Low suitability
sensors need a direct contact with the Silicon x for large-scale
monitored medium. The dielectric charac- thin-film production, (x)
current examples
teristics of which practically always affect Diffusion x
◼◼ Suitability
the calibration of such capacitive pressure resistors
considered
sensors. This means that the calibration
would then not only be dependent upon With regard to the particular measuring
the medium in question, but would also be effect’s magnitude and type, the strain-gage
impossible without it (that is, in the “dry” techniques listed above have widely vary-
state). Clear separation of the sensor from ing characteristics. The gage factor (K) de-
the monitored medium has up to now only fines the magnitude of the measuring effect
been achieved at the cost of considerable of strain-gage resistors. It gives the relative
technical outlay. change in such a resistor’s resistance R re-
ferred to the relative change in its length l
Strain gage pick-off (equation 1):
The elongations of the diaphragm occur-
ring on the deflection of a diaphragm-type (1) K = –––––––– r/r
DR/R = 1 + 2?n + d––––––––
Dl/l «
sensor are detected using strain-gage
technology (strain-gage strips or strain- Here, the symbol « (elongation) is often
gage resistors). Strain-gage resistors are used for the relationship ∆l/l, and is used
affixed to the diaphragm (e.g. vapor- in multiples of 10–6 (ppm) as “micron” or
deposited). Their electrical resistance “micro strain”.
214 Sensor measuring principles | Pressure sensors
n is the material’s Poisson ratio, r repre- sure applied to its top and bottom sides.
senting its electric conductivity. n charac- This means that there are four different
terizes the reduction in cross-section of the basic types of pressure sensor (Table 3):
material on elongation and, in the ideal- • Absolute pressure
ized case of a volume being kept constant, • Reference pressure
amounts to n = 0.5 (in reality n = 0.3 to 0.4). • Barometric pressure and

Whereas the conductivity term in equa- • Differential pressure
tion 1 is of hardly any importance in the
case of metallic resistors, with regard to Transfer to a force sensor
Si resistors it plays a dominant role. Instead of directly using the force taken
up by their diaphragm, a number of sen-
One refers to a longitudinal K factor when sors transfer it to a force sensor whose
the resistor is expanded in the current di- measurement range can remain constant
rection and to a transverse K factor when due to the fact that the purely mechanical
it is expanded transverse to the current di- diaphragm has already performed the
rection (Fig. 2). Table 2 provides an over- adaptation to the pressure-measuring
view of typical values for the most impor- range. Perfect linkage from measuring
tant gage factors. diaphragm to force sensor (for instance
“Creep” (slight mechanical give under by a tappet) must be ensured though.
the effects of long-term unidirectional load-
ing) is a highly-feared phenomenon which, Examples of application
when it occurs at all, is only encountered on • Thick-film pressure sensors
glued foil strain gage. The other strain gage • Micromechanical pressure sensors
techniques all apply non-glued techniques • Si combustion-chamber pressure
and are not affected by this phenomena. sensors

To be precise, a diaphragm’s deformation • Metal-diaphragm high-pressure sensors
depends upon the difference in the pres-

2 Gage factors (K) for different materials
Material Gage factors (K)

2 Gage factor (K), physical quantities
longitudinal transverse
a F
Foil strain gage 1.6 to 2.0 ≈0
Thick-film 12 to 15 12 to 15
Metal thin film 1.4 to 2.0 –0.5 to 0
Δl
I Si thin film 25 to 40 –25 to –40
R
Si-monocrystalline 100 to 150 –100 to –150
Table 2 ΔR/R
Kl =
l εl
εl = Δl 3 Basic sensor types for pressure measurement
l Pressure on Pressure on upper side of diaphragm pO
Table 3
diaphragm Measured Ambient Vacuum
F bottom side pressure pressure
Fig. 2 pU
a Longitudinal Measured Differen- Reference Absolute
b Transverse b pressure tial pres- pressure pressure
F Force R I sure
I Current F F
Ambient Reference – Baromet-
R Resistance
ΔR/R pressure pressure ric pres-
l Length Kt =
UAE0819Y
εt sure
w Width
« Elongation εt = Δw
w
Vacuum Absolute Barometric –
w Δw pressure pressure
K Gage factor
Sensor measuring principles | Force and torque sensors 215
Force and torque sensors Since any form of measurement of only

part of the force or torque is very prob-
Measured variables lematical and can easily lead to false re-
The following list underlines the wide vari- sults, force and torque sensors must be
ety of applications for force and torque directly connected into the power flux
sensors in automotive engineering: (in other words, the complete measured
• In the commercial-vehicle sector, cou- variable must pass through them). Sensors
pling force between the tractor vehicle that measure force are extensive sensors,
and its trailer or semitrailer for the in other words, their size depends directly
closed-loop controlled application of on the measuring range.
the brakes, whereby neither push nor Although, in line with automotive-indus-
pull forces are active at the drawbar try demands, there are compact force and
• Damping force for use in electronic torque sensors on the market, these only
chassis and suspension control measure accurately enough when the
• Axle load for electronically controlled forces are introduced to the sensor in a
braking-force distribution on commer- precisely defined manner, a stipulation
cial vehicles that normally can only be complied with
• Pedal force on electronically-controlled under laboratory conditions. The toler-
brake systems ances and misalignment normally encoun-
• Braking force on electrically actuated, tered in practice generally dictate the con-
electronically-controlled brake systems nection of homogenization elements which
• Drive and brake torque in turn then lead to the sensors becoming
• Steering and steering servo torque too large.
• Finger protection on power windows If force and torque-transfer components
and electrically operated sliding sun- must be cut in order for previously test-
roofs able sensors to be fitted, this generally re-
• Wheel forces sults in an interface problem. This can only
• Weight of vehicle occupants (for occu- be solved by close cooperation between
pant-protection systems) the sensor supplier and the various suppli-
ers of the parts (or vehicle manufacturer)
In many cases, initial developments failed which must be cut in order to install the
to lead to the expected results since gener- sensors. Up to now, this problem has not
ally the costs involved in achieving the been encountered with other sensor types,
stipulated accuracy were excessive for at least not with this severity and with such
the systems in which the sensors were to wide-ranging implications.
be installed. Contrary to expectations, it Even if the force and torque-transfer
proved impossible to force down the costs components do not have to be cut, and me-
for the production of good torque sensors chanical elements are used as measuring
below those for pressure and acceleration springs which only need modifying for in-
sensors. In fact the torque sensors cost stallation of the sensor elements, very pre-
more. Matters are aggravated, and this ap- cise alignment is still necessary.
plies particularly to torque sensors, when Even though hardly any series produc-
the measured variable has to be trans- tion force and travel sensors are currently
ferred using non-contact methods from available from Bosch for the motor vehicle,
a rotating shaft (e.g. steering spindle or there follows a brief overview of the most
drive shaft) to a sensor mounted on the important measuring principles and pro-
chassis. Collector rings are not acceptable posed developments that have made it to
for the motor vehicle. the shortlist.
216 Sensor measuring principles | Force and torque sensors
Measuring principles being considered (e.g. for electronic

Basically speaking, when considering power-assisted steering systems), which
force and torque measurement, a differ- simultaneously permit torque measure-
ence must be made between static and dy- ment and high-resolution angle of rotation
namic measuring principles, and between measurement.
measuring principles based on displace-
ment and mechanical strain. Past require- Strain measuring force sensors
ments were predominantly for static sen- Magnetoelastic principle
sors, and in the case of force sensors, Under the influence of a magnetic field,
generally non-resilient strain-measuring ferromagnetic materials change their
principles were preferred. “Soft”, resilient length in the direction of the field (magne-
sensor systems are only accepted for tostriction effect). The length can either
torque sensors, particularly in the current increase (positive magnetostriction) or re-
example of steering-wheel torque sensing, duce (negative magnetostriction) here de-
which can also be implemented using an- pending on the material, for the same field
gle-measurement pick-offs. This is possi- direction. The inversion of this effect, the
ble, particularly because this characteris- change in the magnetic characteristics un-
tic proved to be tolerable in earlier hy- der the effect of tensile and compressive
draulic systems which were not equipped stresses, or elongation and compression,
with sensors. The use of magnetically ac- is known as the magnetoelastic effect.
tive coil systems previously dominated This effect is expressed in an anisotropic
both sectors, force and torque sensing, (directional) behavior of relative magnetic
in automotive and industrial uses. permeability mr (relationship between
More recently, however, displacement magnetic induction B and magnetic field
measuring magnetostatic sensors (Hall strength H). If this still has the same value
sensors) are being used here, too, to allow in all directions (isotropic) in a force-free
the application of micro-structured ele- situation, it takes on a somewhat different
ments manufactured in series production value (mrl) in the direction of the force un-
(e.g. measurement of the passenger’s der the influence of a force F introduced
weight) and, despite the familiar problems than that transverse to it (mrq) (Fig. 1). The
with dirt and the complexity of their instal- effect is not only observed on crystalline
lation, optoelectronic pick-offs are also or polycrystalline materials, but also in
amorphous substances.
1 Magnetoelastic anisotropic effect In fact, the permeability change in the
direction of force is a true reflection of the
sign of the force. Even though practically
a F b all ferromagnetic materials demonstrate
m this effect, it can be optimized by using a
specific alloy composition. Unfortunately
Fig. 1
mrl though, the materials which have good lin-
a Magnetoelastic
1
measurement mrq earity, low hysteresis, and low temperature
structure sensitivity, are not identical with the ones
mrq
b Measuring effect which have a high measuring effect.
mrl
F Force
Whereas the maximum measuring effects
mr Relative magnetic
F observed up to the present are about 30 %
permeability
(referred to the fundamental isotropic val-
UAE0835-2Y
mrq Transverse to the

F ues), and can be utilized without electron-
direction of force
mrl In the direction of ics being necessary, the effect displayed by
force materials which have been optimized from
the point of view of measurement tech- kept as small as possible. Often, this means
niques is only in an area of a few percent that the magnetically active measuring cir-
and needs electronic amplification. cuit is closed with ferromagnetic material,
The advantages of the magnetoelastic even when this is not included in the
effect are to be found in its broad tempera- power flux.
ture range and the fact that technical ap- Figure 2 shows the two most important
plications are possible up to temperatures possibilities of evaluating the magneto-
as high as approximately 300 °C. On the elastic effect: If a coil is arranged on the
other hand, it represents a marked volume measurement structure so that its direc-
effect. This means that the coils used for tion of field coincides with the direction of
detection not only register a local variation applied force, the change in inductance L
of permeability (as with a strain gage, for can be picked-off and applied directly. In-
instance) caused by the application of force, dependent of the magnitude of the applied
but more or less the effects throughout force, the excitation field strength H and
the whole of the coil’s cross section. This the induction B always have the same di-
makes the sensor somewhat less sensitive rection (Fig. 2a).
to the possibility of force being applied If the field strength H of the supply coil
asymmetrically. is not axially parallel to the applied force,
Since the changes in permeability as a the effect of the latter not only changes the
function of applied force are practically magnetic induction B, but also its direction
always registered with the help of alter- (due to the anisotropy of the permeability,
nating fields, the penetration depth of Fig. 2b). Assuming that with no force ap-
these fields, which is highly frequency- plied the directions of H and B are super-
dependent, must be observed: posed one upon the other in the normal
Only those mechanical forces can con- manner, these assume increasingly differ-
tribute to the measuring effect which are ent directions when force is applied and
present in the measuring field’s effective increased. This can be exploited particu-
penetration depth. In order to be able to larly advantageously in the variation of the
put the measuring effect to maximum use, magnetic coupling of two coils affixed at
the magnetically active air gap should be 90° to one another (Kreuzduktor) (Fig. 3).
2 Influence of the magnetoelastic effect 3 Magnetoelastic tensile-force/compressive-force sen-

sor for measuring brake torque (design specimen)
a F b F
H B (F 0)
H
B (F=0)
B (F 0) Fig. 2
B (F=0) a With direction of
force parallel to the
a
direction of field
b For different
F F directions of
UAE0836-1Y
field strength H
NAE0838Y
and force F
B Induction
a Enclosed angle
Strain gage principle (piezoresistive) surface of the selected elastic member that
Strain-gage measuring resistors (strain- they follow its surface elongation perfectly.
gage strips) represent the most wide- The change in resistance resulting from
spread and probably the most reliable and the resistor’s elongation is defined by the
precise method for measuring force and particular gage factor K of the resistor in
torque (Fig. 4). Their principle is based question (refer to “pressure sensors”):
on the fact that in the zone of the elastic-
member material to which Hooke’s Law (2) DR/R = K?«
applies there is a proportional relationship
between the mechanical strain s in the The K factor generally does not exceed a
member, caused by the introduction of value of 2 for metal-film resistors, but is
force, and the resulting elongation «. In generally somewhat below this in practice.
this case, in accordance with Hooke’s law: Strain gages are designed such that they
(in conjunction with a specific elongating
(1) « = Dl/l = s/E material (substrate) and their thermal
elongation) have no inherent temperature
whereby the proportionality constant E is sensitivity where possible (TCR ≈ 0). Any
the modulus of elasticity. Since it does not residual temperature sensitivity is usually
directly measure the strain resulting from eliminated by depositing the resistors on
the applied force, but rather the – locally – the elastic member in the form of a half
resulting elongation, the strain-gage or full bridge. Since temperature effects
method can be regarded as an indirect result in same-direction changes on the
measuring method. For instance, if the strain gage this results in no output signal.
modulus of elasticity decreases by 3 % The auxiliary bridge resistors can be (but
above 100 K, a figure which is normal for need not be) located within the elastic mem-
metals, then the force indicated by the ber’s elongation zone. They can also be
strain-gage method is 3 % too high. fitted as purely compensation resistors
Strain-gage resistors in the form of film (Fig. 4c). It must be noted that often the
resistors – by contrast with cemented-on K factor itself also has a temperature coef-
strain gages – are so closely bonded to the ficient (TCK). Usually, this decreases along
4 Strain-gage force sensors
a b Rl Rq
F F
F F
Fig. 4 F
a Rod-shaped
b Toroidal-shaped Rq
c Electronic F
evaluation
F Force Rl
Rl,q Metal-film resistors, F
lengthwise, c Rl
Rq
crosswise
U0 -
UAE0839-2Y
R Auxiliary bridge
F R R
resistors
+ UA
U0 Supply voltage
UA Output voltage
with increasing temperature, which means (and only this strain) at the point at which
that in favorable cases it can compensate it is applied. This, of course, no longer
for the signal increase caused by the mod- holds true in cases in which the strain-
ulus of elasticity. Apart from this, signal gage structure is distributed across the
reductions caused by the K factor are usu- elastic member’s complete surface. This
ally compensated for by a bridge supply necessitates extremely precise and repro-
voltage U0 which increases accordingly. ducible application of the force to be mea-
sured if measuring errors due to uneven
In addition to the longitudinal K factor Kl force introduction are to be avoided.
described above, there is also the transverse
K factor Kt, which must then be applied if the Application:
direction of strain is transverse to the cur- As a rule, in order to carry out force mea-
rent flow. It has an inverse sign (negative) surements, very small strain-gage resistors
and is no greater than 0.5 in value. must be attached to larger force-carrying
parts or elastic members. The traditional
Disadvantages/limitations: method of applying the strain-gage resis-
Notwithstanding their high levels of accu- tors to the elastic member by means of a
racy and reliability, since the elongations foil backing (as applied in such devices as
and therefore also the changes in resis- high-precision scales), is not inexpensive
tance (at least in the case of metal film re- enough for large-batch “low-cost” produc-
sistors) are only in the percentage range tion. First attempts are therefore being
of the original state, the strain-gage sen- made to apply low-cost, large batch film/
sors only generate output signals in the layer techniques by depositing the strain-
mV range so that in general local amplifi- gage resistors on small metal wafers which
cation is required. A further disadvantage are then pressed into, or welded onto the
of small strain-gage sensors is the fact elastic element.
that they measure precisely the strain
Orthogonally printed resistors
Practically all electrical film resistors
5 Piezoresistive behavior of various resistance mate- change their resistance not only under the
rials when orthogonal compression is applied
influence of laterally applied deformation
strain but also when compression is ap-
plied vertical to the film plane (orthogo-
4 1 nal). Here, the so-called “conductive plas-
Relative resistance change DRp/R0
2 tic”, commonly used in potentiometers,

2
features very high sensitivity. “Cermet”
and “carbon layers” are also highly sensi-
0
3 tive (Fig. 5). Up to a certain limit, the resis-
–2 tance of the above materials decreases
4 5 along with increasing compression. The
–4 values that can be achieved without per-
manent resistance change are similar to Fig. 5
–6
those which apply for lateral elongation. 1 84.5 Ag 15.5 Mn
6 In both cases, the limit is a function of 2 Manganin
–8 3 Cu
8 7 substrate strength and not of the resis-
4 Au
–10 tance material. Of course, sensors of this
UAE0840-1E
0 400 800 bar 5 Ag

type are almost only suitable for loading 6 Carbon film/layer
Pressure p
by compression but not by lateral strain. 7 Cermet
8 Conductive plastic
Travel measuring force sensors current) require a certain length l of the

The question of whether an integrated torsion shaft via which the torsion angle
travel-measuring weight sensor still has a (approximately 0.4 to 4°) can be picked-off.
little additional elasticity or not is not im- The mechanical stress s proportional to
portant when it is built into seats for motor the torque is aligned at an angle of less
vehicle occupants, which are anyway sus- than 45° to the shaft axis (Fig. 6).
pended. The principles described below are all
This principle is used in the iBolt™ force suited for the non-contacting (proximity)
sensor (see sensor types). Four of these transfer of measured values, even from ro-
sensors are built into the frame of the front tating shafts. In the case of steering-torque
passenger seat for measuring the weight measurement, it is also desirable for the
of the passenger. This makes it possible employed measuring system to measure
to modify the triggering of the airbag. the steering angle very accurately (through
a full rotation of 360°). This is in the sense
Examples of application for force sensors of modular integration and is to be imple-
• Magnetoelastic load measuring bolts mented with only slight modification.
(agricultural tractor with plow force
control) Strain-measuring sensors
• iBolt™ force sensor for sensing passen- Even though magnetoelastic sensors which
ger weight enclose the shaft are available on the mar-
ket, these involve very high costs. Since it
Torque sensors is often impossible to optimize the shaft
A fundamental distinction is made in material with regard to its magnetoelastic
torque measurement, too, between meth- properties, a search is being made for
ods using angle or strain measurement. methods to coat the measuring shaft with
In contrast to strain-measurement meth- a magnetoelastic layer. Such a coating,
ods (strain-gage resistors, magnetoelastic), which has to exhibit good measuring quali-
angle-measurement methods (e.g. eddy ties, has still not been found.
This is why the strain gage principle has
become predominant (Fig. 7): a strain-gage
6 Torque measurement: basic principle bridge measures the mechanical stress.
The bridge is powered using transformer
principles (by a rectifier and control-elec-
tronics circuitry on the shaft, independent
M r of the air gap). Further local electronic
components on the shaft permit the mea-
surement signal to be amplified and con-
F verted to an AC form which is independent
of the air gap (for instance, frequency-ana-
log) which can then also be outputted us-
ing transformer principles.
l When large-batch quantities are in-
volved, the electronic circuitry on the shaft
Fig. 10
can be integrated in a single chip without
1 Torsion rod
problems. The strain-gage resistor can be
F Torsion angle
inexpensively formed on a prefabricated
UAE0708-3Y
s Torsional stress
M Torque round steel plate (for instance, using thin-
+s s 1
r Radius film techniques) which is then welded to
l Rod length the shaft. The two ring transformers re-
quired for power and signal transmission Up to now, since adequate accuracy de-
can be manufactured at an extremely mands extremely precise bearings, to-
low cost from a soft-magnetic molding gether with the necessity to provide corre-
compound. High precision levels can be spondingly accurate angular or incremental
achieved with such a configuration in subdivision around the complete periph-
spite of reasonable manufacturing costs. ery, this method has been regarded as too
complicated. Nevertheless work is forcing
Angle-measuring (torsion-measuring) sensors ahead with solving this problem (magneti-
Sensors for measuring angular difference cally or optically), because such a system
It is a relatively easy matter to determine would incorporate two distinct advantages:
the torsion angle when two mutually-inde- • Possibility of simultaneous measure-
pendent incremental speed sensors, or an ment of the angle of rotation with the
absolute-measuring (analog or digital) same system
non-contacting angular-movement pick- • Possibility of measurement without far-
off, are provided at each end of the reaching modifications to the torsion
section of the shaft subject to torsion shaft being necessary, so that essentially
(L ≈ 5 to 10 cm) (Fig. 8). The difference the sensor could be in the form of a
in their outputs w2 - w1 is a measure for plug-in sensor providing an efficient
the torsion angle: interface for a supplier component
(3) M = const?L?(w2 - w1) The angle may be measured with high

where L = length of the precision using the vernier principle, for
section subject example: this is done by marking the cir-
to torsion cumference of the shaft with a sinusoidal
track, of which an integer number of
N cycles fits in one circumference.
7 Strain-gage torque sensor with non-contacting 8 Determining torque by measuring angular

(proximity) transformer pick-off difference
M
M Fig. 7
w1 1 Torque indicator
s Torsional stress
1 M Torque
3 U0 Supply voltage
f 1 R1 to R4 Strain-gage
U measuring resistors
9.81 Nm
R2 R4 s +s
Fig. 8
Strain 2
l 1, 2 Angle/speed
gage
R1 R3 sensors
3 Angle markings
w2 l Torsion-
measurement
UAE0709-2Y
UAE0844-1Y
section
U0 M Torque to be
measured
w1,2 Angle signals
By contrast with a pure incremental track, cated in the second sine cycle. If the value
which would permit only a resolution of an is between 216° and 252°, one is located
N-th of the circumference, this sinusoidal in the 6th cycle. The precise measuring
signal structure, however, theoretically angle w is then formed by adding another
permits as fine a resolution per circumfer- angle, in the first case of 360° and in the
ence as is required by the application of second case of 216° to the measured, but
the arcsin function within a single sine not unambiguous fine signal w.
cycle. This fine resolution is, however, In practice however, the arcsin function
only useful if it is always known in which is not used to resolve the individual sine
of the N similar cycles one is located. cycles finely, as it would be necessary to
The unambiguity can easily be created guarantee a constant and normalized sig-
if a second track with a little less spatial nal amplitude for this. It is rather the case
frequency in which only N-1 sine cycles that a second, cosine-wave track, phase-
fit on the circumference is applied around shifted against the first by precisely 90°,
the circumference. Although the two sig- is applied next to the sine-wave marking.
nals do not supply an unambiguous signal Because of the spatial vicinity, it may be
for an angle-of-rotation measurement, assumed that both tracks are sensed with
it is, however, possible to draw firm con- the same amplitude u, so that the angle w
clusions from the phase difference Dw be- within the sine cycle can be determined by
tween the two signals, in which sine cycle the arctan function from the two individual
one is currently located. The pitch differ- signals u1 and u2 independently of u:
ence between the two signals varies over
the entire circumference by precisely 360°. (4) w = arctan (u1/u2)
Figure 10 shows an example with N = 10. where u1 = u?sin w and
If the phase difference between the two u2 = u?cos w
signals is, for instance, in the range be-
tween 36° and 72°, then one is clearly lo-
9 Optoelectronic angular difference steering torque 10 Angle measurement using the vernier principle
sensor
9 Dw
1.0 360
8
7
Fig. 9 6 0.5 180
Signal values
Optical pick-off of the 5

Dw
angle marking tracks
4 0 0
1 Steering shaft with
torsion rod
2 Code disks with 3
–0.5 –180
bar code
3 Steering gear 1
housing –1.0 –360
4 Lens 2 0 0.2 0.4 0.6 0.8 1.0
5 Optical waveguides Dw Circumference
UAE1030-1E
UAE1029-1Y
6 Optical ASIC
7 LED
8 Sensor module
9 PCB
Thus a total of four tracks are required for must be precisely manufactured and as-
the determination of a precise and unam- sembled in order to achieve the stipulated
biguous angle of rotation w. The steering accuracy. The associated electronics are
angle and TAS (Torque Angle Sensor) steer- appropriately accommodated very near
ing-angle sensor illustrated in Figure 9 thus to the coils.
requires a total of eight tracks, which are Although this sensor principle was de-
read optoelectronically, to measure two veloped at Bosch for motor vehicle appli-
angles w1 and w2. Its angular precision over cations, it was only used in electric power
360° is 1°, the angular resolution is 0.0055°, tools (torque sensing in industrial power
the resolution of the difference angle Dw is screwdrivers). A license-holder in Japan
0.0044° with a measuring range of + 9°. has further developed this, very cost-ef-
fective, sensor principle for motor vehicle
Eddy-current sensors applications to product maturity.
Two slotted sleeves of electrically con-
ducting aluminum are attached to each Examples of application for torque sensors
end of a sufficiently long section of the • Strain gage steering torque sensor (de-
measuring shaft. These are inserted one veloped up to A prototype for electrical
inside the other (Fig. 11). Each sleeve is power-assisted steering)
provided with two rows of slots so that • Angle-measuring eddy-current torque
when the shaft is twisted, it becomes in- sensors (electric power tools, third-
creasingly visible through one row of slots, party product in Japan)
and is hidden more and more by the other • Angle-measuring optoelectronic steering
row. This leads to increasing, or decreas- angle sensor (electrical power-assisted

ing, damping of the two high-frequency steering)
coils (approx. 1 MHz) situated above each
row of slots so that coil inductance also
varies accordingly. The slotted sleeves
11 Eddy-current torque sensor
Fig. 11
2
1 Slotted sleeves
UAE0710-2Y
1 2 Air gap
3 High-frequency
coils
M Torque
224 Sensor measuring principles | Flowmeters
Flowmeters flow is 1:50 to 1:100. Because of the higher

idle air demand in diesel engines, these
Measured variables ratios must be assumed to be 1:20 to 1:40.
The purpose of flow measurement in the The severe exhaust gas and fuel-consump-
motor vehicle is the detection of the intake tion requirements dictate accuracies of
air flow rate. This air flow rate must be 2 to 3 % of the measured value. Referred
known precisely so that the engine-man- to the measuring range, this can easily
agement system – both in diesel and in correspond to a measuring accuracy of
gasoline engines – can set a defined air- 2?10-4, which is unusually high for a motor
fuel mixture. This value can be determined vehicle.
by a flowmeter. The sensors which are The air though, is not drawn in continu-
used for measuring air flow rate or gas ously by the engine, but rather in time with
flows in general are also referred to as the opening of the intake valves. Particu-
“anemometers”. larly with the throttle valve wide open
As such, the often-used term “air quan- (WOT) in gasoline engines, this leads to
tity” is incorrect because it does not stipu- considerable pulsation of the air-mass
late whether volume or mass is concerned. flow, also at the measuring point which is
Since the chemical processes involved in always in the intake tract between air filter
fuel combustion are clearly based on mass and throttle valve, or between air filter and
relationships, the object of the measure- turbocharger (Fig. 1). Intake-manifold res-
ment is the mass of intake air. On gasoline onance leads to the pulsation in the mani-
engines, the air-mass flow rate is the most fold sometimes being so pronounced that
important load parameter. In diesel en- brief return flows can occur. This applies
gines, the exhaust-gas recirculation rate in particular to 4-cylinder engines in
is regulated using the air-mass flow. which there is no overlap of the air-intake
Depending upon engine power, the aver- phases. An accurate flowmeter must be
age (over time) maximum air-mass flow capable of registering these return flows
rate to be measured is between 400 and with the correct direction.
1,200 kg/h. Due to the low air require- On a 4-cylinder engine, the pulsations
ments at engine idle in modern gasoline are generated at twice the crankshaft
engines, the ratio of minimum to maximum speed. This means that they can easily be
in the range 50 to 100 Hz. With a flowme-
ter featuring a linear characteristic curve
1 Pulsating air-mass flow QLM in the intake tract and a narrower frequency bandwidth than
above, it would suffice for it to follow the
mean value of these rapidly fluctuating air
kg/h
flows. The mean value is in any case posi-
1 crankshaft rotation tive, so that the meter need not necessarily
Air-mass flow QLM
300 detect the correct sign.

Practically all of the air-mass meters ac-
200 tually in use feature a characteristic curve
QLM
Fig. 1 which is far from linear, so that the mea-
At full load with 100 surement signal must be linearized elec-
engine speed
tronically before it can be evaluated.
n = 3, 000 rpm,
Averaging before linearization can lead
intake-manifold 0 5 10 15 20 ms
to significant dynamic errors (mean-value
UAE0802-1E
pressure Time t
pS = 0.96 bar, errors). This fact means that such air-mass
mean air-flow rate meters must be sufficiently rapid in follow-
QLMm = 157.3 kg/h ing the pulsations which mostly have a
Sensor measuring principles | Flowmeters 225
pronounced non-sinusoidal characteristic, between the laminar and turbulent profiles

and therefore also have a considerable illustrated will be established for the case
harmonic content. This necessitates a of laminar flow – caused by the short in-
bandwidth of about 1,000 Hz. Apart from flow. The laminar or turbulent characteris-
this considerable bandwidth, the air-mass tic of the flow profile depends on whether
meters must also have a high switch-on Re is below or above a value of approxi-
time constant in order for them to be able mately 1,200.
to measure correctly during the engine A further important influencing factor is
starting phase. the degree of turbulence in the flow, which
As is the case in all flowmeters, those is determined by the aerodynamic design
used in the motor vehicle are calibrated of the inflow to the flowmeter. If the lami-
for flow through a pipe with a symmetrical nar-turbulent transition is within the mea-
flow profile. This means that a flow with a suring range, marked irregularity of the
velocity vector v at practically any point of characteristic curve can be expected at
the flow cross-section of area A depends this point. As far as motor vehicle applica-
only on the radius to the central axis, and tions go, a purely turbulent flow (square
not on the angle at the circumference. The profile: v = constr) cannot be assumed at
flow profile (laminar or turbulent, Fig. 2) the smallest air masses (gasoline engine
is related to the Reynolds number Re: idling). A careful design of the inflow, to
achieve square profiles even in the event
Re = v?D/h of laminar flow, is then necessary.
where Assuming a homogeneous density r, and
D typical cross-section dimension and assuming a square profile, the flow is sim-
h kinematic ductility of the medium. ple to calculate as follows:
The flow profiles illustrated by way of ex- QV = v?A Volume flow rate
ample only establish themselves after the QM = r?v?A Mass flow rate
flow has been uniformly formed for some
time. In the motor vehicle, a profile lying
2 Flow profiles
y (r) Fig. 2
1 Laminar flow
Q Q profile
2 Turbulent flow
R r
profile
A 1 2 A A Cross-sectional
area of the tube
Q Flow
UAE0803-1Y
R Tube radius
r Distance from
the tube center
v(r) Flow profile
Whereas in measurement techniques, Variable orifice plates (sensor plates)

long, straight, advance and overshoot The calculation of the pressure drop across
sections of constant cross-section are fixed orifice plates is based on two physical
stipulated in order to guarantee a sym- laws:
metrical profile, such conditions cannot
be complied with in a vehicle’s cramped Continuity equation:
under-hood installation space. If pro-
nounced asymmetries occur, the flowme- r?1?v1?A1 = r?2?v2?A2 = const
ter must be calibrated as a function of the
actual installation conditions. Bernoulli’s equation:
1 1
Measuring principles p1 + ___ ?r1?v12 = p2 + ___ ?r2?v22 = const
2 2
Up to now, of the practically unlimited
variety of flowmeters on the market, only These laws are to be applied for two different
those which operate according to the im- measuring cross-sections A1 and A2 (Fig. 3).
pact-pressure principle have come to the Assuming constant density r = r1 = r2, this
forefront for air-quantity measurement in results in the pressure drop:
the vehicle. This principle still depends
1 1
upon mechanically moving parts, and in Dp = QV2?r?(____ – ____)
A22 A12
principle correction measures are still
needed to compensate for density fluctua- This pressure drop can be measured either
tions. directly with a differential-pressure sensor,
Today, true air-mass meters applying or by means of the force acting against a
thermal methods (hot-wire or hot-film air so-called sensor plate. It should be noted
flowmeters) are used which can follow here that impact-pressure gages of this
sudden flow changes without mechanically type measure a flow value that corresponds
moving parts. neither to the volume flow rate nor to the
mass flow rate, but to a geometrical mean
value of the two:
Fig. 3 ––
QSt = const ? √ r ? v = const ? √QV ? QM
a Ring orifice
b Sensor plate
1 Orifice plate
AS Disk cross-section 3 Impact-pressure flowmeter 4 Impact-pressure air-mass meter
A1, 2Measuring cross-

section a Dp 1
p1 p2
p1, 2 Measured
pressures
1
Dp Pressure drop
QLM Air-mass flow QLM A1 A2 QLM
QL
AS QL
Fig. 4 b Dp 2
p1 p2
1 Sensor plate
2 Air-temperature 3
sensor
3 To control unit QLM A1 AS 1 A2 QLM
UAE0299-2Y
UAE0804-1Y
4 Potentiometer
5 Damping chamber 4 5 6
6 Compensation flap
QL Intake-air flow
Because of the square relationship to the Hot-wire/hot-film anemometers

flow rate, fixed orifice plates permit only a When current IH flows through a thin wire
variation in measured variable of 1:10, as with electrical resistance R, its tempera-
otherwise the accuracy of the differential ture increases. If at the same time a me-
pressure recording is no longer adequate dium with density r flows across it at ve-
and significant pressure losses would result locity v, a balance is set up between the
which would lead to power losses in the en- electrical power input Pel and the power PV
gine. When larger ranges are to be covered, drawn off by the air flow, whereby
several orifice plates must be used, or such
versions which automatically adapt them- Pel = IH2?R = PV = c1?l?Dq
selves to the measuring range by opening
up a larger flow cross-section A2 in line with Here, the power drawn off by the air flow
the impact pressure. is proportional to the temperature differ-
With such variable, moving sensor ence Dq established and the coefficient of
plates it is an easy matter to increase the thermal conductivity l. The following ap-
variation to 1:100. Here, the increasing plies in close approximation:
air flow causes the sensor plate to be de-
–––––– ––––––
flected (usually against a constant coun- l = √ r?v + c2 = √ QLM + c2
terforce) into an area whose cross-section
is specifically shaped so that the resulting Although l is primarily a function of the
deflection/angle relationship complies mass flow QLM, with the static medium at
with the desired characteristic. In other standstill (v = 0), a certain heat loss takes
words, linear for K-Jetronic and non-linear place (convection) represented by the ad-
for L-Jetronic. The sensor plate’s (Fig. 4) ditive constant c2. This results in the famil-
setting is then a measure for the air flow iar interrelationship between the heating
which is in relationship to the impact pres- current IH and the mass flow QLM.
sure defined above. ––––
––––––––––––––
The limit frequency for such sensors is IH = c1 · √(√QLM + c2) · √ DRq
typically approx. 10 Hz. Such sensor plates
though are unable to follow the high pulsa-
tion frequencies which often occur. From With the supply of a constant heat IH2?R,
the point of view of the pulsation, they can a temperature rise Dq decreasing recipro-
be regarded as fixed orifice plates with a cally with the root of the air-mass flow QLM
square-law characteristic curve. Under would establish itself. If, on the other
certain load conditions this leads to con- hand, the heating current IH is controlled
siderable mean-value errors which can such that a constant temperature increase
only be compensated for roughly by the (for instance, Dq = 100 K) is maintained
use of suitable software. even when the flow rate increases, this will
If the density r of the intake air changes lead to a heating current which increases
because of temperature fluctuations or the approximately with the fourth root of the
altitude, the measurement signal here will mass flow, and at the same time serves as a
––
only change by √ r. An air-temperature measure for the mass flow.
sensor and a barometric pressure sensor The advantage of such a control loop lies
must also be used to determine the density in the fact that the electrical heating resis-
fluctuation fully. tor always remains at the same tempera-
ture so that its heat content need not be
changed by means of time-wasting heat
transfer. In fact, with a 70 µm platinum
wire for instance, it is possible to achieve
time constants in the 1 ms range for average out any asymmetries in the flow
changes in air-flow rate. In cases where profile (Fig. 6). Adequate service lives only
closed-loop control is not used the time became possible when the platinum wire
constants would be 40 to 100 times higher. was stabilized by alloy additives so that
its resistance no longer changed due to
If the heater temperature were to be main- deposits and cracks on its surface. This
tained constant simply by keeping its (tem- meant though that the deposits on the hot
perature-dependent) resistance constant, wire had to be burned-off automatically
with constant mass flow and higher me- (approximately 1,000 °C) following every
dium temperature, this would result in a operating phase.
current drop and therefore a false mea-
surement. In practice, this error is avoided Notwithstanding a number of functional
by using a bridge circuit containing a sec- advantages, this sensor concept was far
ond high-ohm compensation resistor RK of too costly. A thick-film version (HFM2 hot-
the same type (e.g. platinum). Here, the film air-mass meter) was able to combine
heating resistor is kept at a constant overall the resistors concerned with the mea-
temperature Dq compared to the medium surement on a single ceramic substrate.
(Fig. 5) by a control loop. In case of a sud- Due to the ceramic substrate’s consider-
den jump in the medium temperature, the able thermal capacity, it was difficult not to
sensor reacts with a long time constant exceed the maximum permissible switch-
since in this case the heat content of the ing constants. Furthermore, a complicated
hot wire must be changed. The heating saw cut had to be made to reduce the un-
current generates a control voltage UM desirable heat coupling between heating
proportional to the air-mass flow on a pre- and compensation resistors. On the other
cision resistor (measuring resistor RM). hand though, this version permitted the
The heating resistors in the first air- burn-off process to be dispensed with since
mass meters (hot-wire air-mass meters) the special flow conditions no longer led to
used for automotive applications were of unwanted deposits.
very fine platinum wire. This wire was In contrast to both its predecessor
stretched in trapezoidal form across the types, a further silicon-based microme-
Fig. 5 flow cross-section so that it was able to chanical version (HFM5 and HFM6 hot-
QLM Air-mass flow
UM Measurement
voltage
RH Hot-wire resistor 5 Hot-wire air-mass meter (circuit) 6 Hot-wire air-mass meter (components)
RK Compensation
resistor
IH
RM Measuring resistor
R1,2 Calibration QM tL
resistors
RK RH
1 2 3
Fig. 6
1 Temperature R1
compensation
resistor RK
2 Sensor ring with
RM
hot wire RH
RM UM QM RM
UMK0311-2Y
UMK1302-1Y
R2
3 Precision RH
measuring
resistor (RM)
QM Air-mass flow
film air-mass meters) fulfilled practically this flowmeter must ensure that the partial
all expectations. In particular, this version flow it measures represents the same frac-
is able to measure in both flow directions tion of the total flow throughout the whole
with the correct sign (Fig. 7). This means measuring range. This precondition can be
that the brief return flows that occur as a achieved by means of a flow rectifier (grill)
result of pulsation no longer lead to mea- built into the measuring pipe.
suring errors (Fig. 8).
To this end, in addition to the heater Examples of application
control circuit used in the previous ver- • HFM5 and HFM6 hot-film air-mass
sions, a temperature sensor is located on meters for gasoline and diesel engine
each side of the heating resistor, in other applications
words upstream and downstream. This
principle is similar to the Thomas process
often encountered in the literature. When
there is no flow (QML = 0), each of these
sensors indicates the same temperature.
When flow starts though, since the up-
stream sensor is cooled by the medium,
the higher the flow the higher the temper-
ature difference between the two resistors.
The output signal derived from the tem-
perature difference has a similar charac-
teristic curve to the anemometers used up
to now, whereby its sign is a clear indica-
tion of the flow direction.
Due to its small size, the micromechani-
cal flowmeter is only a partial-flowmeter.
In other words, it is no longer in any way
able to average-out any non-homogeneity
in the flow velocity as a function of the
flow cross-sectional area. Rather,
7 Micromechanical hot-film air-mass meter with 8 Determination of the pulsating air-mass flow
air-quantity measurement in both directions in a 4-cylinder engine
5 kg/h
V
4 100
Air-mass flow QLM
Signal voltage U
Fig. 7
3 QR Return flow
QR 50 1
2
0 Fig. 8
1 At full load and
2
engine speed
0 -50 n = 900 rpm
-100 0 100 200 300 kg/h
UAE0807-1E
UAE0808-1E
0 10 20 30 40 50 ms 1 Hot-wire
Air-mass flow QLM Time t air-mass meter
2 Hot-film
air-mass meter
230 Sensor measuring principles | Gas sensors and concentration sensors
Gas sensors and The introduction of the fuel cell as an auto-

concentration sensors motive drive means that further gas sen-
sors will have to be developed, for instance
Measured variables for the detection of hydrogen.
The concentration of a given material or
medium defines the mass or volume per- Measuring principles
cent of a given material in another given Measured mediums occur in gaseous, liquid,
material or in a mixture or combination of or solid state, so that in the course of time
other materials. With a concentration sen- countless measuring methods have been de-
sor (also known as a concentration probe, veloped. For automotive applications, until
the important thing is that in the ideal case now only the gas-analysis area, and in par-
it is sensitive to only one medium, while at ticular the measurement of gaseous humid-
the same time practically “ignoring” all ity, has been of any interest. Table 1 presents
other mediums. Of course, in practice, ev- an overview of the processes applied in gen-
ery concentration sensor has its own cross eral measurement techniques.
sensitivity to other mediums even though,
as is often the case, “temperature” and Gas measurement in general
“pressure” are maintained constant. Gas sensors are usually in direct unpro-
In the vehicle, the following parameters tected contact with the monitored medium
must be measured: (in other words with foreign matter) so
• Oxygen content in the exhaust gas that the danger of irreversible damage
(closed-loop combustion control, exists. This form of damage is referred to
catalytic-converter monitoring) as sensor “contamination”. For instance,
• Carbon-monoxide and nitrogen-oxide the lead that may be contained in fuel or
content, as well as air humidity inside the exhaust gas can make the electrolytic
the vehicle (air quality, misting of vehi- oxygen concentration sensors (Lambda
cle windows) oxygen sensors) unusable.
• Humidity in the compressed-air brake
system (air-drier monitoring) Moisture measurement

• Dampness of the outside air (black ice In addition to the outstanding significance
warning) of the Lambda oxygen sensor in dealing
• Concentration of soot in diesel-engine with exhaust gases, moisture measurement
exhaust gas. A still unsolved problem. also plays an important role.
In contrast to the above-mentioned gas
concentrations, this is a particle concen- In the broader sense, moisture indicates
tration. The difficulties inherent in the the moisture content of gaseous, liquid, or
measuring assignment are further ag- solid substances. In the narrower sense,
gravated by the possibility of the sensor we are dealing here with the gaseous-wa-
being blocked by particles so that it no ter (water vapor) content in gaseous media –
longer functions. above all in the air.
1 Gas-analysis processes (without particular attention being paid to the moisture-measurement process)
(X) = for automotive applications
Physical process Physical-chemical process Chemical process
Thermal conductivity Heat of reaction Selective absorption
Magnetic processes Heat of absorption Selective absorption with prior
chemical conversion
Radiation absorption Characteristic color reaction
Gas chromatography Electrolytic conductivity X
Table 1 Radioactive processes Electrochemical processes X
Sensor measuring principles | Gas sensors and concentration sensors 231
When a damp gas is cooled in an isobaric Absolute humidity:

process, it reaches its saturation point at
mw
x = ––––– Mw pw
a specific temperature (known as the dew = ––––– · ––––––– (in %)
mtr Mtr p–pw
point t).
A number of important definitions are
mw
x = fa = –––––
given below in connection with humidity (volume-related)
Vtr
measurement (refer also to Figure 1):
Relative humidity:
mw Mass of water
pw
ms Mass of water in the F = ––––– (in %)
ps
saturation state
mtr Mass of dry gas
For low-cost applications (for instance in
Mw Mole mass of water the vehicle), resistive and capacitive sen-
Mtr Mean mole mass of the sors are used almost exclusively. They are
dry gas provided with hygroscopic layers which
can store water as a function of the relative
p Total pressure of the humidity (and release it again), and thus
gas mixture trigger a usually drastic change in a resis-
pw Partial pressure of the tor’s value or in the value of a planar ca-
water vapor pacitor.
ps Saturated vapor pressure On capacitive humidity sensors, a hygro-
(vapor pressure of the water scopic insulating layer (e.g. Al2O3, or a
at mixture temperature) polymer plastic), which can possibly also
1 Temperature/humidity diagram for air
5% 10% 20% 30%

C 130 40%
120 50%
60%
110 70%
80 90%
100 100%
Damp-air temperature t
90
80
70
60
y
idit
50 hum
la tive
40 Re
30
20
10
0
Water level Absolute humidity
Dew point τ –10 0 10 20 30 40 50 60 70 80 90 100
(C)
Partial pressure Pw
2.14 4 6 8 10 20 30 40 50 60 80 100 150 200 300 400 500 600 760
(Torr)
Absolute humidity ƒa
2.35 3 4 6 8 10 20 30 40 50 60 80 100 130 150 200 300 400 500 597
SAE0901E
(g/m3)
Absolute humidity χ
1.8 2 4 6 8 10 20 30 40 60 80 100 120 150 200 300 500 1000 2000 5000
(g/kg)
232 Sensor measuring principles | Gas sensors and concentration sensors
2 Humidity-measurement procedures. X = technologically important
Processes Serial no. Measuring device Measuring method

Saturation processes 1 X Dew point hygrometer Direct methods
2 X LiCI dew point hygrometer (measurement of absolute
Evaporation processes 3 X Psychrometer humidity)
Absorption processes 4 Volume hygrometer
5 X Electrolysis hygrometer
6 Condensate volume hygrometer
Energetic processes 7 X Infrared hygrometer
8 Microwave hygrometer
9 Electrical-discharge hygrometer
10 Diffusion hygrometer
Hygroscopic processes 11 X Electrical conductive-foil Indirect methods
12 hygrometer (measurement of relative
13 X Capacitor hygrometer humidity)
14 X Hair hygrometer
15 Bimetal strip hygrometer
16 Color hygrometer
17 Quartz hygrometer
Table 2 Gravimetric hygrometer
be the mounting wafer, serves as the
2 Capacitive sensor wafer with comb electrodes
dielectric of a capacitor. Either one of the
1 2 electrodes is permeable to water, or the
electrodes have a comb-like structure
(Fig. 2). With increasing humidity, the
dielectric absorbs more water and the
sensor’s capacitance increases consider-
ably (relative dielectric constant of water
«rW ≈ 81, Fig. 3).
Fig. 2 In the case of the resistive sensor, an insu-

1 Plastic wafer lation substrate is coated with a layer of
SAE0903Y
2 Comb-like, hygroscopic salt (LiCi) held in a paste

embossed gold-foil
binder and located between an electrode
electrodes
pair. The layer’s conductivity changes
3 Resistive and capacitive humidity sensors drastically along with the relative humidity
(typical characteristic curve)
(Fig. 3). Unfortunately this resistance
change is also highly dependent upon tem-
10,000 perature so that normally compensation
k
is needed. The dew point can also be de-
1,000 nF
fined, and with it the absolute humidity,
Resistance R
by an additional measurement of the air

Capacity C
100 100
C temperature. The typical time constants
of these sensors are about 30 s.
10 10
R
Table 2 gives an overview of the numerous
1 1
measurement processes that have been
0.1 0.1
developed in the course of time for humid-
SAE0902E
0 20 40 60 80 % 100 ity measurement.

Relative humidity φ
The piezoelectric effect 233
▶ The piezoelectric effect
In 1880 Pierre Curie and his brother Jacques The change in length Dx when a voltage U is
discovered a phenomenon that is still very lit- applied results from the following:
tle known today, but is present in the everyday U / d = Dx (Example of quartz: deformation of
lives of millions of people: the piezoelectric approximately 10–9 cm at U = 10 V)
effect. For example, it keeps the hands of a
quartz clock operating in time. The piezoelectric effect is not only used in
quartz clocks and piezo-inline injectors, it has
Certain crystals (e.g. quartz and turmaline) are many other industrial applications, either as a
piezoelectric: Electric charges are induced on direct or inverse effect:
the crystal surface by exerting a compression Piezoelectric sensors are used for knock
or elongation force along certain crystal axes. control in gasoline engines. For example, they
This electrical polarization arises by shifting detect high-frequency engine vibrations which
positive and negative ions in the crystal rela- is a feature of combustion knock. Conversion
tive to each other by exerting force (see Fig., of mechanical vibration to an electric voltage
Item b). The shifted centers of charge gravity is also used in the crystal audio pickup of a
within the crystal balance each other out auto- record player or crystal microphones. In the
matically, but an electric field forms between piezoelectric igniter (e.g. in a cigarette lighter),
the end faces of the crystal. Compressing mechanical pressure produces the voltage
and elongating the crystal create inverse field needed to generate a spark.
directions. On the other hand, if an alternating volt-
On the other hand, if an electrical voltage age is applied to a piezoelectrical crystal, it
is applied to the end faces of the crystal, the vibrates mechanically at the same frequency
effect reverses (inverse piezoelectric effect): as the alternating voltage. Oscillating crystals
The positive ions in the electric field migrate of this type are used as stabilizers in electrical
toward the negative electrode, and negative oscillating circuits or as piezoelectric acoustic
Principle of the
ions toward the positive electrode. The crystal sources to generate ultrasound, for instance.
piezoelectric effect
then contracts or expands depending on the When used in clocks, the oscillating quartz (illustrated on the
direction of the electric field strength (see is excited by an alternating voltage whose example of a unit cell)
Fig., Item c). frequency is the same as the quartz’s natural
frequency. This is how an extremely time-con- a Quartz crystal SiO2
The following applies to piezoelectric field stant resonant frequency is generated. In a
b Piezoelectric effect:
strength Ep: calibrated quartz, it deviates by only approx.
When the crystal
Ep = d Dx/x 1/1,000 second per year.
is compressed,
Dx/x: relative compression or elongation
negative O2– ions
d: piezoelectric coefficient, numeric values shift upward,
109 V/cm to 1011 V/cm positive Si4+ ions
shift downward:
electric charges
a b c +
Si4+ are induced at the
+ crystal surface.
O2¯ ¯ ¯ + ¯ ¯ +
¯ ¯ ¯ ¯ ¯ ¯ c Inverse piezoelectric
effect:
+ + + + When an electrical
+ +
SAN0170Y
+ + ¯ + + voltage is applied,
¯
O2– ions shift
¯
upward, Si4+ ions
¯
shift downward: the
crystal contracts.
234 Sensor measuring principles | Temperature sensors
Temperature sensors Generally speaking, a temperature sensor

should reflect this dependency correctly,
Measured variables in other words, it should reflect as accu-
Temperature is defined as a nondirectional rately as possible the local distribution of
quantity which characterizes the energy the temperature and its change as a func-
state of a given medium, and which can be tion of time.
a function of time and location:
T = T (x, y, z, t) (1) In special cases, possibly for functional
where: x, y, z are the spatial coordinates, reasons, this stipulation can be relaxed
t is time, and T is measured according to somewhat. The call for high local resolu-
the Celsius or Kelvin scale. tion and for high-speed response both de-
mand that the sensor should be as small as
Generally speaking, with monitored media possible, since it should not falsify the
which are in gaseous or liquid form, mea- temperature readings by itself absorbing
surements can be taken at any point. In the heat. In other words it should have a low
case of solid bodies, measurement is usu- thermal capacity.
ally restricted to the body’s surface. With In order to ensure that the temperature
the most commonly used temperature sen- assumed by the sensor remains indepen-
sors, in order for it to assume the medi- dent of the usually very different temper-
um’s temperature as precisely as possible, ature of its mounting, it should be ther-
the sensor must be directly in contact with mally well insulated from it. Since this
the monitored medium (direct-contact also falsifies the measurement, the heat
thermometer). In special cases though, developed by the majority of sensors in
proximity or non-contacting temperature the active state should be kept to a mini-
sensors are in use which measure the me- mum (e.g. < 1 mW).
dium’s temperature by means of its (infra- The temperature sensor’s dynamic re-
red) thermal radiation (radiation ther- sponse is specified by the time constant t.
mometer = Pyrometer, thermal camera). This defines the time taken by the sensor
to reach 63 %, 90 %, or 99 % of its final
reading when subjected to a jump in
temperature. This time depends not only
1 Temperature measuring points in the motor vehicle
upon the sensor’s thermal capacity, but
Measuring point Temperature range °C
also to a great extent upon the heat-trans-
Intake/charge air –40 to 170 fer coefficient between the sensor and
External environment –40 to 60 the monitored medium. The higher it is,
Interior –20 to 80 the faster the sensor reaches its final
Exhaust air/heating system –20 to 60 reading. Naturally, this figure is far higher
Evaporator –10 to 50 for liquid mediums than for gaseous me-
(air-conditioning system)
diums. It must also be noted that the heat-
Cooling water –40 to 130
transfer coefficient is highly influenced
Engine oil –40 to 170
by any flow velocity y in the medium. It
Battery –40 to 100 ––
increases with approximately √ y. In other
Fuel –40 to 120
words, the time constant of a temperature
Tire air –40 to 120
sensor should always be specified with
Exhaust gas 100 to 1,000
reference to a specific flow velocity of a
Brake caliper –40 to 2,000
Table 1 defined medium.
Sensor measuring principles | Temperature sensors 235
Temperature measurement in the vehicle Measuring principles for

makes use almost exclusively of the tem- direct-contact sensors
perature-dependence of electrical resis- The fact that practically all physical pro-
tance material with positive (PTC) or nega- cesses are temperature-dependent means
tive (NTC) temperature coefficient. The that there are almost just as many methods
direct-contact thermometers apply this for making temperature measurements.
phenomena. For the most part, the conver- The preferable methods though are those
sion of the temperature change to an ana- in which the temperature effect is very dis-
log voltage takes place by adding a second tinctive and dominant and as far as possi-
resistor to form a voltage divider (which ble features a linear characteristic curve.
also has a linearizing effect). Recently Furthermore, the measuring elements
though, non-contacting (pyrometric) tem- should be suitable for inexpensive mass-
perature sensing has been considered for production, whereby they should be ade-
occupant-protection (passenger-position quately reproducible and non-aging.
monitoring for airbag triggering), and for Taking these considerations into account,
comfort and convenience (air conditioning the following sensor techniques have come
in accordance with skin-temperature mea- to the forefront, some of which are also
surement, prevention of windscreen mist- applied in automotive technology:
up). It needed the advent of microsystem
technology for this (pyrometric) method to Resistive sensors
become feasible from the costs viewpoint. In the form of 2-pole elements, tempera-
Table 1 presents a listing of the temperature-dependent electrical resistors are par-
tures which have to be measured in the ticularly suitable for temperature measure-
vehicle. ment, no matter whether in wire-wound,
sintered ceramic, foil, thin-film, thick-film,
It is not only the highly differing measur- or monocrystalline form. Normally, in or-
ing ranges that demand a variety of differ- der to generate a voltage-analog signal
ent sensor concepts and technologies, but they are combined with a fixed resistor R V
the different accuracies and dynamic re- to form a voltage divider, or load-indepen-
sponse requirements also lead to very dif- dent current is applied (Fig. 1). The volt-
ferent sensor shapes. At many locations, age-divider circuit changes the original
temperature is also measured in order sensor characteristic R(T) to a slightly
that it can be compensated in cases where different characteristic U(T):
temperature variations trigger faults or
act as an undesirable influencing variable.
R(T)
U(T) = U0 · –––––––––––––– (2)
R(T) + RV
1 Methods for resistance/voltage conversion
On the other hand, the application of a

U0
supply current I0 permits the resistor Fig. 1
curve to be reproduced exactly: I0 Current supply
RV I0 U0 Supply voltage
U(T) = I0 · R(T) (3) RV Temperature-
independent
series resistor
The measurement sensitivity is reduced
R(T) Temperature-
R(T) UA(T) R(T) UA(T) more or less depending upon the con-
NAE0820Y
dependent
struction of the voltage-divider circuit. measuring
Notwithstanding this fact, in the case of resistor
resistance characteristics which display a UA(T) Output voltage
slightly progressive curve it does have a oxidized mixed crystals. These are sin-
linearization influence (an effect which is tered in bead form or disc form (Fig. 3) and
usually very welcome). Very often, in this have a polycrystalline structure. Because
connection, the auxiliary resistor is so di- of their very steep temperature curve, they
mensioned that it equals the measuring re- are also known as NTC resistors or therm-
sistor at a given reference temperature T0 istors. To a good approximation, and by
(e.g. 20 °C): applying the exponential law, their charac-
teristic curve can be defined as follows:
RV ≈ R (T0) (4)
If the precision is inadequate, a resistance R(T ) = R0 · e

1–1
(
B · –– ––
T T0 ) (5)
sensor can be calibrated to the setpoint

value both with regard to its resistance (re- Where R0 = R(T0),
ferred to a reference temperature), and its B = 2,000 to 5,000 K = const,
temperature coefficient (TC), by means of T absolute temperature
an adjustable parallel resistor RP and a
serial resistor RS (Fig. 2). Of course, when Here, the characteristic-curve gradient
fixed resistors are added the TC is reduced (TC) or the percentage resistance change
and the characteristic changes somewhat. as a function of temperature, is highly
dependent upon the working point.
Sintered-ceramic NTC resistors This means that it can only be defined
As a result of their pronounced measuring at given points:
effects and inexpensive manufacture, the
most common semiconductor resistors in TC = – B/T 2 (6)
use are based on heavy-metal oxides and
2 Calibration of a resistor to its setpoint value 3 NTC resistors (examples)
a
a 1 RS RNi
Fig. 2
a Resistance sensor
b Characteristic R(T) t b
curves
1 Auxiliary contacts T
2 Bridge
RNi Nickel-film resistor 2 RP c
Rtot(T) Total resistance b 10 5
RP Increase
referred to Ω
temperature T
RP Adjustable
10 4
Resistance R T
parallel resistor
Curve-slope
RS Adjustable
calibration
series resistor
Total resistance Rtot
10 3
RS Increase
Fig. 3
a Pearl form 10 2
Curve-position
UAE0821E
UAE0822E
b Disk form
c Characteristic
calibration
-50 0 50 100 150
curve with limits
Temperature T Temperature T
of variation
It reduces considerably along with increas- Even though the b coefficient is generally
ing temperature, and the resistance value very small for metals, it cannot be ignored
itself frequently varies by 4 to 5 powers of completely. This is why the measurement
ten, e.g. typically from a few 100 kΩ to a sensitivity of such sensors is usually char-
few 10 Ω. This pronounced temperature- acterized by means of a mean TC, the
dependence means that applications are “TC 100”. The TC 100 corresponds to the
restricted to a “window” of about 200 K. mean curve gradient between 0 °C and
This though can be selected in the range 100 °C (Table 2 and Figure 4).
from –40 to approximately 850 °C. Tighter
tolerances of up to ± 0.5 K at a given refer-
R(100 °C) – R(0 °C)
ence point are complied with by using a se- Where TC 100 = –––––––––––––––––––––––––– (8)
R(0 °C) · 100K
lection process, or possibly even by grind-
ing in oil, a method which of course has an
affect upon costs. The aging stability of
these sensors has been vastly improved
compared to earlier versions which means
that it is quite possible for the very close 2 Temperature coefficient TC 100
tolerances to apply throughout the sensor’s Sensor TC 100 Characteris- Measuring

useful life. material tic curve range
10–3/K
Nickel 5.1 slightly –60 to 320
PTC thin-film/thick-film metallic resistors
(Ni) progressive
The thin-film metallic resistors placed with
Copper 4.1 slightly –50 to 200
two additional temperature-neutral cali-
(Cu) progressive
bration resistors on a common substrate
Platinum 3.5 slightly –220 to 850
wafers have particularly high accuracy val- (Pt) degressive Table 2
ues. They can be tightly toleranced in re-
spect of their characteristic curve and can
be manufactured to be stable over the long 4 Definition of the mean temperature coefficient
TC 100 = a100
term and are additionally trimmed by laser
cuts. The use of layer technology makes it
possible to adapt the substrate (ceramic,
glass, plastic foil) and the upper layers
(plastic molding or paint, sealed foil, glass
and ceramic coatings) to the respective
application, and thus provide protection
against the monitored medium. Although R100
metallic layers are less sensitive to thermal
variations than the ceramic-oxide semi-
Resistance R
conductor sensors, both linearity and re- α(100C)

α 100
producibility are better: The following ap- R(T)
plies for the computational definition of α(0C)
these sensors: R0
R(T) = R0 (1 + a · ∆T + b · ∆T 2 + ..) (7) 0 100 C

Measurement temperature TM
where ∆T = T – T0 and
UAE0823E
T0 = 20 °C (reference temperature),
a linear temperature coefficient (TC),
b square temperature coefficient.
Although platinum (Pt) resistors have the Monocrystalline silicon semiconductor

lowest TC, they are not only the most pre- resistors (PTC)
cise resistive temperature sensors, but can Basically speaking, with temperature sen-
also boast the best aging stability. They are sors of monocrystalline semiconductor
available on the market under the designa- materials such as silicon (Si), it is possible
tion “PT 100” or “PT 1000” (100 Ω or to incorporate further active and passive
1,000 Ω nominal resistance at a reference circuit elements on the sensor chip (initial
temperature of 20 °C) in a variety of differ- signal conditioning at the measuring
ent tolerance classes (up to 0.1 °C) (Fig. 5). point). Due to the closer tolerances in-
At temperatures up to about 1,000 °C, volved, their production uses the “spread-
thick-film platinum sensors are suitable ing resistance” principle (Fig. 6a). The cur-
whose platinum layer has been stabilized rent flows through the measuring resistor
by special additives. and through a surface-point contact before
arriving at the silicon bulk material. It then
Thick-film resistors (PTC/NTC) proceeds, widely distributed, to a counter-
Thick-film pastes with both high resis-
tance (low surface-area requirement) and
positive and negative temperature coeffi- 6 Spreading-resistance principle (dual-hole version)
cients are generally employed as tempera-

ture sensors for compensation purposes.
They have a non-linear characteristic
curve (which though is not as “bent” as a R(T)
the curve of the solid NTC resistors), and
1 2 1 3
are suitable for laser-trimming. The mea-
suring effect is improved by forming
voltage-divider circuits from NTC and
PTC materials.
5 Pt resistor (tolerance diagram) 4

b
kΩ
2.0
3
Temperature tolerance ΔT
1.6
Resistance R
1 1.2
Fig. 6
0
a Design 0.8
b Characteristic -1
curve 0.4
1 Contacts -2
2 Passivation 0
-3
(Nitride, oxide) -100 0 100 200 C
3 Si substrate -200 0 200 400 C Measurement temperature TM
4 Counter- Measurement temperature TM
UAE0824E
UAE0825E
electrode without
connection
R(T) Temperature-
dependent resistor
electrode covering the base of the sensor Thermocouples

chip. In addition to the high reproducibil- Thermocouples are used in particular for
ity material constants, the high current measuring ranges ≥ 1,000 °C. They rely on
density behind the contact point (high ac- the Seebeck effect, according to which
curacy thanks to photolithographic pro- there is a voltage across the ends of a me-
duction) mainly defines the sensor’s resis- tallic conductor when these are at differ-
tance. In order to make the sensor highly ent temperatures T1 and T2. This “thermo-
independent of polarity, these sensors are electric voltage” Uth depends (independently
usually series-connected in pairs in oppo- of the development of this) exclusively on
site alignments (dual-hole version, Fig. 6). the temperature difference ∆T at the ends
The base electrode can be in the form of of the conductors (Fig. 8). It is expressed
a temperature contact (no electrical func- by the equation:
tion).
Measurement sensitivity is practically dou- Uth = c (T2 – T1 ) = c ∆T, (9)
ble that of a Pt resistor (TC = 7.73 · 10–3/K).
The progressive bend of the temperature where the proportionality constant is ma-
curve is more pronounced than on a metal- terial-specific and termed the Seebeck
lic sensor. The measuring range is limited coefficient.
to approx. +150 °C by the material’s intrin- Since the instrument leads used to mea-
sic conductivity (Fig. 6b). There are special sure this voltage across the metallic con-
versions (Fig. 7) available for uses up to ductor must themselves be equipped with
300 °C. terminals (for instance made of copper),
7 Spreading resistance sensor (unipolar version for 8 Seebeck effect

temperatures up to 300 °C)
Uth
T2 > T1
Fig. 8
1 1 Heat source
2 Metallic conductor
A 2
+ High,
3 – low thermal velocity
of the electrons
1 2
NAE0827Y
T2 High temperature
1
T1 Low temperature
Uth Thermoelectric
voltage
3 Thermoelectric voltage Uth of a number of metals

n+ Material Thermoelectric voltage Uth
Detail A mV/100 °C
Constantan –3.40
Nickel –1.90
Paladium –0.28
Platinum 0.00 Fig. 7
Copper +0.75 1 Metal wire

NAE0826Y
2 Glass
Manganite +0.60
3 Si crystal
Iron +1.88
Silicon +44.80 Table 3
these are also subject to the same temper- of the thermoelement configuration are at
ature difference, so that unfortunately the same (reference) temperature, other-
only the difference between the metallic wise the temperature difference at the free
conductor and the connecting cables is ends will also be included in the measure-
measured. Thermoelectric voltages are ment. Thermocouples, therefore, always
always listed based on Platinum as the measure only the temperature difference
reference material (Table 3). to a given reference point. If the measur-
In order that the generated voltages are ing point’s absolute temperature is to be
as high as possible, a number of material measured, other devices (such as resistive
pairs have established themselves (Fig. 9, sensors) must be used to also measure the
e.g. iron/constantan etc.). It is important temperature at the reference point.
Fig. 9
that the “limbs” of such a thermocouple The characteristic curve of the thermo-
1 Copper/Constantan
are joined at the end to which the heat is electric voltage against temperature is usu-
2 Iron/Constantan
3 Nickel-chromium/ applied in such a manner that the joint is ally not as linear as that given in equation
Nickel electrically conductive (by means of twist- (9). The sensor signals are usually small, and
4 Platinum rhodium/ ing, welding, soldering, etc., Fig. 10). ICs are used for their amplification and for
Platinum their linearization. In order to increase the
Thermocouples themselves are usually measurement voltage, it is common practice
short, and the extensions up to the point to connect a number of identical thermocou-
Fig. 10
A/B Material pair
where the signal is captured can be made ples in series. These have their “hot” junc-
(thermocouple with compensating cables which use the tions at the temperature to be measured,
legs) same material pair as in the thermocouple and their “cold” junctions at the reference
1 Measuring point itself. It is important that both free ends temperature (Fig. 11, thermopile).
(electrically
conductive 9 Common thermocouples (characteristic curves) 11 Thermocouples connected in series
junction)
2 Connection head mV
Thermoelectric voltage Uth
3 Compensating
cable 40 a TM
4 Reference point 1 2 3
30 2
5 Connection cable
(Cu)
20
TM Measurement
temperature 4
10 3
TR Reference
temperature 0
NAE0828E
TR
Uth Thermoelectric 0 400 800 1,200 C
voltage Measurement temperature TM
Fig. 11 b
10 Thermocouple measuring set-up
a Principle of the
thermopile
b Example of A 3 TR
application
1 Sensitive surface TM
2 “Hot” junctions at
the measurement 1 Uth
temperature TM 1 2 3 4
(TM ) (TR )
NAE0829E
3 “Cold” junctions
NAE0830Y
at the reference
B 2 3 4 5
temperature TR
4 Thermopile
Although thermocouples are robust (for sensor, and may necessitate additional cal-
instance, high-level EMC due to low inter- ibration elements in order for a precise
nal resistance), they are not particularly measurement to be made. Above all, it is
accurate as a measuring device. Their the temperature-dependent saturation
deviation can easily be in the 5° to 15 °C current Isat which is responsible for the
range, and they are not outstanding re- negative temperature coefficient (NTC).
garding their resistance to aging, which This current increases dramatically with
means that individual calibration does rising temperature. Here, the silicon’s in-
not result in a permanent improvement trinsic conductivity limits this sensor’s ap-
of their accuracy. plication to temperature ranges <150 °C.
Of course, thermocouples can be manu- Sometimes, emitter-coupled transistor
factured using both thin-film and thick- pairs are used in a similar manner for tem-
film techniques. Metallic films stacked one perature measurement. With this form of
on top of the other provide for excellent temperature measurement, the ratios of
thermal contact, and extremely small ther- the collector currents to each other repre-
mocouples can be produced by applying sent a very good reproducible measure for
microsystem technology. Thermocouples the temperature. Usually, an integrated
are particularly suitable for use in thermo- supplementary circuit is used for the
piles comprised for instance from 50 to “on-chip” conversion to an analog output
100 individual thermocouples. They are voltage.
used in non-contacting radiation ther-
mometers (pyrometers). Zener diodes operated in the reverse di-
rection can also be used as highly practical
Semiconductor depletion layers temperature sensors. Their voltage changes
Presuming a constant current, the forward are highly dependent upon the Zener volt-
voltage of semiconductor depletion layers age. Here, the option exists of various lev-
(Fig. 12) such as those in diodes and in the els of voltage reduction at Zener voltages
basis-emitter path of a transistor, demon- <4.7 V, and voltage increases at Zener volt-
strate very good linearity as a function of ages >4.7 V.
temperature: Such sensors are often used for temper-
ature compensation on the chip itself.
k · T · ln (––––
UF(T) = ––––––– IF + 1) (10)
q Isat
Where:
Isat = Isat (T) and IF = constant, 12 Semiconductor depletion layers
q = 1.6 · 10 –19 C (elementary charge),

k = 1.88 · 10 –23 JK–1 (Boltzmann’s constant),
T absolute temperature. a b
C Fig. 12
a Diode
It is advantageous here when the sensor IF b Transistor
IF
directly outputs a voltage which is a func- B B Base
UF C Collector
tion of the temperature. The two-pole sen-
E Emitter
sor is of course dependent upon polarity.
IF Conducting-state
Whereas, for all sensors, the forward volt- UBE
NAE0831Y
current
age decreases by almost exactly 2 mV/°C, E UF Forward voltage
the absolute voltage at each depletion UBE Voltage between
layer differs considerably from sensor to base and emitter
Measuring principles for non-contacting ture of the body being measured. A given
temperature measurement temperature difference at the object often
The radiation emitted by a body is used for corresponds to only 1/1,000th of this dif-
the non-contact measurement (pyrometry) ference at the measuring point. Neverthe-
of its temperature. This radiation is for the less, the object’s temperature can be deter-
most part in the infrared (IR) range (wave- mined with an accuracy of 0.5 °C.
length: 5 to 20 µm). Strictly speaking, the
product of the radiated power and the Bolometer
emission coefficient of the body is mea- The Bolometer is a highly sensitive resis-
sured. The latter is a function of the mate- tance temperature sensor for measuring
rial, but for materials which are technically minute temperature increases (Fig. 13).
of interest (including glass) it is usually A further sensor is needed for measuring
around 1, although for reflective and IR- the temperature of the sensor housing. If
permeable materials (e.g. air, silicon) it is this device is to operate efficiently across
far less than 1. a wide temperature range, it is necessary
The measuring point is reproduced on a though for both these sensors to feature
radiation-sensitive element, which heats an extremely high degree of synchronism.
up slightly with respect to its environment The Bolometer housing is therefore usu-
(typically 0.01 to 0.001 °C). The element’s ally thermostatically controlled (and well-
temperature is a measure for the tempera- insulated to the housing) so that the pri-
Fig. 13 mary detecting element (sensor) always
1 Lens housing with 13 Prototype of a bolometric sensor array for operates at the same temperature.
automotive applications
lens
2 Connections Thermopile sensor
3 Infrared detector
When a very extensive temperature range
housing
is concerned, it is more practical for the
4 Infrared window
5 Detector temperature difference generated by the
radiation from the object to be measured
4 5
using thermocouples. In order to increase
Fig. 14 the measuring effect, a number of thermo-
a Principle of the
couples are connected in series (thermo-
measuring cell 3
2 pile). Such a thermopile sensor (Fig. 14) is
UAE0832Y
1 Silicon chip
2 Thermocouples
inexpensive to manufacture micromechan-
1 ically. All its “hot” junctions are located on
connected in series
(e.g. Al/poly-Si) a thermally well insulated thin diaphragm,
3 SiN diaphragm
4 Thermopile 14 Pyrometric sensor produced using micromechanical techniques, with thermopile pick-off.
terminals
5 Absorber layer
a 2 5 b 5 6 7 8
b Sensor type
1 Thermocouple
2 “Cold” junction 1
3 Diaphragm 2
4 Absorber
5 Heat radiation
6 Electrical
connection
UAE0833Y
7 Si3N4 layer
8 SiO2 layer 3 4 1 3 4 9
9 Heat sink
and all its “cold” junctions are in contact sive curved mirror is usually ruled out due
with the thicker chip rim (heat sink). Typi- to it needing too much room. Glass lenses
cally, the sensor’s response time is approx. are impermeable for IR light, and plastic
20 ms. Using such a so-called “single-pixel lenses can only be used for operating tem-
sensor”, it is an easy matter to determine peratures of up to approx. 85 °C max. On
the windshield’s surface temperature so the other hand, Si lenses are highly suit-
that measures can be taken to prevent able for thermal radiation and up to diam-
misting should the temperature fall below eters of approx. 4 mm micromechanical
the dew point. techniques can be used to inexpensively
manufacture them in the form of a Fresnel
Single-point sensors, imaging sensors or refraction lens. Fitted in the cover of
If a number of pixels are combined on a a TO5 housing, these then also serve to
single chip (e.g. 4 x 4) to form an array, protect the sensor against direct damage
this provides the basis for a rough form (Fig. 15). Even though filling the housing
of image analysis (Fig. 16). However, there with an inert gas improves the crosstalk
should not be too much insensitive surface between the individual pixels, it also nega-
area between the pixels and the pixels tively affects their response time.
themselves must be thermally well insu-
lated against each other. As all the pixels
can optionally respond electrically, the 16 Micromechanical thermopile array
chip has a large number of terminals. On

a TO5 housing, for example, the ASIC for
3 2
signal preamplification and series connec-
tion of the signal must be located directly
adjacent to the sensor chip. Usually, in the
case of thermopile sensors, this ASIC also
includes a reference-temperature sensor
which measures the pixel’s absolute tem-
perature. This permits object tempera-
tures to be measured with an accuracy
UAE0714-1Y
of approx. ±0.5 K. Fig. 16

In order to reproduce a scene thermally 4 4 1 1 Silicon chip
on the sensor array, the array requires an 2 Pixel
IR optical imaging unit. The very inexpen- 3, 4 Pixel connections
15 Thermal image analysis Fig. 15

a IR imaging sensor
a α b 1 Silicon IR lens
2 TO5 housing
3 Terminal posts
1 4 Sensor chip
1 2 3 4 5
5 Evaluation ASIC
2
a Viewing angle
b Simple IR camera
1 Electronics
2 Lens system
3 Camera’s field
UAE0834Y
of view
3 4 5 4 Si IR lens
5 Sensor array
244 Sensor measuring principles | Imaging sensors (video)
Imaging sensors (video) Imaging sensors are a special case of mul-

tisensor structures comprising light-sensi-
In particular, imaging sensors are begin- tive elements (photodiodes), arranged as
ning to gain a hold in the motor vehicle us- pixels in a row or matrix form and receiv-
ing visible light or infrared light. They can ing the light through a conventional imag-
be used for passenger-compartment moni- ing optical system. In the case of the sili-
toring, but are principally aimed at obser- con imaging sensors principally available
vation outside the vehicle. (CCD, Charge-Coupled Devices, Figs. 1 and
All of these sensors have one objective 2), free charge carriers are generated pro-
in view, and that is the simulation of the portionally to the intensity and exposure
superior capabilities of the human eye time by the light arriving through a trans-
and its mental recognition capabilities parent electrode; they are then collected
(of course, only to a very modest degree in a “potential pot” (Si-SiO2 boundary
at first). These were introduced in large layer). Further electrodes move these
numbers to industrial measurements some charges into an opaque zone which are
time ago – in particular on robot handling further transported in analog shift regis-
equipment. The costs of imaging sensors, ters (bucket chain principle) line-by-line
and the associated very high-performance into an output register which is read seri-
processors needed for the interpretation ally at a high clock rate.
of a scene, are already of interest for appli-
cations in the automotive sector. 1 CCD principle
In contrast to the human eye, common

imaging sensors are sensitive in the near 1 2 3 4 5 6
IR range (wavelength approximately
1 mm). With appropriate non-visible SiO2
IR illumination, therefore, all imaginable
applications in the vehicle become feasi-
ble, including nighttime operation. For
p-Si
applications in the visible light range, the
imaging sensors generally have an IR filter
in front of them to prevent false recordings
due to color.
t1
In future, imaging sensors will be able to
play a highly variegated role for the obser-
vation of the vehicle’s interior (seating po-
sition, forward shift in case of a crash, size
of seat occupants etc.), and of the vehicle’s t2
surroundings (lane guidance, collision
prevention, parking and reversing aids,
road-sign recognition). A night vision de-
vice, that takes the IR image recorded for t3
the road and displays it on a screen in the
dashboard, thus offering enhanced driving
Fig. 1
safety in poor visibility (darkness, fog, etc.)
1 Photodiode
is already in series production. SH
2 Light
UAE0716-2Y
3 Storage electrode F1
4 Shift gate
5 Transfer electrode t1 t2 t3
6 Optical masking
Sensor measuring principles | Imaging sensors (video) 245
While CCD sensors can only be used in 3 CMOS imaging sensors with random pixel access
motor vehicles to a very limited degree
because of their limited brightness/dark-
ness dynamics (50 dB), their read time and
15 bit address bus
their temperature range (< 50 °C), newer,
smart, imaging sensors based on CMOS
technology appear to be fully suitable for Gap Gap
use in the motor vehicle. The logarithmic decoder decoder
1 of 128 1 of 128
brightness/signal curve possible here,
plexer and buffer
plexer and buffer

Black reference
corresponds to the human eye and has a
Analog-multi-
Analog multi-
Line decoder
Line decoder
Sensor Sensor
1 of 128
1 of 128
element
dynamic response of 120 dB. This makes field 1 field 0
aperture control, for instance, unneces-
128 x 128 pixels
sary and supplies a constant contrast reso-
lution across the entire brightness range.
These sensors permit a digitally controlled,
random access to the individual pixels in
Video switch
a matrix sensor (Fig. 3) with simultaneous
higher sensitivity (higher read rate).
The first steps in preprocessing the Buffer Buffer Buffer Buffer
signals on the imaging-sensor chip have
already been implemented (intelligent Fig. 3
imaging sensor). A resolution of – + – + – + Pixel processors at
1,280 × 1,024 pixels with typical indi- AOUT_1 AOUT_c AOUT_0 each pixel in the active
vidual dimensions of 6 µm × 6 µm is pixel matrix ensure
3 differential analog outputs
logarithmic signal
already possible in sensors currently
compression.
on the market. Iphoto X Y
Reading is random
SAE1080E
logU access and non-

destructive to the
signal, transmission
is loss-free.
2 CCD imaging-sensor structure
2
A1
1
A2
3
Fig. 2
1 Column clock pulse
A1/A2
2 Photosensors
3 CCD array
UAE0717-3Y
B1 4 Line clock pulse

4 6 B1/B2
B2 5 Output register
5
6 Video output
246 Sensor types | Engine-speed sensors
Sensor types
Engine-speed sensors engine speed (Fig. 2). The amplitude of the

alternating voltage increases strongly
Application along with increasing pulse-wheel speed
Engine-speed sensors are used in engine- (several mV to >100 V). At least about
management systems for 30 rpm are needed to generate an ade-
• Measuring the engine speed and quate amplitude.
• Determining the crankshaft position
(position of the pistons) The number of teeth on the pulse wheel

depends on the particular application. In
The engine speed is calculated from the in- Motronic systems, a 60-pitch pulse wheel
terval between the speed sensor’s signals. is normally used, although 2 teeth are omit-
ted (Fig. 1, Item 7) so that the pulse wheel
Inductive speed sensors has 60 – 2 = 58 teeth. The space where the
Design and operating principle missing teeth would be situated is allocated
The sensor is mounted directly opposite a to a defined crankshaft position and serves
ferromagnetic pulse wheel (Fig. 1, Item 7) as a reference mark for synchronizing the
from which it is separated by a narrow air control unit.
gap. It has a soft-iron core (pole pin, 4), The geometries of the pulse-wheel teeth
which is enclosed by a winding (5). The and the pole pin must be matched to each
pole pin is also connected to a permanent other. An evaluation circuit in the control
magnet (1), and a magnetic field extends unit converts the sinusoidal voltage, which
through the pole pin and into the pulse is characterized by strongly varying ampli-
wheel. The level of the magnetic flux tudes, into a constant-amplitude square-
through the coil depends on whether the wave voltage for evaluation in the control
sensor is opposite a pulse-wheel tooth unit microcontroller.
or space. Whereas the magnet’s leakage
flux is concentrated by a tooth, and leads Active speed sensors
to an increase in the magnetic flux through Active speed sensors operate according to
the coil, it is attenuated by a space. When the magnetostatic principle. The amplitude
the pulse wheel rotates, these changes in of the output signal is not dependent on the
magnetic flux induce a sinusoidal output rotational speed. This makes it possible for
voltage in the coil which is proportional to very low speeds to be sensed (quasistatic
the rate of change of the flux and thus the speed sensing).
1 Inductive engine-speed sensor (design) 2 Signal from an inductive engine-speed sensor

Fig. 1
1 2 3
1 Permanent magnet 2 cm
2 Sensor housing 1
3 Crankcase
4 Pole pin
S
5 Winding
Output voltage
6 Air gap
7 Pulse wheel with
N
reference mark 4
UMZ0138-5Y
5 3
6
2
UAE0727-1E
Fig. 2 7
1 Tooth
2 Tooth space Time
3 Reference mark

Sensor types | Engine-speed sensors 247
Differential Hall sensor AMR sensors

A voltage UH proportional to the magnetic The electrical resistance of magnetoresis-
field (Hall voltage) can be picked off hori- tive material (AMR, Anisotrop Magneto
zontally to the current direction at a cur- Resistive) is anisotropic, i.e. it depends on
rent-carrying plate which is permeated the direction of the magnetic field to which
vertically by a magnetic induction B. In it is exposed. This property is utilized in
a differential Hall sensor, the magnetic an AMR sensor. The sensor is located
field is generated by a permanent magnet between a magnet and a pulse wheel.
(Fig. 3, Item 1). Two Hall sensor elements The field lines change direction when the
(2 and 3) are situated between the magnet pulse wheel rotates (Fig. 4). This generates
and the pulse wheel (4). The magnetic flux a sinusoidal voltage, which is amplified in
by which these are permeated depends on an evaluation circuit in the sensor and
whether the speed sensor is opposite a converted into a square-wave signal.
tooth or a space. Determining the differ-
ence between the signals from two sensors GMR sensors
reduces the magnetic interference signals The use of GMR technology (Giant Mag-
and achieves and improved signal/noise neto Resistive) is a further development
ratio. The edges of the sensor signal can of active speed sensors. Because of the
be processed without digitization directly higher sensitivity by comparison with the
in the control unit. AMR sensors, larger air gaps are possible,
Multipole wheels are used instead of the which makes applications in difficult cir-
ferromagnetic pulse wheel. Here, a magne- cumstances possible. The higher sensitiv-
tizable plastic is attached to a nonmagnetic ity also results in less background noise
metallic carrier and alternately magne- from the signal edges.
tized. These north and south poles adopt All two-wire current interfaces previ-
the function formerly performed by the ously used for Hall speed sensors may
teeth of the pulse wheel. also be used on GMR sensors. Fig. 3
a Configuration
b Signal of Hall
sensor
– high amplitude
with small air gap
– low amplitude
3 Principle of differential Hall sensor 4 Principle of speed sensing with an AMR sensor
with large air gap
c Output signal
UAE0993-1Y
a 1 S
N a 1 2 3 1 Magnet
2 3 2 Hall sensor 1
3 Hall sensor 2
4 4 Pulse wheel
b 5 6 b
Fig. 4
7 a Configuration at
various times
b Signal from
c AMR sensor
c Output signal
UAE0994-1E
c 1 Pulse wheel
Time 2 Sensor element
3 Magnet
248 Sensor types | Hall phase sensors
Hall phase sensors piston is in the compression or exhaust

stroke. The phase sensor on the camshaft
Application provides the control unit with this infor-
The engine’s camshaft rotates at half the mation. This is required, for example, for
crankshaft speed. Taking a given piston on ignition systems with single-spark ignition
its way to TDC, the camshaft’s rotational coils and for sequential fuel injection (SEFI).
position is an indication as to whether the
Design and operating principle
1 Hall rod sensor (design) Hall rod sensors
Hall rod sensors (Fig. 1a) utilize the Hall
a
1 effect: a rotor of ferromagnetic material
2
(7, pulse wheel with teeth, segments or
aperture plate) rotates along with the cam-
Fig. 1
a Positioning of shaft. The Hall IC (6) is located between
2 cm
sensor and single- 3 the pulse wheel and a permanent magnet

track pulse wheel 4 (5), which generates a magnetic field
S
b Output signal
5 N strength perpendicular to the Hall ele-
characteristic UA
6 a ment.
If one of the pulse-wheel teeth (Z)
1 Electrical Z
connection (plug)
now passes the current-carrying sensor
7 L
2 Sensor housing ϕ element (semiconductor wafer), it changes
3 Crankcase the magnetic-field strength perpendicular
4 Sealing ring to the Hall element. This results in a volt-
5 Permanent magnet b age signal (Hall voltage) which is indepen-
6 Hall IC L L
UA ϕs High dent of the relative speed between sensor
7 Pulse wheel with
and pulse wheel. The evaluation electron-
UMK1768Y
tooth/segment (Z)
Z
and space (L) Low ics integrated in the sensor’s Hall IC condi-
a Air gap tion the signal and output it in the form of a
Angle of rotation ϕ
w Angle of rotation square-wave signal (Fig. 1b).
2 Generations of camshaft sensors
PG-1 PG-3-3 PG-3-5 PG-3-8

TIM no no yes yes
Fig. 2
TIM = Twist Intensive
TPO no yes no yes
Mounting (i.e. the Accuracy low medium medium high
sensor may be
twisted around its
axis any number of
times, without losing
accuracy. Important
for minimizing type
diversity).
TPO = True Power On ess
(i.e. the sensor detects progr
cal
directly on switching
ologi
on whether it is located hn
over a tooth or a space. Tec
UAE0995E
Important for short

synchronization times PG-1 PG-3-3 PG-3-5 PG-3-8
between crankshaft and
camshaft signals).
Sensor types | Speed sensors for transmission control 249
Speed sensors for Because of the very compact design of the

transmission control transmissions, the mechanical customer
interface cannot generally be covered
Application by standard geometries. This means that
Transmission speed sensors determine custom sensor types are needed for each
the shaft speeds in AT, ASG, DKG and transmission model. Module-integrated
CVT transmissions. These are the turbine types differ in respect of insertion length,
and output shaft speeds on AT transmis- direction of recording and mounting
sions with a hydrodynamic torque con- flange (Fig. 1). A further variation found
verter, the speed of the primary and sec- in stand-alone sensors is the location of
ondary pulleys in CVT transmissions and the mounting socket and the design of
the speeds of the two input shafts and the the connector.
driven shaft of double-clutch transmis- Hall ASICs (Application Specific Inte-
sions. The speed of the power take-up grated Circuit) with varying complexities
element is also detected for high dynamic of the evaluation algorithms are used to
demands on power take-up control. cover the complete range of the functional
Direction of rotation detection can also requirements (Fig. 2).
be necessary in high-end transmissions in If a ferromagnetic trigger wheel or
order to optimize clutch management and a trigger area (toothed, stamped or em-
for rollback prevention. bossed) is available on the rotating trans-
Both stand-alone sensors and sensors mission component, the magnetic field
built into electronics modules which proj- required to operate the Hall sensor is
ect into the transmission from the outside generated by a back-bias magnet.
or are mounted internally are used.
Requirements 1 Sensor types
Transmission speed sensors are exposed

to very severe operating stresses because a
of
• Extreme ambient temperatures between
-40 and +150 °C
• An aggressive operating environment
caused by the transmission oil, also
known as ATF (contains special addi-
tives for transmissions and has a low
condensate content) 1
• High mechanical stress with vibrational
accelerations up to 30 g
• Abraded metallic materials and a build-
up of particles in the transmission b c
1 1
The package of the electronics used in the
sensor is hence subject to high demands
as a consequence of these severe stresses. Fig. 1
a Bottom read
A service life of more than 15 years in
b Side read
transmission oil has been achieved by
SAE1081Y
c Slant read
using a suitable oil-resistant package.
1 Direction of
detection
250 Sensor types | Speed sensors for transmission control
It is mounted in the sensor immediately Operating principle

behind the ASIC. Transmission speed sensors use the differ-
Compact transmissions increasingly ential Hall effect. The difference between
require speed measurement over greater the Hall voltages from two Hall plates on
distances (magnetic air gap) by rotating, the ASIC is determined. The majority of
nonmagnetic components or through a the similar interferences can be compen-
housing wall. Multipole wheels (magne- sated for in this way. The difference signal
tized rings) are used for these applica- is first amplified on some ASIC types,
tions, the back-bias magnet is omitted then converted into a digital signal by
from the sensor. trigger algorithms of varying complexities.
This constitutes the controlled variable for
Design modulating the output current by means
The Hall ASICs used in the transmission of a power source. A digital signal with
speed sensors are mounted in a retainer two current levels (typically 7 mA for the
with or without a back-bias magnet –
depending on the magnetic interface –
with the electrical contact being made 3 Hall sensor with two-wire current interface
by a welding process, then inserted in

a housing, potted with epoxy resin or – in IS US
the case of types mounted on the outside
of the transmission – have an oil-tight en- UV
RM URM
closure sprayed onto them (Fig. 3).
The sensor has a two-wire interface,
which combines optimum diagnostic op-
tions with a minimum number of electrical
connections. The two terminals both sup-
ply the Hall IC and transmit the signal.
SAE0907-1Y
2 Requirements complexity
Vibration immunity
Direction of rotation detection
High speed range Vibration suppression

Sensor requirement
Speed accuracy Additional third Hall area
Air gap changes Higher cycle frequency
Large Noise suppression

air gap
area Switching threshold tracking
Offset tracking Multiplier tracking

SAE1082E
ASIC complexity
Sensor types | Speed sensors for transmission control 251
low level and 14 mA for the high level) is Some transmission controls include func-
obtained, the modulation frequency of tions that require detection of a standstill.
which corresponds to the frequency of For this application, the sensor must have
tooth change on the trigger wheel and as great an immunity as possible against
thus represents the speed of rotation. vibration-induced variations in the air gap
The sensor signal is evaluated in the elec- and the torsional vibration of the trigger
tronic control unit by means of a measur- wheel. This property of the sensor, known
ing shunt RM, which converts the sensor as immunity against vibration, can only
current IS into the signal voltage URM. be implemented to a very limited degree
On principle, it does not matter for in differential Hall sensors with only two
the operation of a differential Hall ASICs Hall plates, by the use of adaptive trigger
whether the sensor is operated with a thresholds, for example. Two phase-
steel trigger wheel or on a multipole shifted differential signals become avail-
wheel (Figs. 4a and 4b). able with the use of a third Hall plate.
This makes both the detection of the di-
rection of rotation (Figs. 4c to f) and addi-
tional function algorithms for increasing
4 Transmission speed sensor operating principle immunity against vibration possible.
a Rotation
The typical values for “value” and “high Fig. 4
feature” sensors differ in the air gap range a Configuration with
1 achievable (distance between the sensitive trigger wheel
L C R area on the sensor and the trigger wheel), b Configuration with
2
the signal frequency range and the addi- multipole wheel
3 c Sensor signal
tional functionalities implemented (Table 1).
(difference signal
b Rotation between Hall plates
The complexity of transmission type, R and C)
4
N S N
mounting space restrictions including d Sensor signal
all the derived marginal design conditions (difference signal
L C R
2 and the functional requirements leads to between Hall plates
application-specific solutions in most ap- C and L)

c
e Output signal for
plications. These are characterized by a
Difference
R-C clockwise rotation

combination of ASIC, package design and
signal
f Output signal for

sensor mechanical and magnetic interface counterclockwise
suitable for the system requirements. rotation
d
Difference
C-L 1 Trigger wheel

2 Hall plates L and
signal
R (C optional
for direction of
5 1 Typical characteristics rotation detection)
e
Type Value High feature 3 Permanent magnet
(back-bias)
Maximum air gap
Output
signal
4 Multipole wheel
at trigger wheel 2.5 mm 3.5 mm
5 Phase shift
at pole wheel 5 mm 7 mm
dependent on
f Signal frequency 0 to 8 kHz 0 to 12 kHz
the direction of
Direction of rotation No Yes
SAE1083E
rotation
detection
Output
signal
Trigger wheel – ± 1.5°

vibration Table 1
252 Sensor types | Wheel-speed sensors
Wheel-speed sensors Passive (inductive) wheel-speed sensor

A passive (inductive) speed sensor con-
Application sists of a permanent magnet (Fig. 2, 1) with
Wheel-speed sensors are used to measure a soft-magnetic pole pin (3) connected to
the rotational speed of the vehicle wheels it, which is inserted into a coil (2) with sev-
(wheel speed). The speed signals are eral thousand windings. This setup gener-
transmitted via cables to the ABS, TCS ates a constant magnetic field.
or ESP control unit of the vehicle which The pole pin is installed directly above
controls the braking force individually the pulse wheel (4), a gear wheel attached
at each wheel. This control loop prevents to the wheel hub. As the pulse wheel turns,
the wheels from locking up (with ABS) or the continuously alternating sequence of
from spinning (with TCS or ESP) so that teeth and spaces induces corresponding
the vehicle’s stability and steerability are fluctuations in the constant magnetic field.
maintained. This changes the magnetic flux through
Navigation systems also use the wheel- the pole pin and therefore also through
speed signals to calculate the distance the coil winding. These fluctuations in the
traveled (e.g. in tunnels or if satellite magnetic field induce an alternating volt-
signals are unavailable). age in the coil suitable for monitoring at
the ends of its winding.
Design and operating principle The frequency and amplitude of this
The signals for the wheel-speed sensor are alternating voltage are proportional to
generated by a steel pulse generator that wheel speed (Fig. 3) and when the wheel is
is fixed to the wheel hub (for passive sen- not rotating, the induced voltage is zero.
sors) or by a multipole magnetic pulse gen- Tooth shape, air gap, rate of voltage rise,
erator (for active sensors). This pulse gen- and the control unit input sensitivity de-
erator has the same rotational speed as the fine the smallest still measurable vehicle
wheel and moves past the sensitive area speed and thus, for ABS applications,
of the sensor head without touching it. the minimum response sensitivity and
The sensor “reads” without direct contact switching speed.
via an air gap of up to 2 mm (Fig. 2).
The air gap (with strict tolerances) en-
sures interference-free signal acquisition. 1 Passive (inductive) wheel-speed sensors
Possible interference caused for instance

by oscillation patterns in the vicinity of the a
brakes, vibrations, temperature, moisture,
installation conditions at the wheel, etc. is
therefore eliminated.
Since 1998 active wheel-speed sensors

have been used almost exclusively with
new developments instead of passive
b
(inductive) wheel-speed sensors.
Fig. 1
a Chisel-type
pole pin
(flat pole pin)
b Rhombus-type
SAE0974Y
pole pin
(lozenge-shaped
pole pin)
Sensor types | Wheel-speed sensors 253
2 Figure illustrating the principle of the passive wheel-speed sensor
1 2 3 4
Fig. 2
1 Permanent magnet
SAE0975Y
2 Solenoid coil
3 Pole pin
5
4 Steel pulse wheel
5 Magnetic field lines
Various pole-pin configurations and Active wheel-speed sensor

installation options are available to adapt Sensor elements
the system to the different installation con- Active wheel-speed sensors are used al-
ditions encountered with various wheels. most exclusively in today’s modern brake
The most common variants are the chisel- systems (Fig. 4). These sensors usually
type pole pin (Fig. 1a, also called a flat consist of a hermetic, plastic-cast silicon
pole pin) and the rhombus-type pole pin IC that sits in the sensor head.
(Fig. 1b, also called a lozenge-shaped pole In addition to magnetoresistive ICs
pin). Both pole-pin designs necessitate (the electrical resistance changes as the
precise alignment to the pulse wheel magnetic field changes) Bosch now uses
during installation. Hall sensor elements almost exclusively.
These sensors react to the smallest
changes in the magnetic field and there-
fore allow greater air gaps compared to
passive wheel-speed sensors.
3 Signal output voltage of passive 4 Active wheel-speed sensors

wheel-speed sensor
Umax Umin
a
t
Fig. 3
b a Passive wheel-
speed sensor with
t pulse wheel
Voltage
b Sensor signal
c at constant
wheel speed
SAE0976E
SAE0977Y
c Sensor signal
Time t at increasing
wheel speed
254 Sensor types | Wheel-speed sensors
Pulse wheels A steel pulse wheel can also be used in-

A multipole ring is used as a pulse wheel stead of the multipole ring. In this case a
for active wheel-speed sensors. The multi- magnet is mounted on the Hall IC that gen-
pole ring consists of alternately magne- erates a constant magnetic field (Fig. 7b).
tized plastic elements that are arranged in As the pulse wheel turns, the continuously
the shape of a ring on a nonmagnetic metal alternating sequence of teeth and spaces
carrier (Fig. 6 and Fig. 7a). These north and induces corresponding fluctuations in the
south poles adopt the function formerly constant magnetic field. The measuring
performed by the teeth of the pulse wheel. principle, signal processing and IC are
The sensor IC is exposed to the continu- otherwise identical to the sensor without
ously changing fields generated by these a magnet.
magnets. The magnetic flux through the
IC therefore changes continuously as the Characteristics
multipole ring turns. A typical feature of the active wheel-speed
sensor is the integration of a Hall measur-
5 Explosion diagram with multipole pulse generator ing element, signal amplifier and signal
Fig. 7
conditioning in an IC (Fig. 8). The wheel-
a Hall IC with
multipole pulse
speed data is transferred as an impressed
generator current in the form of square-wave pulses
b Hall IC with steel (Fig. 9). The frequency of the pulses is
pulse wheel and 4 proportional to the wheel speed and the
magnet in sensor speed can be detected until the wheel is
practically stationary (0.1 km/h).
1 Sensor element 3
2 Multipole ring
The supply voltage is between 4.5 and
3 Magnet 20 volts. The square-wave output signal
4 Steel pulse wheel level is 7 mA (low) and 14 mA (high).
Fig. 5 2 7 Diagram illustrating principle for measuring speed

1 Wheel hub
SAE0978Y
2 Ball bearing IMR

1
3 Multipole ring a
4 Wheel-speed I
sensor
1
6 Sectional drawing of active wheel-speed sensor
2 1 b
IMR
I
Fig. 6 1
1 Sensor element
2 Multipole ring
SAE0979Y
SAE0980Y
with alternating
4
north and south
magnetization
Sensor types | Wheel-speed sensors 255
This type of data transmission using digital Compact dimensions combine with low
signals is less sensitive to interference weight to make the active wheel-speed
than the signals from passive inductive- sensor suitable for installation on and even
type sensors. The sensor is connected to within the vehicle’s wheel-bearing assem-
the control unit by a two-conductor wire. blies (Fig. 10). Various standard sensor
head shapes are suitable for this.
8 Block diagram of Hall IC Digital signal conditioning makes it possi-

ble to transfer coded additional informa-
"Vcc"
tion using a pulse-width-modulated output
Power Supply signal (Fig. 11):
Regulator
• Direction of wheel rotation recognition:
Main
Comp This is especially significant for the
Oscillator “hill hold control” feature, which relies
(Clock) Signal
on selective braking to prevent the vehi-
Right
Speed cle from rolling backwards during hill
PGA
ADC starts. The direction of rotation recogni-
Gain Range
Digital
Offset Circuit tion is also used in vehicle navigation
Center
DAC systems.
• Standstill recognition:
Direction
ADC This information can also be evaluated
SAE0981Y
Left
= (Left + Right)/2 - Center
by the “hill hold control” function.
The information is also used for self-
diagnosis.
• Signal quality of the sensor:
9 Signal conversion in Hall IC Information about the signal quality of
the sensor can be relayed in the signal.
a b If a fault occurs the driver can be ad- Fig. 9
vised that service is required. a Raw signal
Os1 b Output signal
Voltage
Us1
OS1 Upper switching
Time t Time t
SAE0982E
threshold
11 Coded information transfer with
pulse-width-modulated signals US1 Lower switching
threshold
a 90μs
10 Wheel bearing with wheel-speed sensor Fig. 10

1 Wheel-speed
1 b 180μs
sensor
c 1440μs Fig. 11
a Speed signal when
reversing
0,74 s b Speed signal when
driving forwards
c Signal when vehicle
d 45μs
is stationary
SAE0983Y
SAE0984Y
d Signal quality
of sensor, self-
diagnosis
256 Sensor types | Micromechanical yaw-rate sensors
Micromechanical A further area of application is rollover

yaw-rate sensors detection in airbag electronic control units
for triggering restraints (e.g. sidebags/win-
Applications dow airbags, rollover bar) in the event of
In vehicles with electronic stability pro- the vehicle rolling over. MM2 yaw-rate
gram (ESP), the rotation of the vehicle sensors have been specially developed
about its vertical axis is registered by mi- for this area. These are particularly suited
cromechanical yaw-rate (or yaw-speed) for the detection of rotary motion around
sensors (also known as gyroscopes) and the longitudinal axis because of their opti-
applied for vehicle-dynamics control. mized sensing direction. They permit a
This takes place during normal cornering, very compact design and space-saving in-
for instance, but also when the vehicle stallation in airbag control units, installed
breaks away or goes into a skid. These sen- along the longitudinal axis of the vehicle.
sors are reasonably priced as well as being Packaging the sensor element and the
very compact. They have forced out the evaluation electronics together in a stan-
conventional high-precision mechanical dard IC housing contributes to the cost
sensors. reductions in this area.
1 Structure of the MM1 yaw-rate sensor MM1 micromechanical yaw-rate sensor

200 m A mixed form of technology is applied
in order to achieve the high accuracies
needed for vehicle-dynamics systems.
That is, two somewhat thicker oscillating
3
elements (mass plates) which have been
machined from a wafer using bulk micro-
mechanics oscillate in counter-phase to
Fig. 1 their resonant frequency which is defined
1 Retaining/guide
by their mass and their coupling-spring
spring 1
stiffness (>2 kHz). On each of these oscil-
UAE0790-1Y
2 Part of the
oscillating element 2 lating elements, there is a miniature, sur-
3 Coriolis face-type micromechanical capacitive ac-
acceleration sensor celeration sensor. When the sensor chip
Fig. 2 2 MM1 micromechanical yaw-rate sensor
1 Frequency-determin- V
ing coupling spring
2 Permanent magnet
3 Direction of
oscillation
4 Oscillating element
3y
5 Coriolis acceleration
sensor
2 N
6 Direction
of Coriolis
S
acceleration
1
7 Retaining/guide B
spring
UAE0706-2Y
V Yaw rate
v Oscillation velocity
B Permanent-magnet
7 6 5 4 0 2 mm
field
Sensor types | Micromechanical yaw-rate sensors 257
rotates about its vertical axis at yaw rate V, below 40 dB). Here, the drive and mea-
these register the Coriolis acceleration in surement systems are rigorously decou-
the wafer plane vertical to the direction of pled from each other.
oscillation (Figs. 1 and 2). These accelera-
tions are proportional to the product of MM2 micromechanical yaw-rate sensor
yaw rate and the oscillation velocity which If the silicon yaw-rate sensor is manufac-
is maintained electronically at a constant tured completely in surface micromechan-
value. ics (SMM) and, at the same time, the mag-
To drive the sensor, all that is required is netic drive and control system is replaced
a simple, current-carrying conductor track by an electrostatic system, the decoupling
on each oscillating element. In the perma- of drive and measuring system does not
nent-magnet field B vertical to the chip need to be so thorough. Comb-like struc-
surface, this oscillating element is sub- tures (Figs. 3 and 4) electrostatically force
jected to an electrodynamic (Lorentz) a centrally mounted rotary oscillator to os-
force. Using a further, simple printed cillate. The amplitude of these oscillations
conductor (which saves on chip surface), is held constant by means of a similar ca-
the same magnetic field is used to directly pacitive pick-off. Coriolis forces result at
measure the oscillation velocity by induc-
tive means. The different physical con- 3 MM2 yaw-rate sensor: structure
struction of drive system and sensor sys- 50 m

tem serves to avoid undesirable coupling
between the two sections. In order to sup-
press unwanted external acceleration ef-
2
fects (common-mode signal), the opposing
sensor signals are subtracted from each
other. The external acceleration effects
can be measured by applying summation. 1
The high-precision micromechanical con-
struction helps to suppress the effects of
UAE0791-1Y
high oscillatory acceleration which is sev- Fig. 3

1 Comb-like structure
eral factors of 10 higher than the low-level
2 Rotary oscillator
Coriolis acceleration (cross sensitivity far
4 MM2 surface-micromechanical yaw-rate sensor
CDrv1
CDrv2 1
CDrv Det1
CDrv Det2
0,5 mm
Fig. 4
y
1 Comb-like structure
2 2 Rotary oscillator
CDet1 CDet2 3 Measuring axis
CDrv Drive electrodes
V -FC +FC CDet Capacitive pick-off
3
UKI0044-3Y
FC Coriolis force
v Oscillation velocity
V = DCDet, measured
yaw rate
258 Sensor types | Micromechanical yaw-rate sensors
the same time in an out-of-plane tilting navigation (Travel Pilot), adaptive cruise
movement, the amplitude of which is pro- control (ACC), roll over mitigation (ROM),
portional to the yaw rate V, and which is electronic active steering (EAS), active
detected capacitively by the electrodes un- suspension control (ASC), Steer-by-Wire.
derneath the oscillator. To avoid excessive DRS MM3.7k is the base variant of the
damping of this movement, the sensor MM3 generation for ESP applications.
must be operated in a vacuum. Although It comprises a yaw-rate sensor and an
the chip’s small size and the somewhat integrated lateral acceleration module.
simpler production process result in con-
siderable cost reductions, this miniaturiza- Operating principle
tion is at the expense of reductions in the The new micromechanical measuring ele-
measuring effect, which in any case is ment for measuring yaw rate belongs to
not very pronounced, and therefore of the familiar group of vibrating gyroscopes
the achievable precision. It also places operating according to the Coriolis prin-
more severe demands on the electronics. ciple (CVG = Coriolis Vibrating Gyros).
The system’s high flexural stability, and It consists of an inverse tuning fork with
mounting in the axis of gravity, serve to two linear vibration modes perpendicular
mechanically suppress the effects of un- to one another, the drive circuit and the
wanted external acceleration from the evaluator circuit. Drive and evaluation
side. take place electrostatically in a comb-like
structure. The Coriolis acceleration is
DRS MM3x sensor cluster measured electrostatically using elec-
A new generation of micromechanical ele- trodes that extend into each other. The
ments are used in the DRS MM3x sensor measuring element consists of two masses
cluster. These can measure and digitally connected by a coupling spring. The reso-
process angular velocities and accelera- nant frequency the same for the two vibra-
tions. They create a modular concept for tion modes. It is typically around 15 kHz
hardware and software based on circuit and is thus outside the usual interference
board technology with many new safety spectrum in the vehicle and is hence resis-
features which lead to a versatile and tant to interfering vibration. The evalua-
reliable solution for many applications tion circuit (ASIC) and the micromechani-
in the vehicle. cal measuring element are contained in a
prefabricated housing with 20 terminals
Application (Premold 20).
The ESP system, the connection to further
chassis comfort systems and the develop- The acceleration module is comparable
ment of advanced driving stability systems in form with the yaw-rate sensor module
called for inertia signals with high require- and consists of a micromechanical mea-
ments, in particular in respect of signal suring element, an electronic evaluating
quality and robustness, additional measur- circuit and a housing with 12 terminals
ing axes and high reliability. This is why (Premold12). The spring-mass structure is
Bosch has developed a third generation, deflected in its sensitive axis by external
the versatile yet inexpensive DRS MM3.x accelerations and is evaluated by means
sensor cluster, to meet the requirements of a differential capacitor in a comb-like
of functions such as hill hold control form.
(HHC), automated parking brake (APB),
Sensor types | Piezoelectric “tuning-fork” yaw-rate sensor 259
Piezoelectric “tuning-fork” Design

yaw-rate sensor The piezoelectric "tuning fork" yaw-rate
sensor comprises a steel element shaped
Application like a tuning fork. This incorporates four
In order that it can use the digital road piezo elements (two above, two below,
map stored on the CD-ROM to calculate Fig. 1) and the sensor electronics. The
the distance driven, the computer in the tuning forks are about 15 mm long.
vehicle’s navigation system needs informa-
tion on the vehicle’s movements (compos- Operating principle
ite navigation). When voltage is applied, the bottom
When the vehicle is cornering, the navi- piezo elements start to oscillate and excite
gation system’s yaw-rate sensor records the upper section of the “tuning fork”,
the vehicle’s rotation about its vertical together with its upper piezo elements,
axis, thus permitting determination of the which then starts counter-phase oscillation.
direction of travel. The vehicle’s position The frequency is approximately 2 kHz.
can hence be calculated in conjunction
with the determination of the distance Straight-ahead driving
traveled from the tachometer or wheel With the vehicle being driven in a straight
sensor signals by “composite positioning”. line there are no Coriolis forces acting on
Initially in navigation systems, the vehi- the tuning fork, and since the upper piezo
cle’s direction was determined from the elements always oscillate in counter-phase
earth’s magnetic field by a magnetic field and are only sensitive perpendicularly to
sensor. However, these could easily be in- the direction of oscillation, they do not
fluenced by the magnetic interference in generate a voltage.
the vehicle. As oscillation gyroscopes are
based on the measurement of mass forces, Cornering
they are insensitive to magnetic interfer- The rotational movement around the
ence. vehicle’s vertical axis as it travels around Fig. 1
a curve causes the upper portion of the A Sensing section
tuning fork to leave the oscillatory plane of the oscillating
so that an alternating voltage is generated element

B Stimulating section
in the upper piezo elements which is
of the oscillating
transferred to the navigation computer by element
1 Piezoelectric “tuning-fork” yaw-rate sensor an electronic circuit in the sensor housing. 1 Oscillating element
The voltage-signal amplitude is a function 2 Acceleration sensor
A V of both the yaw rate and the oscillation 3 Actuator

5 velocity. Its sign depends on the direction (piezoelectric
+ 6 7
element for
x UA (left or right) taken by the curve.
– vibration excitation)
1 4 Regulator for
2 constant vibration
excitation
8 5 Charge amplifier
B 3 6 Multiplication
(demodulation)
7 Low pass
4 8 Vibration excitation
UAE1068Y
UA Output voltage
(proportional to
the yaw rate)
V Yaw rate
260 Sensor types | Micromechanical pressure sensors
Micromechanical on the external pressure acting on it

pressure sensors (the center of the diaphragm is deflected
by 10 to 1,000 µm). The four strain-gage
Application resistors on the diaphragm change their
Pressure is a non-directional force acting electrical resistance as a function of the
in all directions which occurs in gases and mechanical stress resulting from the ap-
liquids. Micromechanical pressure sensors plied pressure (piezoresistive effect).
detect the pressure of various media in the The four measuring shunts are arranged
motor vehicle, e.g.: on the silicon chip so that when the dia-
• Intake-manifold pressure, e.g. for load phragm is deformed, the resistance of
sensing in engine-management systems two of them increases and that of the other
• Boost pressure for boost-pressure con- two decreases. These measuring shunts
trol are arranged in a Wheatstone bridge cir-
• Ambient pressure for taking into ac- cuit (Fig. 1b) and a change in their resis-
count air density, e.g. in boost-pressure tance values leads to a change in the ratio
control
• Oil pressure to take engine load into 1 Pressure-sensor measuring element with reference
vacuum on the components side
account in the service display
Fig. 1 • Fuel pressure for monitoring the level
a
a Sectional drawing of fuel-filter contamination
b Bridge circuit 1 3
R2
R1 R1
Micromechanical pressure sensors deter-
1 Diaphragm
mine the absolute pressure of liquids and
2 Silicon chip 2 p
3 Reference vacuum gases by measuring the pressure differen-
4 Glass (Pyrex) tial in relation to a reference vacuum. 4
p Measured pressure
U0 Supply voltage Version with the reference vacuum on R2 R1
UM Measurement b
the component side
voltage υM υ0
UAE0017-2Y
Design
R1 Strain-gage resistor R1 R2
(compressed) The measuring cell of a micromechanical
R2 Strain-gage resistor pressure sensor consists of a silicon chip
(elongated) (Fig. 1a, Item 2), in which a thin diaphragm
is micromechanically etched (1). Four
strain-gage resistors (R1, R2), are diffused 2 Measuring element of pressure sensor with cap and
reference vacuum on the components side (design)
into the diaphragm. The electrical resis-
tance in these resistors changes in re-
sponse to mechanical elongation. The
Fig. 2
measuring cell is surrounded and sealed
1, 3 Electrical terminals 1
on its component side by a cap (Fig. 2, 4
with glass-enclosed
lead-in Item 6), which encloses the reference
5
2 Reference vacuum vacuum underneath. 2
4 Measuring A temperature sensor (Fig. 3, Item 1), 6
element (chip) the signals of which can be evaluated inde-
with evaluation pendently, can also be integrated in the
electronics
pressure-sensor housing.
5 Glass base 3 7
UAE0648-2Y
6 Cap
7 Supply for Operating principle
measured The diaphragm of the sensor cell is de- p
pressure p flected to varying degrees depending
Sensor types | Micromechanical pressure sensors 261
of the voltages across them. This leads to The signal-conditioning electronic cir-
a change in the measurement voltage UM. cuitry is integrated on the chip. Its func-
This as yet unamplified voltage is there- tion is to amplify the bridge voltage, com-
fore a measure of the pressure applied pensate for temperature influences, and
to the diaphragm. linearize the pressure curve. The output
The measurement voltage is higher voltage is in the range of 0 to 5 V and is
with a bridge circuit than would be the supplied via electrical connections to
case if an individual resistor were used. the engine control unit (Fig. 3, Item 5).
The Wheatstone bridge circuit thus per- The control unit uses this output voltage
mits a higher sensor sensitivity. to calculate the pressure.
Version with reference vacuum

3 Micromechanical pressure sensor with reference in a special chamber
vacuum on the components side (design)
Design
1 2 3 4 5 A pressure sensor with the reference vac-
uum in a special chamber (Fig. 4) for use
as an intake-manifold or boost-pressure
sensor is easier to install than a sensor Fig. 3
with the reference vacuum on the compo-
(NTC)
nent side: A silicon chip (Fig. 5, Item 6) 2 Housing base
with etched diaphragm and four strain- 3 Manifold wall
gage resistors in a bridge circuit is located 4 Sealing rings
6 – like the pressure sensor with cap and ref- 5 Electrical
erence vacuum on the component side – connection (plug)
7
6 Housing cover
as a measuring cell on a glass base (3). In
7 Measuring cell
contrast to the sensor with the reference
vacuum on the component side, there is
UAE0722Y
no passage in the glass base through which Fig. 4

1 cm the measured pressure can be applied to 1 Manifold wall
the sensor element. Instead, pressure is 2 Housing
3 Sealing ring
4 Micromechanical pressure sensor with reference 5 Measuring element of pressure sensor with
vacuum in a special chamber (design) reference vacuum in a special chamber (design) (NTC)
5 Electrical
connection (plug)
6 Housing cover
7 Measuring cell
1 5
p
Fig. 5
2 1 Protective gel
3 1 2 Gel frame
5
4 2 3 Glass base
6 4 Ceramic hybrid
6
3 5 Chamber with
7
7 reference vacuum
4 6 Measuring cell
(chip) with
UMK1645-1Y
UMK1644-1Y
evaluation
electronics
1 cm 7 Bonded connection
p Measured pressure
262 Sensor types | Micromechanical pressure sensors
applied to the silicon chip from the side on High-pressure sensors

which the evaluation electronics is situated.
This means that a special gel must be used Application
on this side of the sensor to protect it against High-pressure sensors are used in a motor
environmental influences. The reference vehicle to measure fuel pressure and
vacuum (5) is enclosed in the cavity (special brake-fluid pressure, e.g:
chamber) between the silicon chip and the • Rail-pressure sensor for gasoline direct
glass base. The complete measuring element injection (pressure up to 200 bar)
is mounted on a ceramic hybrid (4), which • Rail-pressure sensor for common-rail
incorporates the soldering surfaces for diesel-injection system (pressure up to

electrical contacting inside the sensor. 2,000 bar)
• Brake-fluid pressure sensor in the
A temperature sensor can also be incor- hydraulic modulator of the electronic

porated in the pressure-sensor housing. stability program (pressure up to
It protrudes into the air current, and can 350 bar)
therefore respond to temperature changes
with a minimum of delay (Fig. 4, Item 4). Design and operating principle
High-pressure sensors operate according
Operating principle to the same principle as micromechanical
The operating principle, and with it the signal pressure sensors. The core of the sensor is
conditioning and signal amplification to- a steel diaphragm, onto which strain-gage
gether with the characteristic curve, corre- resistors have been vapor-deposited in the
spond to that used in the pressure sensor with form of a bridge circuit (Fig. 6, Item 3).
cap and reference vacuum on the component The sensor’s measuring range is depen-
side. The only difference is that the measuring dent on the thickness of the diaphragm
element’s diaphragm is deformed in the op- (thicker diaphragms for higher pressures,
posite direction and therefore the strain-gage thinner diaphragms for lower pressures).
resistors are “bent” in the other direction. As soon as the pressure to be measured
is applied to one side of the diaphragm
6 High-pressure sensor via the pressure connection (4), the strain-
gage resistors change their resistance
2 cm values as a result of the deflection of the
diaphragm. The output voltage generated
1 by the bridge circuit is proportional to the
applied pressure. This voltage is transmit-
ted via connecting lines (bonding wires)
2 to an evaluation circuit (2) in the sensor.
The evaluation circuit amplifies the bridge
3 signal to 0 to 5 V and transmits it to the
control unit, which calculates the pressure
Fig. 6 with the aid of a characteristic curve.
1 Electrical
4
connection (plug)
2 Evaluation circuit
3 Steel diaphragm 5
with strain-gage
UMK1576-1Y
resistors
4 Pressure
connection p
5 Mounting thread
Sensor types | Temperature sensors 263
Temperature sensors Exhaust-gas temperature sensor

This sensor is mounted on the exhaust-gas
Application system at points which are particularly
Engine-temperature sensor critical regarding temperature. It is ap-
This is installed in the coolant circuit plied in the closed-loop control of the
(Fig. 1). The engine management uses systems used for exhaust-gas treatment.
its signal when calculating the engine A platinum measuring shunt is usually
temperature (measuring range used (measuring range –40 to +1,000 °C).
– 40 to +130 °C).
Air-temperature sensor Depending upon the particular application,
This sensor is installed in the air-intake a wide variety of temperature sensor types
tract. Together with the signal from the are available. A temperature-dependent
boost-pressure sensor, its signal is applied semiconductor measuring shunt is fitted
in calculating the intake-air mass. Apart inside a housing. This resistor is usually
from this, setpoint values for the various of the NTC (Negative Temperature Coeffi-
control loops (e.g. EGR, boost-pressure cient, Fig. 2) type. Less often a PTC (Posi-
control) can be adapted to the air tempera- tive Temperature Coefficient) type is used.
ture (measuring range –40 to +120 °C). With NTC, there is a sharp drop in resis-
tance when the temperature rises, and
Engine-oil temperature sensor with PTC there is a sharp increase.
The signal from this sensor is used in cal-
culating the service interval (measuring The measuring shunt is part of a voltage-
range –40 to +170 °C). divider circuit to which 5 V is applied.
The voltage measured across the measur-
Fuel-temperature sensor ing shunt is therefore temperature-depen-
Is incorporated in the low-pressure stage dent. It is supplied through an analog-digi-
of the diesel fuel circuit. The fuel tempera- tal (A/D) converter and is a measure of the
ture is used in calculating precisely the temperature at the sensor. A characteristic
injected fuel quantity (measuring range curve is stored in the control unit which
–40 to +120 °C). allocates a specific temperature to every
resistance or output voltage value.
1 Coolant-temperature sensor 2 NTC temperature sensor: characteristic curve
1 2 3 4 5 6 :

2ESISTANCE
Fig. 1

1 Electrical
connection
2 Housing
UMK0124-7Y
UMK1998E
3 Sealing ring
# 4 Thread
4EMPERATURE 5 Measuring shunt
1 cm 6 Coolant
264 Sensor types | Accelerator-pedal sensors
Accelerator-pedal sensors
In Motronic systems with electronic throt-
tle control (ETC), the accelerator-pedal 1 Characteristic curve of an accelerator-pedal sensor
sensor records the travel or the angular

V
position of the accelerator pedal. For this
purpose, potentiometers are used in addi- 4.75
tion to proximity-type sensors. The accel-

erator-pedal sensor is incorporated with
Output voltage
the accelerator pedal in the accelerator- 1
pedal module. These ready-to-install units
make adjustments on the vehicle a thing of 2
Fig. 1
the past.
1 Potentiometer
UAE0724E
(reference
Potentiometric accelerator-pedal sensor 0.75
potentiometer)
2 Potentiometer The engine control unit receives the mea-
Pedal travel approx. 25 mm
(half voltage) sured value picked off at the potentiome-
ter wiper as a voltage. The control unit
uses a stored sensor curve to convert this 2 ARS1 Hall angular-position sensor
voltage into the relative pedal travel or the

angular position of the accelerator pedal
(Fig. 1). 4
A second (redundant) sensor is incorpo- 5
Fig. 2 ϕ
1 Rotor disk rated for diagnosis purposes and for use 1
(permanent- in case of malfunctions. It is a component
magnetic) part of the monitoring system. One sensor 2 3
2 Pole shoe
version operates with a second potentio-
3 Conductive element
meter, which always delivers half the volt-
UAE0770-2Y
4 6
4 Air gap
5 Hall sensor
age of the first potentiometer at all operat-
6 Shaft (soft ing points. Thus, two independent signals
magnetic) are available for fault-detection purposes
(Fig. 1). Instead of the second potentiome-
ter, another version uses a low-idle switch, 3 Principle of ARS2 Hall angular-position sensor
which signals the idle position to the con-

trol unit. The status of this switch and the a b
potentiometer voltage must be plausible. 90° mT
y
Flux density B
For vehicles with automatic transmis- 60

ϕ
30
sions, a further switch can be incorporated 180° 1 0°
Fig. 3 0
for an electrical kickdown signal. Alterna- x 90° 180° 270°
a Principle of NS
tively, this information can also be derived -30
operation 3 -60
b Characteristic curve from the rate of change of the potentiome- 270° Angle of rotation ϕ
ter voltage. A further possibility is to trig- 90° mT
Flux density B
1 Hall IC positioned ger the kickdown function by means of a y 60

in the midpoint of
defined voltage value of the sensor curve; 30 2
the circular path 180° 2 0°
here, the driver receives feedback on a jump x 0
2 Hall IC located 90° 180° 270°
UAE0863-1E
outside the
in force in a mechanical kickdown cell. NS -30
3 -60
midpoint This is the most frequently used solution. 270° Angle of rotation ϕ
(linearization)
3 Magnet
Sensor types | Accelerator-pedal sensors 265
Hall angular-position sensors magnetic field. The sensor is therefore

Hall sensors are used to measure the move- insensitive to component tolerances and
ment of the accelerator pedal on a non-con- is temperature-resistant.
tact basis. In the case of the Type ARS1 Hall As with the accelerator-pedal module
angular-position sensor, the magnetic flux of with a potentiometric sensor, these prox-
a roughly semicircular, permanent-magnetic imity-type systems also contain two sen-
disk is fed back via a pole shoe, two further sors in order to receive two redundant Fig. 4
conductive elements and the similarly voltage signals. a Design
ferromagnetic shaft to the magnet (Fig. 2). b Principle
Depending upon the angular setting, the flux 4 Measuring principle of FPM2.3 c Measurement
is led to a greater or lesser degree through signals
a c
the two conductive elements, in the mag- Hall element
output signals 1 Integrated circuit
netic path of which a Hall sensor is also situ- 400 (IC) with Hall
ated. Using this principle, it is possible to S 200 elements
N
achieve a practically linear characteristic 2
0 2 Magnet (opposing
curve in the measuring range of 90°. 200 magnet not shown
1
The Type ARS2 is a simplified version 400 here)
which does without soft magnetic conduc- 0° 180° 360° 540° 3 Conductive element
Angle of rotation 4 Hall elements
tive elements (Fig. 3). In this version, the b sin α (for recording
magnet moves around the Hall sensor in a 2 N S cos α x-components of B)
circular arc. Only a relatively small section 3 5 Hall elements
of the resulting sinusoidal characteristic Bx 4 5 4 (for recording
curve features good linearity. If the Hall y-components of B)
sensor is located slightly outside the cen- 5
Bx 4 Bx Homogenous
ter of the circular arc, the characteristic
4 magnetic field
curve increasingly deviates from the
(x-components)
UAE0997Y
sinusoidal, and now features a short mea- 5 By Homogenous

suring range of almost 90°, and a longer By magnetic field
measurement range of more than 180° (y-components)
with good linearity. A great disadvantage
though is the low level of shielding against 5 Exploded view of FPM2.3 accelerator-pedal module
external fields, as well as the remaining

dependence on the geometric tolerances 4
of the magnetic circuit, and the intensity Fig. 5
fluctuations of the magnetic flux in the 1 Pedal
3
permanent magnet as a function of tem- 2 Cover
perature and age. 3 Spacer sleeve
2 4 Sensor block with
In the case of the FPM2.3 type Hall 5 housing and plug
angular-position sensor, it is not the field
5 Bearing block
strength but rather the direction of the 6 Shaft with two
magnetic field which is used to generate magnets and
the output signal. The field lines are re- 6 hysteresis elements
corded by four radially arranged measur- 1
(round magnets not
7 8 visible)
ing elements lying in one plane in the
7 Kickdown
x- and y-directions (Fig. 4). The output sig-
(optional)
nals are derived in the ASIC from the raw 8 Two springs
9
UAE0998Y
data (cos and sin signals) using the arctan 9 Stop damper
10 11
function. The sensor is positioned between 10 Thrust member
two magnets to generate a homogenous 11 Floor cover
266 Sensor types | Steering-angle sensors
Steering-angle sensors wheels differ by one tooth which means

that a definite pair of angular variables is
Application associated with every possible steering-
The electronic stability program (ESP) wheel position.
applies the brakes selectively to the indi- By applying a mathematical algorithm
vidual wheels in order to keep the vehicle (a computing process which follows a
on the desired track selected by the driver. defined step-by-step procedure) referred
Here, a control unit compares the steering to here as a modified vernier principle, it
angle and the applied brake pressure with is possible to use the above AMR method
the vehicle’s actual rotary motion (around for calculating the steering angle in a mi-
its vertical axis) and its road speed. If nec-
essary, the brakes are applied at individual
wheels. These measures serve to keep the 1 LWS3 AMR steering-angle sensor (principle)
float angle (deviation between the vehicle

axis and the actual vehicle movement)
down to a minimum and, until the physical
limits are reached, prevent the vehicle 1
breaking away.
Basically speaking, practically all types
of angular-position sensors are suitable
Fig. 1 for registering the steering angle. Safety
1 Steering shaft considerations, though, dictate that only 2 5
2 AMR sensor those types are used which can be easily 6
elements checked for plausibility, or which in the 3
3 Gear wheel with
ideal case automatically check themselves. 4 7
m teeth
Potentiometer principles are used, as well
4 Evaluation
electronics as optical code-registration and magnetic
5 Magnets principles. Whereas a passenger-car steer-
UFL0030-1Y
6 Gear wheel with ing wheel turns through ±720° (a total of
n > m teeth 4 complete turns), conventional angular-
7 Gear wheel with
position sensors can only measure maxi-
m + 1 teeth
mum 360°. This means that with the major-
ity of the sensors actually used for this
purpose it is necessary to continually reg- 2 LWS3 AMR steering-angle sensor (view)
ister and store the data on the steering

wheel’s actual setting.

Steering-angle sensor with AMR element
The LWS3 steering-angle sensor uses AMR
(Anisotropic Magnetoresistive Sensors) for
its operation. The AMR’s electrical resis-
tance changes according to the direction
of an external magnetic field. In the LWS3,
the information on angle across a range
UFL0031-1Y
of four complete rotations is provided by

measuring the angles of two gear wheels
which are rotated by a third gear wheel
on the steering shaft. The first two gear
Sensor types | Steering-angle sensors 267
croprocessor. Here, even the measuring These short connections enhance the
inaccuracy of the two AMR sensors can be sensor’s robustness against external
compensated for. In addition, a self-check interference.
can also be implemented so that a highly The two GMR elements measure the
plausible measured value can be sent to direction of the field lines of the two mag-
the control unit. nets. The steering angle is calculated from
Figure 1 shows the schematic represen- this in the microprocessor. The sensor
tation of the LWS3 steering-angle sensor. element and microprocessor communicate
The two gear wheels, with magnets inserted, via a digital interface (SPI interface).
can be seen. The sensors are located above The calculated steering angle value is
them together with the evaluation elec- placed on the CAN by the microprocessor.
tronics. Because of the greater sensitivity, by
comparison with the AMR effect, the LWS5
A different design is the LWS4 steering- can work with weaker magnets and greater
angle sensor which unambiguously mea- air gaps. This brings considerable cost
sures a steering angle of 180°. It is mounted benefits in materials and design. The 360°
on the shaft end of the steering axle (Fig. 3) angular measuring range of a single GMR
element (a 180° measuring range is typical
Steering-angle sensor with GMR element for AMR), allows the use of smaller gear
The LWS5 is the first steering-angle sensor wheels on the LWS5. It thus requires a sig-
based on the GMR effect (Giant Magneto nificantly smaller installation space than
Resistive). Its mechanical structure and the LWS3. In addition, it has a significant
operating principle have been taken from level of scaleability, reflected in the mea-
the LWS3. The LWS3 and LWS5 are me- suring range (±90° to ±780°) and in the de-
chanically and electrically compatible. gree of redundancy. This means that the
The GMR layers are processed on the sensor can meet the specific requirements
planarized surface of the evaluation circuit of the various vehicle manufacturers as
(vertical integration). The GMR resistor closely as possible.
bridge and the evaluation module are con-
nected by means of plated-through holes.
LWS4 AMR steering-angle sensor for mounting on

3 the end of the steering axle
1 2 3 4
Fig. 3
1 Steering column
UFL0032-1Y
2 Steering gear
3 Steering-angle
sensor
4 Steering rack
268 Sensor types | Position sensors for transmission control
Position sensors for be designed to be sufficiently robust as to

transmission control be able to withstand the severe mechanical
stresses over the service life of the vehicle.
Application Surface-mounted sensors, on the other
The position sensor detects the positions hand, are made in the standard packages
of an actuator (e.g. selector lever shaft, familiar from engine compartment appli-
selector slide or park lock cylinder) in cations. They must meet the correspond-
AT, ASG, DKG and CVT transmissions. ing requirements such as being splash-
The sensors may be fitted stand-alone proof and handling operating tempera-
or built into electronics modules. tures from -40 to 130 °C.
Because of the complex requirements
Requirements (Table 1) arising from the various trans-
Sensors built into the transmission must mission topologies, and the requirements
be designed to cope with the challenging on mounting space and function (safety
operating conditions prevailing there: concept, required precision, etc.) different
• Ambient temperatures between -40 and physical measuring principles (Hall, AMR,
+150 °C, GMR, eddy-current principles) and shapes
• Transmission oil, which is an aggressive (linear and rotary detection) may be used.
ambient medium, These are explained in greater detail be-
• High mechanical stress up to 30 g low using the example of the Hall effect
• Abraded metallic materials and a build- and the eddy current principle.
up of particles in the transmission.
Linear position determination on the
Resistance to the media and temperature basis of Hall switches
in these applications is provided by an oil- Design
resistant enclosure for the electronics and Four digital Hall switches are mounted on
the use of high temperature PCBs and AVT a printed-circuit board in such a way that
for the electronics. The electronic compo- they detect the magnetic orientation of a
nents of ASIC, capacitors, resistors and linearly displaceable, multipolar perma-
interfaces (bonds, soldering points) must nent magnet (Fig. 1). The magnet carriage
is coupled with the linearly actuated selec-
tor slide (hydraulic slide in the transmis-
1 Requirements on position detection sion control plate) or the parking lock
Installation location Inside the transmission cylinder.
On the transmission In addition to the Hall switch, the
Design Linear printed-circuit board carries resistors
Rotational to represent diagnosis functions and
Measuring principle Hall EMC capacitors.
AMR
The sensor electronics are protected
GMR
Eddy current
against the influence of the transmission
Signal acquisition Digital (Gray code)
oil by being potted in a dense, oil-resistant
Analog epoxy resin.
Output signal Digital (2-/4-bit)
PWM
Analog
System security Redundancy
High availability
P/N recognition
Table 1 Shifting type M and E shifting

Sensor types | Position sensors for transmission control 269
Operating principle In some applications, the sensor transmits

In the case of automatic transmission with the controlled variable for starter release,
manual shifting, also known as M-shifting, the P/N signal, directly via a processor-
the position sensor detects the positions of independent hardware route. This in-
the selector slide, P, R, N, D, 4, 3, 2 and the creases the availability in operating states
intermediate ranges and passes these to (e.g. when the vehicle’s electrical system
the transmission control in the form of a voltage is low) in which the electronic con-
4-bit code (Fig. 2). For safety reasons, the trol unit would not yet be operational.
position is encoded in one-step encoding, In the case of automatic transmission
i.e. two-bit changes are always required with electronic shifting, also known as
before a new position is detected. Single- E-shifting, the position sensor detects
bit changes caused by malfunctions can only the positions of the park lock cylinder
be detected as false by the control unit by Pin and Pout and the intermediate range and
plausibility considerations. The sequence passes these to the transmission control
of these bit changes corresponds to a unit in the form of a 2-bit code.
Gray code.
Rotational position determination
1 Position sensor with Hall switch on the basis of eddy currents
Design
a
A position sensor with rotational design
is shown in Figure 3. In this example, a
specially shaped rotor with a return coil
is mounted to rotate on the selector lever
shaft outside the transmission. It repre-
sents the interface to the transmission
actuator. Redundant transmit and receive
coils are mounted on the fixed sensor
board with their associated evaluating
ASICs. In this solution, the actuator (shaft) Fig. 1
passes through the circuit board. a Front view
1 2 b Rear view
1 Oil-resistant
package
b 2 Coding on the magnetic carriage
2 Electrical
connection by
a Magnetic North Pole terminal plate
3 Magnetic South Pole 3 Printed-circuit
1 board potted Hall
elements
4 Carriage with
2
permanent magnet
4 bit code 5 Locating pin
4 b
Fig. 2
P R N D 4 3 2 a Magnetic encoding
b Position ranges
Switching and intermediate ranges
SAE1084E
UTS0363Y
Transition range 1 Moving carriage

5 2 Fixed position of
the Hall elements
270 Sensor types | Position sensors for transmission control
Operating principle For safety reasons, the sensor makes

The sensor uses the eddy-current princi- two independent, opposing output signals
ple. The transmitting coil induces in the available, with the upper and lower 5 %
return coil an eddy current opposite to its of the signal stroke being used as a diag-
cause, the magnetic field of which induces nostic range. In the event of a fault being
a voltage in the receiver coils. The geome- detected inside the sensor, a voltage level
tries of the return coil and the receiver in the upper diagnostic range is output.
coil are tuned to one another, so that con- Faults that can occur on the transmission
tinuously changing positions of the rotor path to the control unit (e.g. open circuit
can be detected. or short-circuit) also lead to voltage levels
Similarly to the example of linear posi- in the diagnostics range. In addition to
tion detection, the rotational position sen- this, the electronic control unit tests the
sor detects the P, R, N, D, 4, 3 and 2 posi- signal for plausibility by adding the two
tions. In this case, however, on the selector signals together. Sensor or transmission
lever shaft projection out of the transmis- errors can be detected by the control unit
sion. The positions of the shaft correspond in this way, and an appropriate emergency
to the positions of the selector slide inside running program can be selected, depend-
the transmission. ing on the fault.
One benefit of analog signal condition- Table 2 compares the two examples of
ing is that the individual shift and interme- position sensor and illustrates the com-
diate ranges are assigned to the angle posi- plexity of the topic.
tions detected by the sensor in the software, The complex requirements on the actua-
and so can easily be adapted to transmis- tor’s mechanical interface, the mounting
sion design variants. space, the environmental influences and
the safety concept frequently lead to appli-
cation-specific sensor system solutions.
3 Position sensor using eddy-current principles
Fig. 3
a Components
1
(without housing)
b Printed-circuit 2 Typical characteristics
board Position sensor Hall switch Eddy current
c Rotor
Measuring principle Hall effect Eddy current
b c (digital) (analog)
1 Selector lever shaft
2
2 Sensor PCB Sensitivity to exter- Yes No
3 Redundant nal magnetic fields
transmitter and Sensitivity to metal- No Yes
3
receiver coil lic surroundings
4 Redundant 5
Flexible position No Yes (software)
electronics
scaling
5 Rotor with
4 Operating voltage 4 to 12 V 4.5 to 5.5 V
return coil
SAE1085Y
range
Sensitivity to the High Low
Table 2 air gap
Sensor types | Axle sensors 271
Axle sensors a homogenous magnetic field. The mag-

netic field generates a Hall voltage in the
Applications Hall IC proportional to the magnetic-field
The automatic headlamp range control strength. When the ring magnet (6) is ro-
(ALWR) adjusts the vehicle’s headlamp tated by the shaft (2), the magnetic field
range automatically. With the headlamps through the Hall IC changes accordingly.
on low beam, ALWR compensates for the When the springs are compressed due
vehicle tilt so that the driver has adequate to loading and/or acceleration/braking,
vision without dazzling oncoming traffic. the connecting rod (Fig. 2, Item 4) trans-
Whereas, the static ALWR compensates fers the spring compression movement to
for the vehicle tilt resulting from the vehi- the pivot lever of the axle sensor so that it
cle loading, the dynamic version also takes is converted into a voltage signal which is
into account the front-end up and down proportional to the angle of rotation.
pitching movement of the vehicle caused The control unit registers the axle-sen-
by braking and acceleration. Here, the axle sor signals and from them generates the
sensors precisely register the bodywork’s difference between front and rear axle.
tilt angle. Using this voltage difference, and taking
into account the vehicle’s speed, the con-
Design and operating principle trol unit now calculates the setpoint value
Axle sensors (angular-position sensors) for the servomotor settings. With the vehi-
are used to measure the vehicle tilt. These cle driving at constant speed, the dynamic
are attached to the bodywork at the front ALWR remains in the status which features
and rear of the vehicle. Spring compres- a high level of damping whereby the step-
sion is transmitted to the sensors by a ping motors are adapted slowly to the
pivot lever which is connected through vehicle tilt. This prevents bumps, uneven-
a connecting rod to the particular axle ness, or holes in the road causing continu-
or to the wheel suspension. The vehicle’s ous correction of the headlamp settings.
tilt is then calculated from the difference When the vehicle accelerates, or the
between the voltages of the front and rear brakes are applied, the system automati-
axle sensors. cally switches to the dynamic mode, and
The axle sensors function according adjusts the headlamp range within a few
to the Hall effect. A Hall IC is built into the milliseconds. The system then switches
stator (Fig. 1, Item 5) where it is located in back to the slower mode.
1 Axle sensor (section) 2 Axle sensor (installed in vehicle)

Fig. 1
1 1 Pivot lever
2 Shaft
2 3 Housing
4 Ring-magnet mount
1 2 3 4
5 Stator with Hall IC
6 Ring magnet
3
Fig. 2
N
1 Bodywork mounting
5 4
UAE0792Y
2 Axle sensor with

6 plug-in connection
UAE0668Y
S
3 Pivot lever
5
4 Connecting rod
5 Vehicle axle
272 Sensor types | Hot-film air-mass meters
Hot-film air-mass meters contains a flow rectifier, which ensures that

the flow in the measuring tube is uniform.
Application The flow rectifier is either a combination
To provide precise pilot control of the of a plastic mesh with straightening action
air-fuel ratio, it is essential for the supplied and a wire mesh, or is a wire mesh on its own
air mass to be exactly determined in the (Fig. 3, Item 8).The measuring tube is in-
respective operating status. The hot-film stalled in the intake tract downstream from
air-mass meter measures some of the actu- the air filter. Plug-in versions are also avail-
ally inducted air-mass flow for this pur- able which are installed inside the air filter.
pose. It takes into account the pulsations
and reverse flows caused by the opening The most important component parts in the
and closing of the engine’s intake and ex- sensor are the measuring cell (Fig. 1, Item 4)
haust valves. Intake-air temperature or in the air intake (8) and the integrated evalu-
air-pressure changes have no effect upon ation electronics (3).
measuring accuracy. The sensor measuring cell consists of
a semiconductor substrate. The sensitive
HFM5 design surface is formed by a diaphragm which
The housing of the HFM5 hot-film air-mass has been manufactured in micromechani-
meter (Fig. 1, Item 5) projects into a mea- cal processes. This diaphragm incorpo-
suring tube (Fig. 3, Item 6) which, depend- rates temperature-sensitive resistors.
ing upon the engine’s air-mass require- The elements of the evaluation electronics
ments (370 to 970 kg/h) can have a variety (hybrid circuit) are installed on a ceramic
of diameters. The measuring tube normally substrate. This principle permits very
compact design. The evaluation electron-
ics are connected to the control unit by
1 HFM5 hot-film air-mass meter (schematic diagram) means of electrical connections (1).
The partial-flow measuring passage (6)
is shaped so that the air flows past the
1 measuring cell smoothly (without swirl
effects) and back into the measuring tube
via the air outlet (7). The length and loca-
tion of the inlet and outlet of the partial-
2 flow measuring passage have been chosen
Fig. 1 to provide good sensor performance even
1 Electric in the event of sharply pulsating flows.
connections (plug)
2 Measuring-tube or Operating principle
air-filter housing 3 The HFM5 hot-film air-mass meter is a thermal
wall
sensor which operates according to the fol-
3 Evaluation
electronics lowing principle: A centrally situated heat-
(hybrid circuit) ing resistor on the measuring cell (Fig. 3,
4 Sensor measuring 4 Item 3) heats a sensor diaphragm (5) and
cell maintains it at a constant temperature.
5 Sensor housing 5 8 The temperature drops sharply on each
6 Partial-flow QM side of this controlled heating zone (4).
measuring passage
UMK1713-1Y
7 Outlet, partial air 6 The temperature distribution on the dia-

flow QM phragm is registered by two temperature-
8 Intake, partial air 1 cm 7 dependent resistors which are mounted
flow QM upstream and downstream of the heating
Sensor types | Hot-film air-mass meters 273
resistor so as to be symmetrical to it Incorrect air-mass readings will be regis-

(measuring points M1, M2). Without the tered if the sensor diaphragm is contami-
flow of incoming air, the temperature nated with dust, dirty water or oil. For the
profile (1) is the same on each side of purpose of increasing the robustness of
the heating zone (T1 = T2). the HFM5, a protective device has been
As soon as air flows over the measuring
cell, the uniform temperature profile at
the diaphragm changes (2). On the inlet 2 Hot-film air-mass meter (characteristic curve)
side, the temperature characteristic is

V
steeper since the incoming air flowing
past this area cools it off. On the opposite
4
side, the temperature characteristic only
Output voltage
changes slightly, because the incoming air

3
flowing past has been heated by the heater
element. The change in temperature distri-
bution leads to a temperature difference 2
(∆T) between the measuring points M1
and M2. 1
The heat dissipated to the air, and there-
fore the temperature characteristic at the 0
UMK1691-1E
measuring cell is dependent on the air mass –100 0 200 400 600 kg/h
Air-mass flow
flowing past. The temperature difference is Reverse Forward
(irrespective of the absolute temperature of flow flow
the air flow past) a measure of the air-flow
mass. It is also direction-dependent so that
the air-mass meter can record both the 3 Hot-film air-mass meter (measuring principle)
amount and the direction of an air-mass

flow. T
Due to its very thin micromechanical 1
diaphragm, the sensor has a highly dy-
namic response (< 15 ms), a point which T1 = T2
2
is of particular importance when the in- T2 Fig. 3
coming air is pulsating heavily.
ΔT 1 Temperature profile
The evaluation electronics integrated in without air flow
T1
the sensor convert the resistance differen- 2 Temperature profile
0
tial at the measuring points M1 and M2 into with air flow
an analog voltage signal of between 0 and 3 Measuring cell
4 Heating zone
5 V. Using the sensor curve (Fig. 2) stored in
5 Sensor diaphragm
the engine control unit, the measured volt- M2
6 Measuring tube
age is converted into a value representing M1 7
with air-mass meter
the air-mass flow (kg/h). 7 Intake-air flow
The shape of the characteristic curve is 8 Wire mesh
such that the diagnosis facility incorporated M1, M2 Measuring
4 points
in the control unit can detect such malfunc-
T1, T2 Temperature
tions as an open-circuit line. An additional
5 values at
3
UMK1652-3Y
temperature sensor for evaluation functions measuring points

can be integrated in the HFM5 hot-film air- 8 M1 and M2
mass meter. It is not required for measuring 6 7 ∆T Temperature
the air mass. difference
274 Sensor types | Hot-film air-mass meters
developed which, in conjunction with a The intake-air temperature is measured

deflector mesh, keeps dirty water and dust by a temperature-dependent resistor,
away from the sensor element (HFM5-CI; which is integrated in the closed control
with C-shaped bypass and inner tube (I), loop for monitoring the heating-zone tem-
which together with the deflector mesh perature. The voltage drop at this resistor
protects the sensor). passes through an analog-digital converter
to give a digital signal representing the in-
HFM6 hot-film air-mass meter take-air temperature. The signals for air
The HFM6 uses the same sensor element mass and intake-air temperature are used
as the HFM5 and has the same basic to address a program map in which the
design. It differs in two crucial points: correction values for the air-mass signal
• The integrated evaluation electronics are stored.
operate digitally in order to obtain
greater measuring accuracy Improved protection against
• The design of the partial-flow measuring contamination
passage is altered to provide protection The partial-flow measuring passage is di-
against contamination directly upstream vided into two sections in order to provide
of the sensor element (similar to the de- better protection against contamination
flector mesh in the HFM5-CI) (Fig. 4). The passage which passes the sen-
sor element has a sharp edge (1), around
Digital electronics which air must flow. Heavy particulates
A voltage signal is generated with a bridge and dirty-water droplets are unable to fol-
circuit from the resistance values at the low this diversion and are separated from
measuring points M1 and M2 (Fig. 3); this the partial flow. These contaminants exit
voltage signal serves as the measure of the the sensor through a second passage (5).
air mass. The signal is converted into digi- In this way, significantly fewer dirt parti-
tal form for further processing. cles and droplets reach the sensor ele-
The HFM6 also takes into account the ment (3) with the result that contamina-
temperature of the intake air when deter- tion is reduced and the service life of the
mining the air mass. This increases signifi- air-mass meter is significantly prolonged
cantly the accuracy of the air-mass mea- even when operated with contaminated
surement. air.
4 HFM6 with improved contamination protection
3 2
Figure 4
1 Diverting edge 1
2 Partial-flow 5
measuring passage
(first duct)
3 Sensor element
SMK2055Y
4 Air outlet
5 Second passage
6 Particulate and 6 4
water outlet
Sensor types | Piezoelectric knock sensors 275
Piezoelectric knock sensors 2 Knock sensor (design and mounting)
Application
Regarding their principle of functioning, 1 2 3 4
knock sensors are basically vibration sen-
sors and are suitable for detecting struc-
ture-borne acoustic oscillations. These F
occur as “knock”, for instance, in gasoline
engines when uncontrolled combustion V
takes place. They are converted by the
Fig. 2
knock sensor into electrical signals (Fig. 1) 5 1 Piezoceramic
F
and transmitted to the Motronic control element
6
unit, which counteracts the engine knock 2 Seismic mass
by adjusting the ignition angle. with compressive
forces F
3 Housing
7 4 Bolt
UMK1634-1Y
Due to its inertia, a mass (Fig. 2, Item 2)
5 Contact washer
excited by a given oscillation or vibration 6 Electric connection
exerts a compressive force on a toroidal 7 Engine block
1 cm
piezoceramic element (1) at the same V Vibration
frequency as the excitation oscillation.
These forces effect a charge transfer Mounting
within the ceramic element. An electrical In four-cylinder engines, one knock sensor
voltage is generated between the top and is sufficient to record the knock signals for
bottom of the ceramic element which is all the cylinders. Engines with more cylin-
picked off via contact washers (5) and ders require two or more knock sensors.
processed in the Motronic control unit. The knock-sensor installation point on
the engine is selected so that knock can
be reliably detected from each cylinder.
The sensor is usually bolted to the side
1 Knock-sensor signals (oscilloscope display) of the engine block. It must be possible
for the generated signals (structure-borne-
noise vibrations) to be introduced without
Without resonance into the knock sensor from
knock
the measuring point on the engine block.
a A fixed bolted connection satisfying the
b following requirements is required for
this purpose:
c • The mounting bolts must be tightened to
a defined torque
With • The contact surface and the bore in the
knock
engine block must comply with prespec-
a ified quality requirements
• No washers of any type may be used
b
Fig. 1
a Cylinder-pressure
UMZ0121-1E
c characteristic curve
b Filtered pressure
signal
c Knock-sensor signal
276 Sensor types | SMM acceleration sensors
SMM acceleration sensors packaging stage (module). Figure 1 shows

a module in a SO16 housing. This module
Application on a small PCB, with further circuit com-
Surface micromechanical (SMM) accelera- ponents, is pressed into the second pack-
tion sensors have a variety of acceleration aging stage, a plastic housing. This housing
detection roles in the vehicle. These in- is closed and sealed with a welded plastic
clude the detection of acceleration values lid. This sensor is intended for installation
for a frontal or lateral impact in order to in the side or fender of the vehicle after
trigger the seat-belt pretensioners, the air- calibration and testing and sends its accel-
bag and the rollover bar in vehicle occu- eration signal via plug and cable to the
pant-protection systems. Surface micro- central airbag control unit.
mechanical (SMM) acceleration sensors The functional layers of the sensor ele-
are used in the field of active vehicle safety ment representing a spring-mass system
in ABS, ESP and HHC (Hill Hold Control, are applied to the surface of the silicon
low-speed traction control) systems. Fur- wafer by an additive method (Fig. 3,
ther possible applications of acceleration surface micromechanics).
sensors are found in chassis and suspen- The mass with its comb-like electrodes
sion control (active suspension) and vehi- (Figs. 2 and 3, Item 1) is connected via
cle alarm systems, where a change in the spring elements (2) to the anchor points
vehicle tilt will trigger an alarm. (Fig. 3, Item 4) in the core of the sensor.
There are fixed comb-like electrodes (3, 6)
Design and operating principle on the chip on each side of these movable
The surface micromechanical sensor ele- electrodes. A useful capacitance in the
ments are designed for different measur- range of 300 fF to 1 pF results from the
ing ranges, depending on the application. connection in parallel of the individual
These measuring ranges lie between 1 g capacitances formed by the fixed and
and 400 g (1 g ≈ 9.81 m/s2). The design of moving electrode fingers. The two rows
an acceleration sensor for side or frontal of electrode fingers connected in parallel
impact sensing is described below. The result in two useful capacitances (C1 – CM
SMM sensor element is fitted together with and C2 – CM), which change in opposite di-
the evaluation electronics (ASIC) in a first rections as the central mass is deflected.
1 Surface micromechanical acceleration sensors for airbag triggering (example)
Fig. 1 1
1 First packaging 1
stage (module)
4
in SO16 housing
5 2
2 Evaluator chip
(ASIC)
3 SMM sensor
element
6 3
SAE1079Y
4 Populated PCB
5 Second packaging
stage (device)
6 Plug
Sensor types | SMM acceleration sensors 277
The acceleration applied to the spring- Up to three acceleration sensors (for ESP
mass system deflects it by an amount lin and HHC) and one or two yaw-rate sen-
early related to the acceleration applied- sors (for ESP) are incorporated in the ESP
over the spring return force. An electrical sensor cluster. This clustering of sensor
output signal that is linearly dependent on modules reduces the number of individual
the acceleration is obtained by evaluation components and signal lines by compari-
of this differential capacitor. son with separate sensor devices. It also
The acceleration signal obtained in means that fewer mountings and less in-
the first stage of the evaluation circuit is stallation space are needed in the vehicle.
further conditioned in the ASIC, i.e. it is
amplified, filtered and prepared for the
output interface.
Analog voltages, pulse-width-modulated
signals, SPI protocols or current loop in
terfaces are usual as the output interface. -
Tolerances arising from the manufacture
of the sensor element, the evaluation cir
cuit and package influences on sensitivity -
and zero point are eliminated by means of 3 Comb-like structure of the sensor measuring cell
a comparison of programmable memory 100 m

cells at the end of the production line.
A self-test function tests the full mechani
cal and electrical signal path. In the course -
of this self-diagnosis, an electrostatic force 3
deflects the sensor structure, i.e. simulat 2
ing an acceleration in the vehicle, and the- 1
response of the measurement signal is
compared with a setpoint value. Fig. 3
4 1 Spring-mounted
seismic mass
UAE0800-1Y
with electrode
2 Spring
3 Fixed electrodes
4 Anchor areas
2 Surface micromechanical acceleration sensors with capacitive pick-off
1 2 3 Fig. 2
1 Spring-mounted
seismic mass
with electrodes
2 Spring
3 Fixed electrodes
a with capacity C1
4 Printed Al
conductor track
5 Bond pad
6 Fixed electrodes
with capacity C2
UAE0801-2Y
7 Silicon oxide
C1 C2 CM
a Acceleration in the
sensing direction
4 5 6 7 C2 CM C1
CM Measuring capacity
278 Sensor types | Micromechanical bulk silicon acceleration sensors
Micromechanical bulk silicon This configuration corresponds to a series

acceleration sensors circuit of two capacitors C1–M and C2–M.
Alternating voltages are applied across
Application terminals C1 and C2, and their superimpo-
Micromechanical bulk silicon acceleration sitions picked off between the capacitors
sensors detect the acceleration signals at CM.
needed for ABS, ESP and shock absorber In the idle condition, the capacitances
adjustment. C1–M and C2–M are ideally equal. This means
This type of sensor is today predomi- that the difference DC is equal to zero.
nantly used in the low acceleration ranges If an acceleration a is acting in the measur-
with high demands on signal to noise ratio ing direction, the Si center plate (the seis-
(< 2 gn). mic mass) is caused to deflect. This causes
a change in the distance to the upper and
Design and operating principle lower plates, and with it a capacitance
Anisotropic and selective etching tech- change in the capacitors C1–M and C2–M
niques are used to form the required and thus a difference DC not equal to zero.
spring-mass system for the sensor from This leads to a change in the electrical sig-
the solid wafer (bulk or volume microme- nal at CM which is amplified and filtered in
chanics). Capacitive pick-offs have proven the evaluation electronics.
especially effective for the high-precision The air space between the plates allows
measurement of this seismic-mass deflec- effective damping with low temperature
tion. They require a wafer-thick silicon or sensitivity.
glass plate (Fig. 1, Item 4) with counter-
electrodes, on each side of the spring-held
mass (2). The wafers and their counter-
electrodes here also provide overload
protection.
1 Bulk silicon acceleration sensor with capacitive

pick-off (schematic diagram)
1 2 3
a
Fig. 1
1 Upper Si plate 4
2 Center Si plate
5
(spring-held or
seismic mass)
3 Si-oxide
4 Bottom Si plate
UAE0646-2Y
5 Glass substrate
C1-M C2-M
a Acceleration in the C1 CM C2
sensing direction
CM Measuring capacity
Sensor types | Piezoelectric acceleration sensors 279
Piezoelectric acceleration 2 Piezoelectric acceleration sensor (two channel

sensors sensor for PCB mounting)
Application
Piezoelectric spring elements are used
as acceleration sensors in passenger-
restraint systems for triggering the seat-
belt pretensioners, the airbags, and the
rollover bar.

The core of this acceleration sensor is a
spring element (bending beam). A bonded
structure comprising two piezoelectric
layers of opposing polarity (bimorphous 1
spring element). An acceleration acting
on this element generates a mechanical
tensile stress in one layer and a compres-
sive stress in the second layer (Fig. 1).
The metallized surfaces on the top and
underneath of the spring element act
UAE0797-1Y
as electrodes from which the resultant
voltage is picked off. This structure is
packaged together with the electronics Fig. 2
1 Spring element
required in a hermetically sealed housing
(Fig. 2).
The electronic circuit consists of an im- can be measured (lower cutoff frequency
pedance converter and a tunable amplifier is typically 1 to 10 Hz).
with predefined filter characteristics. Piezo-spring elements do not require
The design means that no static signals any additional seismic mass. Their own
mass is adequate for an easily evaluated
signal.
1 Spring element in a piezoelectric acceleration
sensor
a 1
a =0
UA = 0
Fig. 1
a Not under
acceleration
b 1 a=0 b Under
UA > 0 acceleration a
1 Piezoceramic
UAE0293-2Y
bimorphous
spring element
UA Measurement
voltage
280 Sensor types | iBolt™ force sensor
iBolt™ force sensor nals and transmits the result of the classifi-
cation to the airbag control unit.
Applications
NHTSA FMVSS-208 (National Highway Design and operating principle
Traffic Safety Administration; Federal The iBolt™ sensor’s operating principle
Motor Vehicle Safety Standards and is based on the measurement of the deflec-
Regulations 208) came into force in the tion of a bending beam caused by the pas-
USA in 2004. These regulations were intro- senger’s weight. The amount of the deflec-
duced to prevent or reduce injuries to tion is detected by measurement of the
children traveling in the front passenger magnetic field strength in a special Hall
seat caused by the airbag triggering. Clas- sensor/magnet configuration (Fig. 2a).
sifying the front passenger by means of a The iBolt™ is designed such that it is the
weight measurement makes it possible to z component of the passenger’s weight that
switch the airbag off selectively if a small causes the deflection of the bending beam.
child is occupying the seat. The vehicle’s co-oordinates system here
The BOSCH iBolt™ (intelligent bolt) defines the x axis as the direction of travel,
sensor was developed to perform this the z and y axes being perpendicular and
weight classification reliably and robustly. horizontal to this. The configuration of the
This is done by fitting four iBolts™ in the magnet and the Hall IC in the sensor is se-
seat frame of the front-passenger seat lected such that the static magnetic field
(a sensor in each corner of the seat, Fig. 1). which penetrates the Hall IC results in an
A control unit also built into the seat analy- electrical signal linear to the deflection
ses the four analog, electrical weight sig- of the bending beam. The special design
of the iBolt™ sensor prevents a horizontal
deflection of the Hall IC in relation to the
1 Fitting four iBolts™ in the seat base frame of the magnet, to minimize the influence of trans-
front passenger’s seat
verse forces and moments. In addition,
the maximum stress in the bending beam
is limited by a mechanical overload
stop (Fig. 2b). This protects the iBolt™,
especially in the event of overloads in a
crash.
Calibration
The linear Hall sensor uses the spinning
current measuring principle which allows
calibration of the sensitivity, the offset and
UAE1077Y
the temperature sensitivity of measure-

ment. The calibration data is stored in an
EEPROM incorporated in the substrate
for the Hall sensor.
Linearity of the output signal

A linear output signal has been achieved
by means of a special design feature.
Fig. 1
The force generated by the weight of the
1 Seat base frame front passenger is passed from the upper
1 2
2 BOSCH iBolt™ seat structure through the sleeve into the
sensor bending beam (Fig. 2a). The force is then
Sensor types | iBolt™ force sensor 281
passed from the bending beam into the generated by the seat occupant’s position,
lower seat structure. The bending beam which also depends on the position of the
was designed as a double bending beam seat back. For this reason, the measuring
as this has an S-shaped deformation line. range of the iBolt™ sensor has been de-
The two vertical connection points of the signed such that the forces can be detected
double bending beam remain vertical over in both the positive and negative z direc-
the entire deflection range. This guaran- tions. This permits an unambiguous deter-
tees a linear and parallel movement of the mination of the front passenger’s weight.
Hall IC in relation to the magnet, resulting Thanks to its symmetrical measuring
in a linear output signal (Fig. 3). range, the iBolt™ sensor detects compres-
sion and tensile forces with the same sen-
Symmetrical measuring range sitivity and the same tolerances. This means
Tests of the system in vehicle seats have that the same sensors can be used in both
shown that both positive and negative vertical mounting orientations for all four
forces can act on the sensors. This has a points of connection between the upper
number of causes: on the one hand, nega- and lower seat structures.
tive forces can act on an individual tension
as a result of initial tensions as a conse-
quence of tolerances in the assembly of
the seat and of its installation in the vehi-
cle. On the other hand, negative forces
arise on individual sensors as a function
of the distribution of forces on the sensors
2 Measuring principle of the BOSCH iBolt™ sensor
a 1 FG < 850N 2 3
Fig. 2
a Ratios for weight
FG < 850 N (within
SN the measuring
range)
3 Typical output signal from the sensor
b Ratios for weight
FG > 850 N (outside
5.0 the measuring
V range)
FR 4 5 6 7
1 Oscillating crank
b FG > 850N 2 Air gap
Output voltage UA
3 Sleeve
Overload Overload 4 Seat rail
range range
2.5 5 Double bending
beam
SN 6 Magnet
7 Hall IC
UAE1076E
UAE1075Y
Fig. 3
0
–1,500 –1,000 –500 0 500 1,000 N Output signal as a
FR function of the force
Weight FG
applied
282 Sensor types | Torque sensor
Torque sensor Here the magnetoresitive sensor use sup-

plies two signals over the measuring range
Application which describe a sine and a cosine signal
Electromechanical power-assisted steering plotted against the angle of rotation of the
is increasingly being introduced in small, torsion rod. The angle of rotation and
compact and medium class vehicles. hence of the torque is calculated in
The essential advantages are simple instal- an electronic control unit using an arc-
lation and commissioning in the vehicle, tangent function.
energy saving and the suitability of these As there is always a fixed assignment of
systems in the vehicle’s network of control the two signals over the defined measuring
units for assistance systems enhancing range, a deviation from this allows faults in
comfort and safety. the sensor to be detected and the neces-
sary substitution measures to be intro-
Design and operating principle duced.
In order to sense the driver command in A coil spring with the necessary number
electromechanical power-assisted steer- of contacts is used to provide the electrical
ing, it is necessary to measure the torque contacts for the sensor over the rotational
applied by the driver. In the sensor cur- range of approximately ± 2 revolutions of
rently in series production use for this, a the steering wheel. The power is supplied
torsion rod is fitted in the steering shaft. and the measured values are transmitted
The rod is twisted by a defined amount in via this coil spring.
linear proportion to the torque applied
by the driver when the driver turns the
steering wheel (Fig. 1). This twisting can,
for its part, be measured by suitable means 1 Torque sensor
and converted into electrical signals.

The measuring range required for a torque
sensor to be used in electromechanical a 3
power-assisted steering is usually around
Fig. 1 ± 8 to ± 10 Nm. The maximum angle of ro-
a Sensor module
2
tation is mechanically limited by special
b Measuring principle 1
components to protect the torsion rod 5
c Electrical output
signal
against overload or destruction. 4
A magnetoresistive sensor is mounted 2 6
1 Torsion rod on one side of the torsion rod to allow revolutions
(twisting area measurement of the twisting and hence b 6 7 8
internal) the torque applied. This scans the field of
2 Input shaft (from N
4
a magnetic multipole wheel mounted on S S + Uv
the steering wheel) Signal 1 (sine)
the other side. The number of poles se- N N
3 Housing for the Signal 2 (cosine)
coil spring making lected for this wheel is such that the sen- S
N
S
Ground
the electrical sor issues an unambiguous signal within
9
connections its maximum measuring range and thus c
4 Sensor module with makes a clear statement of the torque V Nm
Signal voltage
magnetoresistive applied at any time.

Torque
4 4
sensor chip and 0
signal amplification 2 –4
UAE1063Y
5 Steering pinion/
0 –12
balancer shaft –8 –4 0 4 8
6 Magnetic multipole Angle of rotation
wheel
Sensor types | Rain/light sensor 283
Rain/light sensor Design and operating principle

A rain sensor comprises an optical trans-
Application of the rain sensor mission and reception path (Fig. 1). An
The rain sensor detects fine raindrops LED (6) emits light at a determined angle
on the windshield, so that the windshield into the windshield (2) which is reflected
wipers can be triggered automatically. on the outer boundary layer (glass/play)
The sensor controls the speed of the wiper (total reflection) and then is evaluated
system according to the amount of rain on an aligned receiver (photodiode, 4).
measured on the windshield (intermittent, If there is moisture on the sensitive outer
speed 1 and speed 2). Together with the surface, a part of the light is coupled out as
electronically controlled wiper drive, a function of the droplet size and number,
intermittent operation can be varied which attenuates the return signal.
infinitely. For instance, when masses of The integrated light sensor generally
rain water flood the windshield when comprises two or three photodiodes (3, 5)
overtaking, the wiper is automatically which receive and evaluate the light com-
switched to the highest wiper speed. ing from various directions. Depending
Various activation strategies are available: on the function of the light sensor, diodes
• Permanently active are used that reflect the receptiveness of
• Activated via steering column switch the human eye (silicon eyes) or have their
• Activation after steering column switch maximum sensitivity rather in the near
movement (transients) infrared range.
Complex filter functions and trigger
This assistance function frees the driver strategies are stored in the rain/light sen-
from the other control operations used on sor. The status information derived from
conventional wiper systems. Nevertheless, this (e.g. wiper speed 1, intermittent wiper,
manual control is retained as an additional intermittent timing, lights on, lights off)
intervention. are made available to other control units
The rain sensor can also be used for via a data bus (e.g. CAN, LIN).
other functions such as closing the win-
dows and the sliding sunroof.
Application of the light sensor

A light sensor is also incorporated in the
rain sensor, it may be used to control the 1 Rain/light sensor for windshields
low beam. This can detect the various situ-

ations (e.g. twilight, driving into and out of
a tunnel, driving under long bridges) and
switch the low beam on or off accordingly.
The light sensor can also be used to con- 1
trol all the lighting functions on the vehi-
cle. These functions include adjusting the
brightness of the instrument cluster light- Fig. 1
6
ing, “coming home leaving home” or selec- 1 Raindrops
tively switching on the rear lamps. 5 2 Windshield

2 3 Ambient-light
sensor
UAE0675-3Y
4 4 Photodiode
3 5 Light sensor set
up for distance
6 LED
284 Sensor types | Two-step Lambda oxygen sensors
Two-step Lambda oxygen functions even at extremely low exhaust-

sensors gas temperatures.
The sensor’s output voltage US is a func-
tion of the oxygen content in the exhaust
Application gas. In the case of a rich air-fuel mixture
These sensors are used in gasoline engines (l,1) it reaches 800 to 1,000 mV. For a
equipped with two-step Lambda oxygen lean mixture (l.1), only approximately
sensor. They extend into the exhaust pipe 100 mV are generated. The transition from
between the engine’s exhaust manifold and the rich to the lean area at Ureg is at about
the catalytic converter, and record the 450 to 500 mV.
exhaust-gas flow leaving each cylinder. The temperature of the ceramic element
Thanks to the Lambda oxygen sensor be- influences its ability to conduct the oxygen
ing heated, it can be installed further away ions, and thus the shape of the output-volt-
from the engine so that long periods of age curve as a function of the excess-air
full-load engine operation are unproblem- factor l (the values in Figure 1 are there-
atical. The LSF4 sensor is also suitable for fore temperature-dependent). In addition,
operation with exhaust-gas control systems the response time for a voltage change
featuring a number of sensors (for instance when the mixture composition changes is
with OBDII). also strongly dependent on temperature.
Two-step Lambda oxygen sensors com- Whereas response times at ceramic tem-
pare the residual-oxygen content in the peratures below 350 °C are in the seconds
exhaust gas with the oxygen content in the range, at optimum temperatures of around
reference atmosphere (surrounding air 600 °C the sensor responds in less than
inside the sensor) and indicate whether a 50 ms. When the engine is started there-
rich (l < 1) or lean air-fuel mixture (l > 1) fore, the Lambda control is switched off
is present in the exhaust gas. The sudden until the minimum operating temperature
jump in the characteristic curve of these of about 350 °C is reached. During this pe-
sensors permits air-fuel control of l = 1 riod, the engine is open-loop-controlled.
(Fig. 1).
Operating principle 1 Voltage curve of a two-step Lambda oxygen sensor

for different operating temperatures
The two-step Lambda oxygen sensors op-
erate in accordance with the principle of
mV
the galvanic oxygen-concentration cell
with solid-state electrolyte (Nernst princi- Air deficiency Excess air
ple). The ceramic element is conductive 1,000 500°C
for oxygen ions from a temperature of ap- 600
700
Sensor voltage U S
proximately 350 °C (safe, reliable opera- 800 800

900°
tion at >350 °C). Due to the abrupt change C
in the residual-oxygen content on the ex- Increasing
600 temperature
haust-gas side in the range of l = 1 (e.g.
9 · 10–15 % vol. for l = 0.99 and 0.2 % vol. 400
for l = 1.01), the different oxygen content
Increasing
on both sides of the sensor generates an 200 temperature
electrical voltage between the two bound-
Fig. 1
UMK0279-1E
a Rich mixture
ary layers. This means that the oxygen 0
0.8 0.9 1.0 1.1 1.2
(air deficiency) content in the exhaust gas can be used as
Excess-air factor
b Lean mixture a measure of the air-fuel ratio. The inte-
(excess air) grated heater ensures that the sensor
Sensor types | Two-step Lambda oxygen sensors 285
Design 3 Configuration of a finger-type Lambda oxygen

LSH25 finger-type sensor sensor in the exhaust pipe
Sensor ceramic element with protective tube

The solid electrolyte is a ceramic element
which is impermeable to gas. It is a mixed
oxide comprising the elements zirconium
and yttrium in the form of a tube closed at
one end (finger, Fig. 3). The surfaces have 8
7
been provided on both sides with elec- 5
trodes made from a microporous, thin
4
noble-metal layer. 1
The ceramic body protrudes into the 2 3

Fig. 3
exhaust pipe, and the platinum electrode 6 1 Sensor ceramic
on its outside surface acts as a catalytic element
converter in miniature. Exhaust gas which US 2 Electrodes
reaches this electrode is processed cata- 3 Contacts
lytically and brought to a stoichiometrical 4 Housing contact

5 Exhaust pipe
balance (l = 1). In addition, the side that is
6 Ceramic protective
exposed to the exhaust gas has a porous,
UMK1684Y
layer (porous)
ceramic multiple layer (spinel layer) to 7 Exhaust gas
protect it against contamination and 8 Outside air
US Sensor voltage
2 LSH25 heated finger-type Lambda oxygen sensor (view and section)

UMK1450Y
Fig. 2
1 Sensor housing
2 Ceramic support
1 2 3 tube
1 cm
3 Connecting cable
4 Protective tube
with slots
5 Active sensor
ceramic
6 Contact element
7 Protective sleeve
8 Heater element
UMK0143Y
9 Clamp-type
connections for the
4 5 6 7 8 9 10
heater element
10 Disk spring
286 Sensor types | Two-step Lambda oxygen sensors
erosive damage. The ceramic element is The sensor’s complete internal structure is
also protected against mechanical impact located and held in place by a protective
and thermal shocks by a metal tube. A metal sleeve (7) which at the same time
number of slots in the protective tube are acts as the support for the disk spring.
specially shaped so that on the one hand The protective sleeve also guards against
they are particularly effective against ex- contamination of the sensor’s interior.
tensive thermal and chemical stresses, The connecting cable is crimped to the
while on the other hand preventing a contact element which protrudes from the
sharp drop in sensor-ceramic temperature sensor, and is protected against moisture
when the exhaust gas is “cool”. and mechanical damage by a temperature-
The sensor’s “open” inner chamber is resistant cap.
connected to the surrounding air, which The tube-type (finger) Lambda oxygen
acts as a reference gas (Fig. 3). sensor is also equipped with an electrical
heater element. This ensures that the ce-
Sensor element with heater element and ramic-element temperature remains suffi-
electrical connection ciently high, even at low engine load and
A ceramic support tube (Fig. 2, Item 2) thus low exhaust-gas temperature.
and a disk spring (10) serve to locate, fix, This external heating is so quick that
and seal-off the active, finger-shaped sen- the sensor reaches operating temperature
sor ceramic in the sensor housing. A con- 20 to 30 s after the engine has started and
tact element (6) between the support tube therefore Lambda control can come into
and the active sensor ceramic element operation. Finally, sensor heating provides
provides the contact between the inner for an optimal ceramic-element operating
electrode and the connection cable. temperature above the operating limit of
The outer electrode is connected to the 350 °C and thus ensures low and stable
sensor housing by the metal sealing ring. exhaust-gas emissions.
Fig. 4
1 Porous protective
layer
2 Outer electrode
4 Planar Lambda oxygen sensor 5 LSF4 planar Lambda oxygen sensor
3 Sensor foil (functional layers) (schematic diagram)
4 Inner electrode
5 Reference-air-
channel foil
6 Insulation layer
1
7 Heater
8 Heater foil 2 2 3 1
9 Connection
contacts
3
4
UA
5
Fig. 5
1 Exhaust gas 6
2 Porous ceramic 7
protective layer
6
3 Measuring cell
with microporous 8
noble-metal layer 4 5
9
UMK1640Y
UMK1789Y
4 Reference-air
channel
5 Heater
UA Output voltage
Sensor types | Two-step Lambda oxygen sensors 287
LSF4 planar Lambda oxygen sensor It is integrated and insulated in the ce-
Regarding their function, planar Lambda ramic wafer and ensures that the sensor
oxygen sensors correspond to the heated heats up quickly even in the event of low
finger sensors with their voltage-jump power input.
curve at l = 1. On the planar sensor
though, the solid-state electrolyte is The reference-air channel inside the
comprised of a number of individual LSF4 Lambda oxygen sensor – operating
laminated foils stacked one on top of the as a reference-gas sensor (Figs. 5 and 6) –
other (Fig. 4). The sensor is protected has access to the ambient air. This enables
against thermal and mechanical influ- it to compare the residual oxygen content
ences by a double-walled protective tube. of the exhaust gas with the oxygen content
The planar ceramic element (measuring of the reference gas. In other words, with
cell and heater are integrated) is shaped the air inside the sensor. This means that
like a long stretched-out wafer with rect- the sensor voltage also features a charac-
angular cross-section. teristic jump in the case of the planar sen-
The surfaces of the measuring cell are sor, in the area of stoichiometric air-fuel
provided with a microporous noble-metal mixture (l = 1) (Fig. 1).
layer, which also has a porous ceramic
layer on the exhaust-gas side to protect
it against the erosive effects of the ex-
haust-residues. The heater is a wave-
shaped element containing noble metal.
6 LSF4 planar Lambda oxygen sensor (view and section)

UMK1606Y
Fig. 6
1 Planar measuring
cell
2 Double protective
1 cm
tube
3 Sealing ring
4 Seal packing
5 Sensor housing
6 Protective sleeve
7 Contact holder
8 Contact clip
9 PTFE grommet
UMK1607Y
10 PTFE shaped sleeve

1 2 3 4 5 6 7 8 9 10 11 12 11 Five connection
cables
12 Seal
288 Sensor types | LSU4 planar wide-band Lambda oxygen sensor
LSU4 planar wide-band powered engines and gas-powered central

Lambda oxygen sensor heaters and water heaters (hence the Ger-
man designation LSU: Lambda-Sonde-Uni-
Application versal = universal Lambda oxygen sensor).
As its name implies, the wide-band In a number of systems, several Lambda
Lambda oxygen sensor is used across a oxygen sensors are installed for even
very extensive range to determine the greater accuracy. Here, for instance, they
oxygen concentration in the exhaust gas. are fitted upstream and downstream of the
The figures provided by the sensor are an catalytic converter as well as in the indi-
indication of the air-fuel ratio in the com- vidual exhaust tracts (cylinder banks).
bustion chamber. The excess-air factor l
is used when defining the air-fuel ratio. Design
The LSU4 wide-band Lambda oxygen sen-
The sensor projects into the exhaust pipe sor (Fig. 2) is a planar dual-cell current-
and registers the exhaust-gas mass flow limit sensor. It features a measuring cell
from all cylinders. Wide-band Lambda ox- (Fig. 1) made of zirconium-dioxide ceramic
ygen sensors make precise measurements (ZrO2), and is a combination of a Nernst
not only at the stoichiometric point l = 1, concentration cell (sensor cell which func-
but also in the lean range (l>1) and in the tions in the same way as a two-step Lambda
Fig. 1
1 Exhaust gas
rich range (l<1). In combination with con- oxygen sensor) and an oxygen pump cell
2 Exhaust pipe trol electronics, these sensors generate an for transporting the oxygen ions. The oxy-
3 Heater unmistakable, continuous electrical signal gen-pump cell (Fig. 1, Item 8) is so ar-
4 Control electronics in the range from 0.7 < l < ` (air with 21 % ranged with respect to the Nernst concen-
5 Reference cell O2) (Fig. 3). This means that the wide-band tration cell (7) that there is a 10 to 50 µm
with reference-air
Lambda oxygen sensor can be used not diffusion gap (6) between them. Here, there
channel
only in engine-management systems with are two porous platinum electrodes: one
6 Diffusion gap
7 Nernst two-step control (l=1), but also in control pump electrode and one Nernst measuring
concentration concepts with rich and lean air-fuel mix- electrode. The diffusion gap is connected
cell with Nernst tures. This type of Lambda oxygen sensor to the exhaust gas by way of a gas-access
measuring is therefore also suitable for Lambda con- passage (10). A porous diffusion barrier
electrode (on the trol with lean-burn concepts on gasoline (11) serves to limit the flow of oxygen mol-
diffusion-gap side)
engines, as well as for diesel engines, gas- ecules from the exhaust gas.
and reference
electrodes (on the
reference-cell side) 1 Planar wide-band Lambda oxygen sensor (schematic design of the measuring cell and installation in the exhaust pipe)
8 Oxygen-pump 2 3 4
cell with pump
electrode
9 Porous protective US IP
+
layer –
10 Gas-access passage UP
11 Porous diffusion URef
1 11
barrier
10
IP Pump current
UP Pump voltage
UH Heating voltage UH
URef Reference
UMK1260-1Y
voltage (450 mV,

corresponds to
l = 1) 9 8 7 6 5
US Sensor voltage
Sensor types | LSU4 planar wide-band Lambda oxygen sensor 289
On the one side, the Nernst concentration cell trols the voltage (UP) across the pump cell
is connected to the atmosphere by a refer- in order that the composition of the gas in
ence-air channel (5), on the other it is con- the diffusion gap remains constant at l = 1.
nected to the exhaust gas in the diffusion gap. If the exhaust gas is lean, the pump cell
The sensor requires control-electronics pumps the oxygen to the outside (positive
circuitry to generate the sensor signal and pump current). On the other hand, if the
to regulate the sensor temperature. exhaust gas is rich, the oxygen (due to the
An integral heater (3) heats up the sensor decomposition of CO2 and H2O at the ex-
quickly so that it soon reaches its operating haust-gas electrode) is pumped from the
temperature of 650 to 900 °C needed for surrounding exhaust gas and into the diffu-
generating a usable signal. This function de- sion gap (negative pump current). At l = 1,
cisively reduces the effects that the exhaust- no oxygen needs to be transported, and the
gas temperature has on the sensor signal. pump current is zero. The pump current
is proportional to the exhaust-gas oxygen
Operating principle concentration and is this a nonlinear mea-
The exhaust gas enters the actual measur- sure for the excess-air factor l (Fig. 3).
ing chamber (diffusion gap) of the Nernst
concentration cell through the pump cell’s 3 Pump current IP of a wide-band Lambda oxygen
small gas-access passage. In order that the sensor as a function of the excess-air factor l
of the exhaust gas
excess-air factor l can be adjusted in the
diffusion gap, the Nernst concentration mA
cell compares the gas in the diffusion gap

with that in the reference-air channel. 1
Pump current Ip
The complete process proceeds as fol-

lows: by applying the pump voltage UP 0
across the pump cell’s platinum elec-

trodes, oxygen from the exhaust gas can be -1
UMK1266-1E
pumped through the diffusion barrier and

into or out of the diffusion gap. With the -2
0.7 1 2 3 4
help of the Nernst concentration cell, an
Excess-air factor l
electronic circuit in the control unit con-
Fig. 2
2 LSU4 planar wide-band Lambda oxygen sensor (view and section)
1 Measuring cell
(combination
of Nernst
concentration
cell and oxygen-
pump cell)
2 Double protective
tube
3 Sealing ring
1 cm
4 Seal packing
5 Sensor housing
6 Protective sleeve
7 Contact holder
8 Contact clip
9 PTFE grommet
UMK1607Y
10 PTFE shaped sleeve

11 Five connecting
cables
1 2 3 4 5 6 7 8 9 10 11 12
12 Seal
290 Electric Actuators | Electromechanical actuators
Electric Actuators
Actuators (final-control elements) form Electromechanical actuators

the interface between the electronic sig-
nal processor (data processing) and the This type of energy conversion represents
actual process (mechanical motion). They one option for classifying electromechani-
convert the low-power signals conveying cal actuators. The energy emanating from
the positioning information into operat- the source is transformed into magnetic
ing signals of an energy level adequate for or electrical field energy, or converted to
process control. Signal transducers are thermal energy. The individual force-gen-
combined with amplifier elements to ex- eration principle is determined by these
ploit the physical transformation princi- forms of energy, and is based on either
ples governing the interrelationships be- field forces or certain specific material
tween various forms of energy (electrical – characteristics. Magnetostrictive materials
mechanical – fluid – thermal). make it possible to design actuators for ap-
plications in the micropositioning range.
1 Actuator chain This category also includes piezoelectric
actuators, which are built according to a
Fig. 1 multilayer design similar to ceramic ca-
2 3 4
1 Information pacitors, and are actuators for high-speed
electr. el.
mech. hydr.
2 Actuator
1 fuel injectors. Thermal actuators depend
3 Converter mech. exclusively on the exploitation of charac-
4 Actuator
teristics of specific materials.
5 Losses 5 5 5
UAN0054-1E
Actuators in a motor vehicle are mostly

6 External electrical
6 7 electromagneto-mechanical converters and,
energy
7 External hydraulic by extension, electrical servomotors, trans-
energy lational, and rotational solenoid actuators.
2 Electromechanical converters (system overview)
Electrical voltage
Electric current
Magnetic energy Electrical energy Thermal energy

Magnetic field Electric field Temperature field
Magneto- Magnetic Piezoelectric Electric Heat Phase Com-

striction field force effect field force expansion conversion bustion
Electro-
magneto-
mechanical
Micro Micro Micro Bi- Memory Airbag
converter
UAE0308-1E
actuator actuator actuator metal metals

Elec. motor
Solenoid
actuator

Electric Actuators | Electromechanical actuators 291
An exception is the pyrotechnic airbag A field coil or a permanent magnet gener-

system. The solenoid actuators can them- ates a DC magnetic field (magnetic flux
selves be the servo element, or they can density B). The electrical energy destined
assume a control function by governing for conversion is applied (current I ) to the
a downstream amplifier (e.g. mechanical- moving armature coil (immersion coil).
hydraulic). A high degree of actuator precision is
achieved by designing the armature coil
Force generation in the magnetic field with low mass and low inductance. The
The distinction between the electrody- two energy accumulators (one on the fixed
namic and the electromagnetic actuator and one on the moving component) pro-
principles stems from the manner in which duce two active force directions via cur-
forces are generated in the magnetic field. rent-direction reversal in the armature
Common to both principles is the magnetic and field coils.
circuit with soft-magnetic material and the The secondary field produced by the
coil for excitation of the magnetic field. armature current flows in an open mag-
A major difference lies in the force which netic circuit, thereby diminishing the ef-
can be extracted from the unit under tech- fects of saturation. Approximately speak-
nically feasible conditions. Under identical ing, the force (torque) exerted by an elec-
conditions, the force produced by applica- trodynamic actuator over its setting range
tion of the electromagnetic principle is is proportional to current and indepen-
greater by a factor of 40. The electrical dent of travel.
time constant for this type of actuator is
comparable to the mechanical time con- Electromagnetic principle
stants. Both force-generation principles The electromagnetic principle exploits
are applied in linear and rotary drive the mutual attraction displayed by soft
mechanisms. iron materials in a magnetic field. The elec-
tromagnetic actuator is equipped with only
Electrodynamic principle one coil, which generates both the field
The electrodynamic principle is based energy and the energy to be transformed.
on the force exerted on moving charges In accordance with the operating princi-
or charged conductors (Fig. 3a, current I) ples, the field coil is equipped with an iron
within a magnetic field (Lorentz force). core to provide higher inductance.
3 Electrodynamic and electromagnetic converters
a b
F
Fig. 3
I a Electrodynamic
B
B principle:
N i d I
Force on moving
F d or charge
F = B · I · l · sin a
S
b Electromagnetic
UAE0309-1E
i principle:
Force between
pole faces
F = B2 · A/2 µ0
However, as the force is proportional to Electromagnetic actuators

the square of the magnetic flux density B, A typical form for translational electro-
the unit is operative in only a single force- magnetic actuators is the switching sole-
transfer direction. The electromagnetic noid (Fig. 4) with a force/travel character-
actuator thus requires a return element istic curve which falls as a function of the
(such as a mechanical spring or a magnetic square of positioning travel. The precise
return mechanism). shape of the curve (Fig. 5) is determined
by the type of working air gap (e.g. conical
Dynamic response or solenoid armature).
The dynamic response of an electrome-
chanical actuator, i.e. the activation and Rotational electromagnetic actuators are
deactivation operations, is defined by the characterized by a defined pole arrange-
equation of mechanical motion, the differ- ment in stator and rotor. When current is
ential equation of electrical circuits and applied to one of the coils, the rotor and
Maxwell’s equations of dynamics. The stator poles respond with mutual attrac-
current- and position-dependent force tion, and in doing so generate a torque.
follows from Maxwell’s equations.
The most basic electrical circuit con- 4 Switching solenoid
sists of an inductance with an ohmic re-
sistor. One means of enhancing the dy-
namic response is through over-excitation
s F
at the instant of activation, while deacti-
vation can be accelerated by a Zener di-
ode, for example. In each case, increasing
the dynamic response of the electric cir-
cuit involves additional expenditure and 1
increased losses in the actuator’s trigger-
ing electronics. 2
Field diffusion is a delay effect which
UAE0310-1Y
3
Fig. 4 is difficult to influence in actuators with
1 Armature high dynamic response. Rapid switching
2 Coil
operations are accompanied by high-
3 Magnetic yoke
frequency field fluctuations in the soft-
magnetic material of the actuator’s mag-
5 Switching solenoid (characteristic curves)
netic circuit. These fluctuations, in turn,
induce eddy currents, which counteract
their cause (build-up and decay of the
magnetic field). The resultant delay in
the build-up or reduction of forces can
1
only be reduced by selecting appropriate
materials with low electric conductivity
Force F
and permeability.
Design 2
Design selection is essentially determined
Fig. 5
by operating conditions (i.e. installation
UAE0311-1E
1 Solenoid plunger 3
2 Conical armature space, required force/travel characteristic
3 Cylindrical curve, and dynamic response). Travel s
armature
Electric Actuators | Electromechanical actuators 293
The single-winding rotary actuator (Fig. 6)

6 Electromagnetic single-winding rotary actuator
incorporates a pair of poles in each of the
two main sections, as well as a coil in the
M stator. Its maximum adjustment range is
w approx. 45°.
1
The torque motor (Fig. 7) is a bidirectional
2 electromagnetic rotary actuator featuring
a stable operating point and without coun-
3 terforces. The rotor is maintained in a sta-
ble position by the excitation field of the
permanent magnet in the stator. The mag-
UAE0312-1Y
netic field generated by one or two stator Fig. 6

windings produces torque and provides 1 Coil
unilateral compensation for the excitation 2 Stator
field. This type of layout is suitable for 3 Armature
applications in which substantial torque

7 Torque motor is required over small control angles.
The relationship between the applied
current and the torque motor’s force is
N S roughly linear (Fig. 8). The torque-motor
1 principle is also employed in translational
actuators.
2
Electrodynamic actuators
In a pot magnet (immersion-coil actuator,
3
Fig. 9) a cylindrical immersion coil (arma-
ture coil) is set in motion in a working air
1 gap. The adjustment range is limited by the
UAE0313-1Y
N S axial length of the armature coil and by the Fig. 7
air gap. 1 Magnets

2 Armature
The short-stroke linear motor (Fig. 10) is
3 Control coils
an actuator with a virtually round disk coil.
8 Electromagnetic rotary actuators (characteristics)

A distinction is made between single-wind-
ing and dual-winding rotary actuators
(Fig. 11). Both types include a permanent
magnet within the rotor and one or two
stator windings. The rotor magnet is mag-
netized at both ends to produce magnetic
L flux density in the rotary magnet’s working
1
Torque M
air gap, which combines with the armature

current to produce a torque. Originating
from the position illustrated, the adjust-
ment range is less than ±45°. The setting
UAE0314-1E
range of the single-winding rotary actuator Fig. 8

2
also varies according to both the torque 1 Single-winding
Control angle w requirement and the angle range in which rotary actuator
the necessary flux density can be provided. 2 Torque motor
The dual-winding rotary actuator can be

9 Electrodynamic immersion-coil actuator
described as a combination of two single-
winding rotary actuators with a 90° pe-
ripheral offset and designed to produce
opposed torque flows. A stable operating
point is achieved at the zero transition 1
point on the resulting torque curve with-
out auxiliary counterforces.
Application
3 2
Electromechanical actuators are direct-
action control elements, and without an
UAE0315-1Y
Fig. 9 intermediate ratio-conversion mechanism,
1 Immersion coil they convert the energy of the electrical
2 Permanent magnet control signal into a mechanical position-
3 Magnetic yoke ing factor/work. Typical applications in-
clude positioning of flaps, sleeves, and
valves. The described actuators are final- 10 Electrodynamic short-stroke linear motor
control elements without internal return
mechanisms, i.e. without a stable operat-
ing point. They are only capable of carry- 1 2 2 3
ing out positioning operations from a
stable initial position (operating point)
when a counterforce is applied (e.g.
return spring and electrical control).
A solenoid armature provides a stable
static operating point when its force/travel
curve is superimposed on the characteris-
tic response of a return spring. A variation
of the coil current in the solenoid shifts UAE0316-1Y
Figure 10
1 Coil the operating point. Simple positioning
2 Permanent magnet
is achieved by controlling the current.
3 Magnetic yoke
However, particular attention must be
paid here to the nonlinearity of the force-
11 Electrodynamic rotary actuator
current characteristic and the positioning
system’s sensitivity to interference factors
(e.g. mechanical friction, pneumatic, and
hydraulic forces). The temperature sensi-
a b
tivity of the coil resistance results in posi-
tioning errors, making corrective current 1
Fig. 11
a Single-winding control necessary. A high-precision posi- 3
rotary actuator tioning system with good dynamic response 5
b Dual-winding must incorporate a position sensor and a
rotary actuator controller. 4
1 Coil 1 2
UAE0317-1Y
2 Coil 2
3 Stator
4 Permanent magnet
5 Shaft
Electric Actuators | Fluid-mechanical actuators 295
12 Electrodynamic rotary actuator (characteristic curves) Fluid-mechanical actuators

a
Hydraulic and pneumatic actuators utilize
similar principles for the conversion and
Torque M
regulation of energy. Table 1 shows the

differences in characteristics and applica-
tions.
0
In most applications, fluid-mechanical
b actuator drives are in the form of hydro-
1 static energy converters. These operate
according to the displacement principle,
(i1 = i2) converting the pressure energy of the Fig. 12
Torque M
0 fluid medium into mechanical work and a Single-winding

vice versa (Fig. 14). rotary actuator
In contrast, hydrodynamic transformers b Dual-winding
3 rotary actuator
2 operate by converting flow energy (kinetic
UAE0318-1E
energy of the moving fluid) into mechani-

-45 0 45 90 1 Coil 1
Control angle w
cal work (example: hydrodynamic clutch). 2 Coil 2
Losses during energy conversion stem 3 Coils 1+2
from leakage and friction. Fluid-thermal
13 Operating points (A) of a linear solenoid losses are caused by flow resistance, in
which throttle action transforms the hy-
draulic energy into heat. A portion of this
heat is dissipated into the environment,
A1 and some of it is absorbed and carried
1
away by the fluid medium:
A2
A3 Qheat = Q1 · p1 – Q2 · p2
Force F
2
With incompressible fluids:
3
Qheat = Q1 · (p1 – p2) Fig. 13
4 1, 2, 3 Characteristic
UAE0319-1E
curves for various

currents
Travel s 4 Characteristic curve
of a return spring
1 Comparison of hydraulic and pneumatic actuators

Hydraulic actuators Pneumatic actuators
Medium – Fluid, mainly oil – Gas, mainly air
– Supplied from tank, oil sump – Supplied from ambient air
– Virtually incompressible – Compressible
– Self-lubricating – Additional lubrication necessary
– Viscosity highly temperature-dependent – Viscosity changes practically irrelevant
Pressure – To approx. 30 MPa – To approx. 1 MPa or greater
section (200 MPa for diesel-fuel injectors) (approx. 0.05 MPa for vacuum actuators)
Line – Supply and return connection – Pressure connection only,
connections (possible leakage connection) return directly to environment
Applications – Positioning applications with high load – Actuators with low power requirement,
rigidity, demanding requirements for positioning by mechanical stops, in open
synchronization and positioning precision control loop
in closed-loop control system
Table 1
296 Electric Actuators | Fluid-mechanical actuators
The flow develops into turbulence at re- Electrical machines

strictions. The flow rate of the fluid is then
largely independent of viscosity. On the Electrical machines are used to convert
other hand, viscosity does play a role in electrical and mechanical energy. An elec-
the case of laminar flow in narrow pipes tric motor converts electrical energy into
and apertures. mechanical energy, and an alternator con-
Fluid-mechanical amplifiers control verts energy in the opposite direction.
the conversion of energy from fluid to Electrical machines (Fig. 16) consist of a
mechanical state. The regulating mecha- stationary component (the stator) and a ro-
nism must be designed for control with tating component (the rotor or armature).
only a very small proportion of the energy There are special designs which depart
flow required for the ultimate positioning from this configuration, such as linear
operation. machines which produce linear motion.
Switching valves open and close the Permanent magnets or several coils
orifice governing the flow to and from (windings) are used to produce magnetic
a fluid-mechanical energy converter fields in the stator and rotor. This causes
(Fig. 15). Provided that the control-ele- torque to develop between the two ma-
ment opens sufficiently, the throttling chine components. Electrical machines
losses remain negligible. Pulse-width- have iron stators and rotors in order to
modulated opening and closing can be control the magnetic fields. As the mag-
applied to achieve quasi-continuous con- netic fluxes change over time, stators and
trol of the fluid-mechanical energy con- rotors must consist of stacks of individual
version process with virtually no losses. laminations which are insulated with re-
In practice, however, pressure fluctuations spect to one another (to minimize eddy-
and mechanical contact between the valve current losses).
elements result in undesirable vibration The spatial arrangement of the coils
and noise. and the type of current used (direct cur-
rent, alternating current, or three-phase
current) permit a number of different elec-
trical machine designs. They differ from
one another in the way they operate, and
therefore have different applications.
14 Fluid-mechanical actuators 15 Energy regulation with 2/2-way switching valves
Fload
a b
w pA QA
s
QA F QA pA
pA A
Fig. 14 M
a Linear actuator
F = pA · A p1 Q1 p2 Q2
s = QA / A
UAH0103-1Y
UAH0104-1E
b Rotary actuator
M = (pA · Vth)/2 p
w = (QA / Vth) · 2 p
Electric Actuators | Electrical machines 297
Direct-current machines 16 Layout of an electrical machine

The stator of a direct-current machine
contains salient poles which are magne-
tized by the direct-current excitation N 1
windings. In the rotor (here also called
the armature), the coils are distributed
among slots in the laminated iron core 2
and connected to a commutator. Carbon 3
brushes in the stator housing wipe against U Fig. 16
1 Stator
the commutator as it rotates, thereby 4
2 Rotor
transferring direct current to the armature 3 Rotor coil
coils. The rotary motion of the commutator 4 Commutator
causes a reversal in the direction of cur-
UAE1061Y
S Stator field generated
rent flow in the coils. The different rota-
by permanent magnet
tional speed vs. torque characteristics re-
or electromagnet
sult from the method selected for connect-
ing the excitation winding and armature.
17 Series-wound machine
Series connection (series characteristics,

Fig. 17)
• Clear relationship between speed and
load
• High starting torque
Rotational speed
• Unacceptably high speed possible
at the moment of load reduction,

requiring therefore a rigid connection M
with the load
• Torque reversal (change in direction
Torque
of rotation) caused by a change in the
UAE0276-1E
current direction in the armature or

excitation winding
• Application, e.g. as vehicle drive motor
or starter for internal-combustion en-

gines
18 Shunt-wound machine
Parallel connection (shunt characteristic,

Fig. 18)
• Speed varies only slightly with load
• Change in direction of rotation caused
Rotational speed
by a change in the current direction in

the armature or excitation winding
• Application, e.g. as drive motor for ma-
M
chine tools or direct-current generators
Torque
UAE0277-1E
298 Electric Actuators | Electrical machines
A shunt characteristic can also be obtained Asynchronous machines

by using a separate power supply for the The laminated rotor contains either a
excitation winding (external excitation) three-phase winding, as in the stator, or a
or by using permanent-magnet excitation rod winding. The three-phase winding is
in the stator. connected to sliprings which are short-cir-
Applications for permanent-field motors cuited either directly or via series resistors
in motor vehicles: starter, windshield (Fig. 19). In the case of the rod winding,
wiper, and low-power motors for various the bars are connected to one another by
drives. two short-circuiting rings (squirrel-cage
rotor). As long as the rotational speed of
If the motor incorporates both series and the rotor deviates from n0, the rotating
shunt excitation windings (compound- stator field induces current in the rotor
wound motor), intermediate levels in the windings, thereby generating torque.
rotational speed/torque characteristic Deviation of the rotational speed of the
can be obtained. rotor n from n0 is termed slip s:
Application: e.g. large starters. s = (n0 − n) / n0
All direct-current machines are easily ca- Continuous-running operation is only

pable of speed control over a wide range. economical in the vicinity of n0 because
If the machine incorporates a static con- losses increase as slip increases (nominal
verter which allows adjustment of the ar- slip ≤ 5%). In this range the asynchronous
mature voltage, the torque and therefore machine has a shunt characteristic. The
the rotational speed are infinitely variable. machine operates as a motor when n < n0,
The rotational speed can be further in- and as an alternator when n > n0. The di-
creased by reducing the excitation current rection of rotation is changed by reversing
(field weakening) when the rated armature two of the phases.
voltage is reached. A disadvantage of di- The asynchronous machine is the most
rect-current machines is carbon-brush frequently used electric motor in the field
and commutator wear which makes regu- of drive engineering. With a squirrel-cage
lar maintenance necessary. rotor it is easy to operate, and requires
little maintenance.
Three-phase machines
A three-phase winding is distributed 19 Three-phase asynchronous motor with squirrel-
among the stator slots in a three-phase cage rotor
machine. The three phases of current
produce a rotating magnetic field. a b
The speed n0 (in rpm) of this rotating
field is calculated as follows:
n0 = 60 ⋅ f / p
f = frequency (in Hz)

p = number of pole pairs
Three-phase machines are either synchro- M M

Fig. 19
nous or asynchronous, depending on rotor
3~ 3~
UAE0279-1Y
a Stator winding,
star-connected design.
b Stator winding,
delta-connected
motor speed based on constant line fre-

2 Examples of rotating-field speeds
quency is desired, or where a reactive
No. of poles Frequency
(2 p) 50 Hz 150 Hz 200 Hz
power demand exists. The automotive
Rotating-field speed in rpm three-phase alternator is a special type
2 3,000 9,000 12,000 of synchronous machine.
4 1,500 4,500 6,000 The rotational speed of all three-phase
6 1,000 3,000 4,000 machines is determined by the stator fre-
8 750 2,250 3,000 quency. Such machines can operate over
10 600 1,800 2,400 a wide range of speeds if used in conjunc-
12 500 1,500 2,000 tion with static converters which vary the Table 2
Synchronous machines frequency.
In the rotor (here, also called the pole
wheel), the poles are magnetized by di- EC motors
rect-current coils. The excitation current The “electronically commutated direct-
is usually transferred via two sliprings to current” or “EC” motor is becoming in-
the rotor. The pole wheel can be made of creasingly popular (Fig. 21). It is essen-
solid steel, because the magnetic flux re- tially a permanent-magnet synchronous
mains constant over time. Constant torque machine, and is brushless. The EC motor
is generated as long as the rotor rotates at is equipped with a rotor-position sensor,
a speed of n0. At other speeds, the torque and is connected to the DC power source
fluctuates periodically between a positive through its control and power electronics.
and a negative maximum value, and exces- The electronic transfer circuit switches the
sively high current is produced. For this current in the stator winding according to
reason, a synchronous machine is not self- rotor position – the magnets which induce
starting. The synchronous machine also the excitation current are attached to the
differs from the asynchronous machine rotor – to provide the interdependence be-
in that the reactive power absorption and tween rotational speed and torque which
generation are adjustable. is normally associated with a separately
The synchronous machine is most fre- excited direct-current machine. The re-
quently used as a generator of energy sup- spective magnetic functions of the stator and
ply in electric power plants. Synchronous rotor are the opposite of what they would be
motors are used in cases where constant in a classical direct-current machine.
20 Star-connected three-phase synchronous generator 21 EC motor
2
Fig. 20
3 S
Slip-ring rotor with
excitation winding
G
N
1
Fig. 21
1 Electrical machine
with rotor-position
UAE0280-1Y
UAE0282-1Y
sensor
2 Control and power
electronics
3 Input
300 Electric Actuators | Electrical machines
The EC motor’s potential applications Machines which are intended only for
are a result of the advantages which this single-phase operation have only a single-
drive principle provides: commutator and phase main winding in the stator, as well
carbon brushes are replaced by electron- as auxiliary starting circuits. The stator
ics, dispensing with both brush noise and also contains an auxiliary winding con-
wear. EC motors are maintenance-free nected in parallel with the main winding
(long service life) and can be constructed for this purpose. The necessary phase dis-
to meet high degrees of protection. The placement of the auxiliary winding current
electronic control feature makes it easy can be achieved through increased wind-
for drive units with EC motors to incorpo- ing resistance (low breakaway torque) or
rate auxiliary functions such as infinitely by means of a capacitor connected in se-
variable speed governing, direction rever- ries with the auxiliary winding (somewhat
sal, soft starts, and antilock protection. greater breakaway torque).
The auxiliary winding is switched off
The main areas of automotive application after the motor starts. The direction of ro-
are in the HVAC (Heating/Ventilation/Air- tation of the motor is changed by reversing
Conditioning) sectors, and for pumps and the two auxiliary or main winding connec-
servo units. In the area of production ma- tions. A motor which has a capacitor in
chinery, EC motors are chiefly employed series with the auxiliary winding is called
as precision drive units for feed-control a capacitor motor. Capacitor motors with
in machine tools. Here the decisive advan- a starting and running capacitor also oper-
tages are freedom from maintenance, ate continuously with capacitor and auxil-
favorable dynamic properties and consis- iary winding. Optimum operation for a
tent torque output with minimal ripple. specific operating point can be achieved
by correct selection of capacitor. An addi-
Single-phase alternating-current tional capacitor is often used in order to
machines increase the starting torque; this capacitor
Universal motors is then disconnected after the motor starts.
The direct-current series-wound motor
can be operated on alternating current if
a laminated rather than a solid iron stator
is used. It is then called a universal motor.
When operated on alternating current,
a torque component at twice the frequency
of the current is superposed on the con-
stant torque component. 22 Two-value capacitor motor
Single-phase asynchronous motors

with squirrel-cage rotor
The simplest design of a single-phase
asynchronous motor is a three-phase
asynchronous machine in which alternat-
ing current is supplied to only two stator
phases. Although its operation remains
largely the same, the power and the maxi- M
mum torque are reduced. In addition, 1~
UAE0281-1Y
the single-phase asynchronous machine

is not self-starting.
Duty-type ratings for electrical machines S7: Uninterrupted duty

(VDE 0530) Operation with starting and braking.
S1: Continuous-running duty
Operation under constant load (rated S8: Uninterrupted duty
output) of sufficient duration to reach Operation with pole-changing.
the thermal steady-state condition. For S3 and S6, the duty cycle time is 10 mins
unless otherwise agreed; and recommended
S2: Short-time-duty type values for relative switch-on duration are
Operation under constant load is so brief 15, 25, 40 and 60 %. For S2, S3 and S6, the
that the thermal steady-state condition operating time or the duty cycle time and
is not reached. The rest period is so long the switch-on duration are to be specified
that the machine is able to cool down to after the rating; the duty cycle time is only
the temperature of the coolant. to be specified if it differs from 10 mins.
Recommended short-time-duty types: Example: S2 – 60 mins, S3 – 25 %.
10, 30, 60 and 90 min.
Relative switch-on duration
S3 to S5: Intermittent-periodic duty The relative switch-on duration is the ratio
Continuous alternating sequence of load of the loading period, including starting
and idle periods. The thermal steady-state and braking, to the cycle time.
condition is not reached during the load
period or during the cooling period of Winding temperature
one duty cycle. The mean temperature t2 of the windings
• S3 intermittent-periodic duty without of an electrical machine can be determined
influence of starting on temperature by measuring the resistance (R2) and refer-
• S4 intermittent-periodic duty with ring it to an initial resistance R1 at a tem-
influence of starting on temperature perature t1:
• S5 intermittent-periodic duty with
R2 – R1
influence of starting and braking on t2 = ––––––––– (τ + t1) + t1
R1
temperature
where
S6: Continuous operation with intermittent
loading 1 − 20 K
τ = ––––
a
Operation with intermittent loading.
Continuous alternating sequence of load a = temperature coefficient
periods and no-load periods, otherwise
as S3.
302 Electrohydraulic Actuators Application, Function, and Requirements
Electrohydraulic Actuators
Electrohydraulic actuators form the interface The actuators are important switching and
between electrical signal processing (data control elements in an electrohydraulic trans-
processing) and the system process (mechan- mission-control system. They control the oil
ics). They convert the low-power command (fluid) flow and the pressure characteristics in
signals into an actuating force with the in- the hydraulic control plate. The following dif-
creased power required for the process. ferent actuator types are available:
1. On/off solenoid valves (on/off, o/o)
Application and Function 2. Pulse-width modulated
solenoid valves (PWM)
The transmission types most commonly in 3. Pressure regulators
use today (AT, CVT, AST) are equipped with (spool-type) (PR-S)
actuators to perform the most varied of 4. Pressure regulators
functions. Table 1 provides an overview of (flat-seat-type) (PR-F)
the most important applications and high-
lights the link between the transmission In most automatic transmissions, these
functions and the actuator types that can be actuators currently serve as pilot-control
used. elements, whose outlet pressure or volumetric
flow is boosted in the hydraulic control plate
before operating the clutches. In contrast, di-
rect-action control elements without this
1 Transmission functions and associated actuators boosting facility supply the clutches with suit-
Automatic transmission (AT) ably high pressure and volumetric flow.
Function Actuator type
Requirements
On/Off
PR-S
PWM
PR-F
The installation location (on the transmission

쐌 Control main pressure X X X
inside the oil pan) gives rise to far-reaching
쐌 Initiate gear change: X X
1-2-3-4-5-6 requirements and demanding operating con-
쐌 Modulate shift pressure X X X ditions for the actuators that are used. Figure
쐌 Control torque converter X X X 1 provides a summary of these requirements.
lockup clutch As ever more transmissions are now being
쐌 Reverse-gear lock X
filled with lifetime oil (i.e., they do not need
쐌 Safety functions X X
to have their oil changed) abrasion and dirt
Continuously variable transmission (pulley CVT)
particles from breaking-in remain in the oil
system throughout the entire operation of the
Function Actuator type
transmission. Even central intake filters and
On/Off
individual filters on the actuators can only re-

PR-S
PWM
PR-F
tain particles over a specific size. Excessively

쐌 Adjust gear ratio X X fine filters would soon clog up.
쐌 Control belt tension X X In addition to these considerations, the
쐌 Control start-up clutch X X X working life of transmissions is being increas-
쐌 Reverse-gear lock X
ingly extended: Transmissions in standard pas-
senger cars are now expected to last at least
Automated shift transmission (AST) 250,000 km (155,000 miles), and much more
Standard electric-motor actuation: for taxis and similar applications.
쐌 Initiate gear change
In contrast to many other electromagnets
쐌 Actuate clutch
쐌 Safety functions (fail-safe)
(e.g. in ABS valves), actuators used in trans-
mission-control applications must be designed
Table 1 over the entire temperature range for 100%

Electrohydraulic Actuators Requirements, Design and Operating Concept 303
ON time, because, for example, they have to PWM switching valves are suitable because
maintain the pressure while the gear is held or of their pulse-width-modulated input signal
control a torque converter lockup clutch dur- (constant-frequency current, variable ratio of
ing driving. This creates a need to limit the ON to OFF time) for use as pressure actua-
power loss and to implement copper windings tors, whose outlet pressure is proportional or
of suitably generous dimensions. inversely proportional to the pulse duty fac-
Vehicle and transmission-specific function tor. The characteristic curve is referred to as
characteristics (switching performance, con- rising or falling.
trol response, technical characteristic data), Pressure-control valves finally are operated
electrical and hydraulic interfaces tailored to with a controlled input current and can
individual applications, and the drive towards likewise be designed with a rising or falling
miniaturization and reducing costs are fur- characteristic. These valves are activated
ther boundary conditions for developing ac- by analog means whereas PWM valves are
tuators for transmission-control applications. activated digitally.
Design and
Operating Concept
Automatic transmissions require switching
valves for the simple on/off switching opera-
tions and/or proportional valves for infinitely 1 Boundary conditions for actuator development
variable pressure control. Figure 2 shows dif-

Dirt tolerance Fatigue limit
ferent ways of converting an input signal (rubbed-off elements, residues) …5000 h (250,000 km)
(current or voltage) into an output signal
Media resistance ON time
(pressure). Basically, actuators can be (ATF with additives 100%
designed with proportional or inversely and water content)
Vibration strength
proportional performance characteristics. Temperature of engine: up to 30 g
On/off valves are usually voltage-controlled, load
(¯40…160°C) Miniaturization
i.e., the battery voltage is applied to the cop- (mass and space
Temperature change reduction)
per winding. The hydraulic section of the
æ STS0286E
valve is designed as either a normally closed

External mounting
contact (n.c.) or as a normally open contact (splash water, salt spray, industrial climate)
(n.o.).
2 Actuators (typical characteristics for switching and proportional valves)
Switching valves Proportional-control valve
On/Off valve Pulse-width-modulated Pressure-control valve

(PWM) valve
Input
(Current I)
Current I
n.c. Time t
Pressure p
Output
(Pressure p) n.o.
Pressure p
æ STS0285E
Time t Duty factor Current I

304 Electrohydraulic Actuators Actuator Types
Actuator Types On/Off Solenoid Valves

On/off solenoid valves (Figure 2) are most
Overview commonly used as simple 3/2 valves (3 hy-
The following types are examples of actuators draulic ports/2 positions). They are clearly
commonly used for transmission-control simpler in design than the 4/3-valves more
purposes together with their characteristic commonly found in stationary hydraulics,
values and features (overview, Figure 1). The and are therefore inexpensive. Nor do they
examples are pilot-control actuators which have the drawbacks of 2/2 valves (such as
operate within a pressure range of 400 to ap- high leakage or limited flow rate, necessitated
prox. 1000 kPa and act on a booster element by their action in combination with an exter-
in the transmission’s hydraulic control sys- nal restrictor).
tem. Spool plungers in the hydraulic control
system boost the pressure and/or the volu-
metric flow. 2 On/off solenoid valves (view and oil flow)
The actuators described can be designed in a

any concrete application at their interfaces in
line with the conditions in the transmission,
e.g.
쐌 mechanically (mounting),
æ UTS0313Y
쐌 geometrically (installation area),
쐌 electrically (contacting), or
쐌 hydraulically (interface to control plate).
b
Fmag
Fig. 2 The function data must be constructively
a View as switching adapted to the requirements in the trans- 1
valve with ball seat mission, particularly with regard to
b Oil flow in valve 쐌 feed pressure and pclutch
1 Return to tank
쐌 dynamics (i.e., reaction speed and control pfeed
Fmag Magnetic force
Ffeed Feed-pressure
stability). Ffeed
æ UTS0287E
force
pfeed Feed pressure
pclutch Clutch pressure
1 Actuators (overview with examples)
Solenoid valves Hydraulic module Pressure regulators
Switching valves Flat-seat valves
PWM valves Spool valves

æ UTS0288E
Electrohydraulic Actuators Actuator Types 305
On/off solenoid valves are mainly used in Figure 4 shows the characteristic of the 3/2
simple 3-speed or 4-speed transmissions switching valve with the pressure switching
without overlap control. However, they are cycles.
being used increasingly less in progressive or
complex transmission-control systems (possi- A switching valve of this type offers the
bly still for safety functions). A pressure regu- following features:
lator controls the gearshift. 쐌 low costs,
쐌 non-susceptible to dirt and contamination,
In the 3/2 switching valve (n.o.) depicted in 쐌 low leakage, and
Figure 3, the feed pressure pfeed generated by 쐌 simple actuation electronics.
the transmission pump is applied before the
flange (P) and seals the ball seat. This feature Areas of application for On/off switching
is referred to as self-sealing. Since no pressure valves are:
is applied at zero current in the working-
pressure channel (A), the valve is closed at 쐌 Gear change
zero current. with multiple use of the same main
In this state, the working pressure, which in pressure regulator.
the end supplies the consumer (e.g. a clutch),
is directly connected to the return line to the 쐌 Safety functions
tank (oil pan) so that a pressure that is ap- e.g. hydraulic disabling of the reverse gear
plied there can reduce or an oil volume con- during forward travel.
tained therein can drain.
When current is applied to the switching- 쐌 Torque converter lockup clutch
valve winding, the magnetic force generated Cutting in and out (often controlled with a
reduces the working air gap, and the arma- PWM valve or pressure regulator for econ-
ture moves with its permanently connected omy or convenience reasons).
push rod towards the ball and opens it. Oil
flows to the consumer (from P to A) and 쐌 Changeover of register pump
builds up the pump pressure there. The Selected for two different flow-rate ranges
return line to the tank is closed off at the via o/o (mainly for CVT transmissions).
same time.
Fig. 3
3 3/2 switching valve (sectional view) 4 3/2 switching valve A Working-pressure
(characteristic with switching cycles)
channel
kPa P Feed
T Return to tank
T 600 pfeed Feed pressure
A
500 Characteristic data
Pressure p
P (typical example):
pfeed 400
Feed 400...600 kPa
pressure
300
Flow rate > 2.5 l · rpm
Operating 9...16 V
200
voltage
100 Resist- 12.5 ⏐
æ UTS0289E
æ UTS0290E
ance
0 Number 2 · 106
0 10 20 30 40 ms of switching
Time t
cycles
PWM Valves These requirements are reflected in the

PWM valves (Figure 5) have essentially the design. The sectional view in Figure 6 shows
same design as switching valves. Since they the incorporation of a rim on the armature,
operate with a frequency of 30...100 Hz, they which ensures a high magnetic force and
must be designed for a higher switching a high hydraulic damping action when
speed (dynamics) and higher mechanical load closing. The relatively long push rod absorbs
(wear). The latter is particularly applicable if pulse forces. An annular seat seals the feed
a PWM valve is used for main-pressure con- pressure in the zero-current state shown
trol and is operated throughout the life of the against the working-pressure channel.
vehicle. Compared with the ball seat, the advantage
of this type is that the feed pressure acts only
on minimal surface areas. These surfaces are
5 PWM valve (view and oil flow) still pressure-compensated to the greatest ex-
tent, i.e., the applied feed pressure has an
opening and closing effect such that essen-
tially the compression spring ensures reliable
closing. Only minimal opening forces have to
be applied here, which results in high dynam-
ics (switching speed) and keeps the coil size
æ UTS0312Y
and inductance low.

Additional requirements in relation to the
accuracy of the characteristics place stricter
Fig. 5 b
demands on precision during the manufac-
Fmag ture of the valve components and during
a Sectional view of
PWM valve their installation than is the case in a pure
1
b Oil flow in valve on/off switching valve.
1 Return to tank
pclutch
Fmag Magnetic force
æ UTS0287E
pfeed
Ffeed Feed-pressure Ffeed
force
pfeed Feed pressure
pclutch Clutch pressure
6 3/2 PWM valve with rising characteristic 7 3/2 PWM valve

(sectional view) (with rising characteristic)
pwork
2
æ UTS0292E
æ STS0293E
Fig. 6
1 Feed from pump
2 Return to tank
pwork Working pressure
In all, PWM valves exhibit the following The ends of the characteristic demonstrate
features: slight discontinuity. This is caused by the
쐌 low costs, transition from the switching to the holding
쐌 non-susceptible to dirt and contamination, operating state (when closed or open). This
쐌 free from hysteresis, inaccuracy is tolerable in these closely limited
쐌 low leakage, and areas.
쐌 simple actuation electronics.
3/2 PWM Valve with High Flow Rate
However, drawbacks are The PWM valve with high flow rate has
쐌 pressure pulsation and essentially the same design as the standard
쐌 dependence of the characteristic on the PWM valve described above, but provides
feed pressure. larger opening cross-sections with a greater di-
ameter and a longer opening stroke of the
Areas of application for the PWM valves with closing element (comparison in Table 1). This
rising characteristic or high flow rate de- design calls for a larger copper winding with a
scribed in the following are: higher magnetic force (Figures 8 and 9).
쐌 control of the torque converter lockup
clutch,
쐌 clutch control, and
쐌 main-pressure control.
3/2 PWM Valve with Rising Characteristic

The pressure-compensated tubular design of 1 Technical data of PWM valves in comparison
the closing element of this PWM valve (Fig- PWM valve type Rising With high
characteristic flow rate
ure 6) results in low inertial forces. Even this
circumstance meets the demand for fast Feed pressure kPa 300…800 400…1200
Flow rate l · rpm > 1.5 > 3.9
switching times, low noise emission, and long (at 550 kPa)
service life. The pronounced linearity of the Timing frequency Hz 40…50 40…50
characteristic (Figure 7) in a broad character- Coil resistance Ω 10 10
istic and temperature range is an important Dimensions
Diameter mm 25 30
advantage for use in the vehicle.
Free length mm 30 42
Table 1
8 3/2 PWM valve with high flow rate 9 3/2 PWM valve with high flow rate
(sectional view) (characteristic)
1000
kPa
800
Working pressure pwork
1
600
pwork
400
200
æ UTS0294E
æ STS0295E
0 Fig. 8
0 20 40 60 80 100 %
1 Feed from pump
Duty factor
2 Return to tank
Pressure Regulator Spool-Type Pressure Regulator PR-S

Two principles are used in the analog valves The spool-type pressure regulator (Figures 11
for pressure control (Figure 10): to 13) operates as a two-notch regulator. The
regulated pressure is tapped between inlet and
쐌 The spool-type pressure regulator opens outlet metering notches, and is produced as a
a metering notch at the feed port and function of their opening ratio. Maximum
simultaneously closes a metering notch pressure is obtained with the inlet opened and
to the tank return line (two-notch regula- the outlet closed, while zero pressure is ob-
tor). The position of the regulator plunger tained under reversed conditions. In between
is derived from the equilibrium of forces, these settings, this pressure regulator changes
depending on the stamped magnetic force, its plunger position in the equilibrium be-
the regulated pressure, and the spring tween force of pressure, spring force, and
force. magnetic force in proportion to the current
쐌 The flat-seat pressure regulator operates as passing through the coil and adjusts the con-
an adjustable pressure relief valve (single- trol pressure accordingly.
notch regulator). The recirculation of the regulated pressure
to the end face of the plunger via an oilway in
With both principles, the regulated pressure the transmission control plate closes the
intervenes directly in the equilibrium of closed control loop (external recirculation).
forces, which is why complete closed control The regulated pressure can also be included by
loops are provided. The following text means of a stepped plunger or other measures
explains the flat seat and spool principles as a resulting force in the equilibrium of forces
by way of examples. and thus in the closed control loop (internal
recirculation).
10 Pressure regulator (basic layout) 11 D30 spool-type pressure regulator (view and oil flow)
Fig. 10
a a
a Spool type pfeed
b Flat-seat type
Fhyd 1
1 Feed from pump
2 Return to tank
Fspring Spring force Fmag
Fhyd Hydraulic-
æ UTS0303Y
pressure force
Fmag Magnetic force Fspring
preg
preg Regulated
pressure to clutch 2
pfeed Feed pressure Fhyd = Fspring+ Fmag
from pump
b pfeed b
1
Fig. 11 Fhyd
pfeed
a Cutaway view of
spool-type 1 Fmag
pressure regulator
b Oil flow in valve
1 Feed from pump
preg
2 Return to tank Fspring preg
æ UTS0296E
æ UTS0297E
preg Regulated 2 2
pressure to clutch
Fhyd = Fspring+ Fmag
pfeed Feed pressure
from pump
The advantages of the spool-type pressure 2 Technical data of spool-type pressure regulator
regulator are: PR-S (typical)
쐌 high accuracy, Feed pressure kPa 700…1600
쐌 non-susceptible to influencing quantities, Regulated pressure kPa typical 600…0
쐌 low temperature sensitivity, Current range mA typical 0…1000
≤ 600
쐌 non-susceptible to system leakage, Activation frequency Hz
Dimensions
쐌 low leakage, and Diameter mm 32
쐌 zero pressure can be achieved. Free length mm 42
Table 2
The disadvantages of the spool-type pressure 12 D30 spool-type pressure regulator (sectional view)
regulator are:
쐌 expensive production of the precision com-
ponents, and
쐌 precision electronics required. pwork
By combining a frictionless mounting 1

(diaphragm spring) with a Teflon-coated
plain bearing, this pressure regulator ensures
minimum hysteresis and optimum accuracy.
The presence of a spool makes the regulator
largely resistant to the influences of system
æ UTS0298E
2 Fig. 12
leakage, fluctuations in feed pressure, or tem- 1 Feed from pump
perature changes. 2 Return from tank
The spool-type pressure regulator shown in pwork Working pressure
Figures 11 and 12, with its typical technical
13 Spool-type pressure regulator (falling characteristic)
data listed in Table 2, exhibits none of the
drawbacks of the flat-seat pressure regulator
discussed below, but is more expensive than kPa
the latter due to its higher-quality individual 200
Regulated pressure pwork
parts (expensive flange, precision-machined

plunger).
As with the flat-seat pressure regulator, 400
there are versions available with rising and
falling characteristics (Figure 13).
200
Flat-Seat Pressure Regulator PR-F

æ STS0299E
Like its spool-type counterpart, the flat-seat 0

pressure regulator is also a proportional valve. 0 0.2 0.4 0.6 0.8 A 1.0
The magnetic force is proportional to the Current I
current passing through the coil. The
hydraulic pressure acts via the sensing surface The flat-seat-type pressure regulator is there-
on the equilibrium of forces. The pressure fore essentially an adjustable pressure limiter
regulator can also be fitted with a compres- with a hydraulic control function.
sion spring to be able to achieve a defined ini- The essential features of the flat-seat
tial state. The regulated pressure is established pressure regulator are:
by a pressure drop caused by oil flowing back 쐌 high accuracy,
to the tank through a varying cross-section. 쐌 inexpensive,
쐌 non-susceptible to influencing quantities,
쐌 non-susceptible to dirt and contamination, precisely matched with the upstream perma-
쐌 high leakage, nent restrictor. Suitable measures can be used
쐌 residual pressure present even to compensate temperature sensitivity to
(dependent on temperature), and a certain extent.
쐌 expensive electronics.
This pressure regulator described by way of
Flat-seat pressure regulator, example has been specially optimized such
falling characteristic that the pressure/current characteristic ex-
Figures 14 to 16 show the example of a D30 hibits low hysteresis and a narrow tolerance
flat-seat pressure regulator with falling range. These properties are achieved by using
characteristic. high-quality magnetic-circuit materials and
The pressure regulator operates in conjunc- modern manufacturing procedures.
tion with a restrictor (diameter 0.8... 1.0 mm) Since the ratio of the hydraulic resistance
which is either arranged externally in the hy- of both restrictors cannot be indiscriminately
draulic-control system or integrated directly small, the pressure/current characteristic of a
in the pressure regulator. The advantage of the typical flat-seat pressure regulator exhibits a
latter type of restrictor lies in the fact that the residual pressure which increases as the tem-
pressure-regulator characteristic (of the vari- perature decreases. The hydraulic-control sys-
able restrictor at the flat seat) can be more tem of the transmission must take this cir-
14 D30 flat-seat pressure regulator, 15 D30 flat-seat pressure regulator,

falling characteristic (view and oil flow) falling characteristic (sectional view)
a 1
æ UTS0365Y
pwork
æ UTS0301E
2
16 Flat-seat pressure regulator (falling characteristic)
Fig. 14
a Cutaway view of b
flat-seat pressure 1
regulator
b Oil flow in valve 2
1 Feed from pump
2 Restrictor
3 Clutch
4 Return to tank
3
4
æ UTS0300Y
æ STS0302E
Fig. 15
1 Feed from pump
2 Return to tank
cumstance into account; accordingly, the ef-

fective range for control does not begin at 0 17 D20 flat-seat pressure regulator, rising characteristic
(cutaway view)
kPa, but rather at a suitably higher pressure.
Another drawback of the flat-seat pres-
sure regulator manifests itself particularly
when several of these pressure regulators are
used in a transmission: The system dictates
that there is a permanent return oil flow
through the opened pressure regulator to
the oil pan, which results in energy losses
and under certain circumstances may re-
quire the use of a transmission-oil pump
æ UTS0314Y
with increased volumetric flow. These draw-
backs can be avoided by increasing the level
of sophistication in the actuator or in the
hydraulic-control system. However, this
cancels out to some extent the main benefits
of the flat-seat pressure regulator (simple
design and low costs). The options offered
by the closed-end function are discussed in
the following. 18 D20 flat-seat pressure regulator, rising characteristic
(sectional view)
Flat-seat pressure regulator, rising

characteristic (miniature version) 1
Thanks to the use of high-precision plastic
technology, the D20 flat-seat pressure regu-
lator (Figures 17 to 19) has undergone a
significant reduction in both size and cost.
With a diameter of little more than
pwork
20 mm, it achieves a high level of character-
æ UTS0304E
Fig. 18
istic precision while taking up the minimal 2 1 Feed from pump
amount of space. 2 Return to tank
3 Technical data of flat-seat pressure regulators 19 D20 flat-seat pressure regulator

in comparison (rising characteristic)
D30 D20
Pressure-regulator type 600
falling rising kPa
characteristic characteristic
Regulated pressure pwork
Feed pressure kPa 500…800 500…800

Regulated pressure kPa 40...540 40...540 400
typical
Current range mA 150…770 150…770
typical
200
Activation frequency Hz 600...1000
Chopper frequency Hz 600...1000
æ STS0305E
Dimensions
0
Diameter mm 30 23 0 0.2 0.4 0.6 A 0.8
Free length mm 33 42
Current I
Table 3
Deciding on the Pressure-Regulator Type 쐌 installation position (horizontal, vertical),

The decision as to whether a flat-seat pressure 쐌 installation location (above or below oil
regulator, a spool-type pressure regulator, or level or varying),
even a PWM valve is used in the transmission 쐌 adjustability of the closed control loop
depends on a whole range of considerations. (end of line programming), and
Some technical criteria have already been dis- 쐌 dirt concentration and dirt composition.
cussed in the descriptions of the individual
types. However, selection criteria such as the
These criteria are again set out and com- following are not to be underestimated:
pared in Table 4. The figures in this table rep- 쐌 The user’s own experience with a particular
resenting accuracy are not absolute, but only type (familiarity level, risk).
relative to each other. They are only 쐌 In-house user traditions which can also
approximate values and the conditions may lead to a wealth of single-sided experiences.
differ completely according to the vehicle 쐌 Existing control-system concepts, for which
type used, for example depending on the reworking expense in changing the reg-
constructive details such as ulator type is deemed to be too great.
쐌 variability of the equilibrium of forces, 쐌 A cost analysis is conducted partly from a
쐌 variability of the magnetic force, standpoint of reducing outsourcing costs.
쐌 use of special tools in the magnetic circuit, An analysis of the overall costs including,
쐌 gap between feed pressure and maximum for example, the expenditure in internal
control pressure, production on control-plate machining is
difficult for the user (existing equipment
and system-conditioned criteria such as . . .).
쐌 type of control,
쐌 damping properties of the follow-up
system,
4 Criteria for actuator selection

Criterion Spool-type pressure Flat-seat pressure 3/2 PWM valve
regulator PR-S regulator PR-F
Sophistication in non-susceptible to Constant feed pressure Constant feed pressure

hydraulic system fluctuations in
feed pressure Feed restrictor Damping
Accuracy:
Comparison value ≈ 7% (recirculation) ≈ 11 % ≈ 13% (control)
(manufacturing ±5...±25 kPa ±5...±30 kPa ±20 kPa
tolerance) (dependent on (dependent on (constant)
characteristic range) characteristic range)
Influence of feed pressure pC = 400 ±0.2 kPa ΔpC ≈ 0.2 · Δpfeed ΔpC ~ Δpfeed
at pfeed = 800 ±50 kPa
Leakage typ. 0.3 l · rpm 0.3...1.0 (...0) l · rpm 0...0.5...0 l · rpm

(without elasticity)
Noise – – If necessary
damping required
Costs High Medium Low

Table 4
Electrohydraulic Actuators Simulations in Development 313
Simulations in Development Computer-aided design of magnetic and

hydraulic circuits thus forms the basis for
Requirements simulation of the static and dynamic proper-
The development times (time to market) for ties of the actuators in the transmission sys-
new transmission generations are getting ever tem. This simulation can be used to optimize
shorter. Testing-compatible actuators must be functions and analyze properties under limit
available with function data specific to the conditions.
transmissions virtually immediately after the
start of any project. The strict demands on Function Simulation
quality and reliability necessitate comprehen- 1D simulation enables, for example, the prop-
sive tests and trials with prototypes. The pre- erties of pressure-control valves to be simu-
viously standard iterative method using related in their system environment. These in-
cursions and modifications will in future not clude pressure/current characteristics, tem-
be practicable from time and cost stand- perature sensitivity, dynamics, etc., under
points. different boundary conditions or geometric
In future, function prognoses and analyses of variables.
conspicuous features in the product-creation
process using virtual prototypes will increas- Function simulation of the actuator in the
ingly take place at an earlier stage and will at (sub)system is performed for example for the
least support experimental development. variables:
쐌 characteristic,
1 Simulation of flow characteristic and pressure 쐌 dynamics,
distribution in a PR-F pressure regulator
쐌 temperature influence,
쐌 influence of feed-pressure fluctuations, and
쐌 worst-case studies, manufacturing
tolerances.
b
Flow Simulation
The Fire tool used by Bosch to calculate
hydraulic losses and flow forces enables,
0…21.7 m·s-1
for example, the geometries of the hydraulic
section of a pressure-control valve to be opti-
mized is such a way that flow influences on
the pressure/current characteristic are mini-
mized under all operating conditions and
a flow cross-sections are optimized.
By way of example, Figure 1 shows the simu-

lated flow characteristic and the pressure dis-
tribution in a flat-seat pressure regulator.
Fig. 1
æ STS0306E
55.5…350 kPa a Pressure regulator

(hydraulic section)
b Flow characteristic
c Pressure distribution
314 Electrohydraulic Actuators Simulations in Development
Magnetic-Circuit Calculation Basic Principles of 1D Simulation

A finite-element program (such as Motion Equation with Force Balance
MAXWELL 2D or Edison) is used to design The motion equation with force balance is as
the magnetic circuit and thus the pressure/ follows:
current characteristic of a pressure-control
valve. But it can also be used to utilize to Ftot = ∑(FM + Fp + FFl + FFr + FD + FS) = 0
optimum effect the available space (size
reduction) and the material of a magnetic cir- With the variables
cuit (magnetic-force increase, Figure 2) or to FM magnetic force,
adapt the force/travel characteristic to the re- Fp pressure force,
quirement. FFl flow force,
FFr frictional force,
The magnetic force is simulated, FD damping force, and
for example, for: FS spring force
쐌 dimensioning,
쐌 the layout for the magnetic-force or according to Figure 3:
characteristic,
쐌 eddy-current losses, and Ftot = mx·· + dx· + cx
쐌 manufacturing tolerances.
With the variables
The following improvements have been m moving mass,
implemented as the result of the simulation d damping,
(example in Figure 2): x·· acceleration,
쐌 position and shape of the working air gap, x· velocity,
쐌 optimized cross-section for a stronger x travel,
magnetic flux, and c spring rate.
쐌 reconfiguration of the working air gap and
parasitic air gap (geometry, position).
2 Optimization of magnetic force with MAXWELL 2D 3 “Force balance” model

(example)
a b
Fp FFr
FM x ˙ x
x
¯ d
FFl
c
æ STS0308E
æ STS0307E
Fig. 2
a Basic model
FM = 5.8 N FM = 10.2 N
b Optimization
FM Magnetic force
Electrohydraulic Actuators Simulations in Development 315
Pressure Calculation in a Chamber 4 “Chamber pressure” model

Pressure calculation in a chamber is governed
by:
dp β
= oil · Q
dt Voil
dp
QE ΣQ β dt p
Where:
Vol
βoil compression module of pressurized
fluid, QA
Voil chamber volume,
Q cumulative volumetric flow.
æ STS0309Y
The cumulative volumetric flow flowing into
the chamber, which is responsible for the pres-
sure build-up, is calculated from (Figure 4):
5 “Restrictor flow” model
Q = (QE – QA)
Where:
Q cumulative volumetric flow,
QA outlet volumetric flow,
QE inlet volumetric flow. Q
Flow-Rate Calculation at a Restrictor

The volumetric flow at a restrictor is 1 0 2 3
calculated from (Figure 5):
æ STS0310Y
2苶
Q = αf · Ao · 冪苶· Δp
ρ
;
where
Δp = p1 – p3
And where:
αf flow-rate coefficient,
Ao cross-sectional area of a restrictor,
ρ density of medium,
Δp pressure differential,
p1 pressure at point 1,
p3 pressure at point 3.
316 Electronic Transmission Control Drivetrain Management
Electronic Transmission Control

In complicated traffic situations, unfamiliar drivetrain-management system calculates the
surroundings, or poor weather conditions conversion into torque and engine speed and
(e.g., heavy rain, snow, or fog), manual implements them. In order for such a strategy
gear changing can distract car drivers to such to be implemented, it is essential for the sys-
an extent as to create situations that are diffi- tem to be equipped with an electrically actu-
cult to control. This also applies to the annoy- ated throttle valve (drive by wire).
ing, incessant process of engaging and disen-
gaging the clutch when driving in Figure 1 shows the organizational structure of
stop-and-go traffic. Automatic transmissions drivetrain management as part of the overall
with electronic control assist drivers in these vehicle structure. The vehicle coordinator for-
and other traffic situations so that they can wards the requested propulsion movement to
concentrate fully on the road conditions and the drivetrain coordinator while taking into
what is happening around them. account the power requirements of other ve-
hicle subsystems (e.g., body or electrical-sys-
Drivetrain Management tem electronics). The drivetrain coordinator
distributes the power demand to the engine,
As the number of electronic systems in the converter, and transmission. In the process,
vehicle increases, so too does the complexity the various coordinators may also have to
of the overall network of the various ECUs. solve conflicts of interest that arise in accor-
Controlling such networked structures re- dance with defined priority criteria. A whole
quires hierarchical order concepts, such as, range of different external influencing factors
for example, the “Cartronic” system from (such as environment, traffic situation, vehi-
Bosch. Coordinated drivetrain manage- cle operating status, and driver type) plays a
ment is integrated as a substructure in the role here.
“Cartronic” system. It facilitates optimally
matched management of the engine and The Cartronic concept is based on an object-
transmission in the various vehicle operating oriented software structure with physical in-
conditions. terfaces, e.g., torque as an interface parameter
The engine is operated as often as possible of drivetrain management.
in the fuel-saving ranges of its program map.
If the driver adopts a sporty driving style,
however, the high, less economical speed
ranges are increasingly utilized. Such a situa-
tion-dependent mode of
operation presupposes on 1 Monitoring architecture of monitored drivetrain management
the one hand that the dri-
ver command is recog-
nized and on the other Vehicle Drive
hand that its implementa- Environmental
tion is left to the elec- ? Torque Engine
variables
Coordinator
tronic drivetrain-manage- Driving-condition Vehicle

Coordinator
? Slip Converter
ment system and a variables motion
higher-level driving strat- Vehicle Body and Trans-
? Gear ratio
egy. When the driver variable interior mission
presses the accelerator User Electrical Provision of propulsion
variables ? system
pedal, the system inter-
æ STS0256E
power and power for other

loads/consumers
prets this action as an
acceleration request.
From this request, the

Electronic Transmission Control Market Trends 317
Market Trends CAFE Requirements

In contrast to Europe, the USA, the most im-
The statutory requirements relating to fuel portant market for automatic transmissions,
consumption and exhaust-gas emissions will has seen no change in the CAFE fuel-con-
play a significant role in transmission devel- sumption requirements (corporate average
opment over the next few years. There follows fuel efficiency) since 1990 (Figure 2). All at-
a brief comparison of the requirements of the tempts to bring about a tightening of these
main markets for this purpose. requirements have proven unsuccessful.
ACEA, JAMA and KAMA

The ACEA (Association des Constructeurs
Européens d’Automobiles, i.e. Association of
European Automobile Manufacturers) has
agreed to reduce the corporate average in 1 CO2-emission requirements
CO2 emissions in the period from 2002 to

2008 from 170 mg CO2 to 140 mg CO2
g/km
(Figure 1). 200
Current field
180 of values
The Japanese and Korean manufacturers’ as-
CO2 emission
165…170
sociations (JAMA and KAMA) have adopted 160
the same limits for the year 2009. In order for 140
this target to be achieved, the next few years 140
will see an increased acceptance of transmis- 120

Consumption leader
VW Polo 3l
sion types such as the 6-speed transmission
CVT (continuously variable transmission) 100
2000 2002 2004 2006 2008 2010 2012
and AST (automated shift transmission).
æ STS0257E
Year
2 CAFE fuel-consumption requirements (passenger cars, light commercial vehicles not included)
Standard Current value

FE/mpg
30
25
Fuel consumption
20
15
10
5
æ STS0258E
0
1978 1990 2000
Year
318 Electronic Transmission Control Control of Automated Shift Transmission AST
Control of Automated Shift the clutch-release system. Instead the clutch-

travel sensor is integrated in the electric-mo-
Transmission AST tor actuator.
Requirements The small electric motor offers a low power
Current market developments reveal a marked density in comparison with a hydraulic pump
trend towards an increase in in-car safety and and accumulator. The electric-motor actuator
operating comfort and convenience. This is is therefore only suitable if it can achieve suf-
accompanied by an increase in vehicle mass ficiently short declutching times for rapid
and in the final analysis an increase in fuel gearshifts.
consumption. The emission guidelines laid
down by legislators (140 g/km CO2 by 2008) Shifting without Torque Correction
only intensify the situation. In the conventional system without torque
correction (Figure 1), the clutch torque is far
The automated shift transmission (AST) in excess of the engine torque. The reason be-
combines the advantages of a manually hind this is that the dry clutch, which has to
shifted transmission with the functions of transfer at least the engine torque under all
an automatic transmission. The automated extreme conditions, normally offers a reserve
version of the classic manual transmission of 50 to 150%. The engine torque drops when
is characterized by its high efficiency. Slip the driver wishes to change gear and at the
losses do not occur as in conventional torque- same time takes his foot off the accelerator
converter transmissions. pedal. Operating the shift lever initiates the
Specific fuel consumption in the automatic intention to change gear, and the clutch must
mode is below the low level of the manual now be moved from the fully closed to the
transmission. fully open position. This defines the declutch-
AST development is founded on the ing time.
knowledge and findings gained in connection
with electric-motor clutch management If the declutching time is too long, the clutch
(ECM). will still transfer torque while the next gear is
being synchronized, a process which may re-
Electric-Motor Clutch Management sult in rattling or damage to the transmission.
(ECM)
Application
Following initial experiences with hydraulic
clutch management, users are now concen-
trating their efforts on the use of electric mo- 1 Shifting without torque correction
tors as clutch actuators in the small-car seg- S1

ment. This new approach allows for savings
to be made on costs and weight, as well as Nm
tD
providing a higher degree of integration. 400
MC
Corresponding ECM systems are used in the 300
Mercedes A-Class, the Fiat Seicento, and the
Torque
Hyundai Atoz. 200

ME
100
Design and Operating Concept
Fig. 1 0
The most important step in minimizing costs
MC Clutch torque
was switching from a hydraulic actuator to an ¯100
æ STS0259E
ME Engine torque 0 2 4 6 8s
tD Declutching time electric-motor actuator. This step removed
Time t
S1 Signal for gearshift the need for a pump, an accumulator, and
command valves as well as the need for a travel sensor in
Electronic Transmission Control Control of Automated Shift Transmission AST 319
Shifting with Torque Correction Design and Operating Concept

The most technically sophisticated solution The electric-motor AST features the auto-
for avoiding transmission damage is to mated clutch operation of the ECM system.
combine a reduced-force clutch with torque With an additional electric-motor actuator
correction. for the transmission, the driver is able to
Figure 2 shows the shifting operation with change gear without there being a mechanical
torque correction, in which the clutch torque connection between the selector lever and the
is only marginally above the engine torque. transmission (shift by wire).
When the driver takes his foot off the acceler-
ator pedal to change gear, the clutch torque In the case of the AST, all modifications to the
drops as well as the engine torque. When the transmission are to be avoided. This will en-
intention to change gear is initiated, the able the transmission manufacturer to mount
clutch is thus already almost open and the re- either manual transmissions or ASTs on the
maining declutching operation follows very production line. As the hardware for this
rapidly. system (e.g., for the Opel Corsa Easytronic),
Figure 3 features a schematic representa- Bosch supplies the electric motors for clutch
tion of electric-motor clutch management engagement, shifting, and selection (see chap-
(ECM) as a partial-automation solution, and ter entitled “Transmission, Automated Shift
the automated shift transmission (AST) as Transmission”), and the ECU. Automated and
the complete automation of the manual cost-effective mass production is made possi-
transmission, both as add-on systems. ble by the use of standard components in all
AST applications.
Electric-Motor Automated Shift
Transmission AST 3 ECM and AST as add-on systems
Application a
Today the AST is used primarily in the lower 1 2 3
torque classes (e.g., VW Lupo, MCC Smart,
Opel Corsa Easytronic, see also chapter enti- Fig. 2
tled “Transmission Types”), where, in com- MC Clutch torque
parison to the fully automatic transmission, ME Engine torque
the cost benefit makes up for the downside of tD Declutching time
S1 Signal for gearshift
the interruption in tractive force.
command
4
2 Shifting with torque correction

Fig. 3
ECM a ECM
S1
b AST
b
tD
Nm AST 1 Available signals
400
1 5 6 2 Clutch actuator with
300 integrated ECM
ECU
Torque
200
MC 3 Gear recognition
100 4 Shift-intention
ME recognition on
0 shift lever
5 Clutch actuator with
¯100
æ STS0260E
æ STS0261E
0 2 4 6 8s integrated AST ECU

7
Time t 6 Transmission
actuator
7 Selector lever
320 Electronic Transmission Control Control of Automated Shift Transmission AST
Software Sharing Shifting Operation and Interruption of

The vehicle manufacturer (OEM), the sup- Tractive Force
plier and if necessary a system integrator The basic problem associated with the AST
share the AST software tasks. The operating is the interruption of tractive force. This is
system, signal conditioning, and the hard- represented in Figure 4 by the “hollow” of the
ware-specific routines for activating the vehicle acceleration between the two shifted
actuators are provided by Bosch. Bosch’s gears. In terms of what is required of the
extensive knowledge and experience in the actuators, these phases can be divided into
field of automatic transmissions is also two blocks:
applied in establishing the AST target gears. 쐌 phases which have an effect on the vehicle
This includes, among others, driver recogni- acceleration,
tion, uphill/downhill recognition, cornering 쐌 phases which represent pure response
recognition, and other adaptive functions times.
(see also chapter entitled “Adaptive Transmis-
sion Control, ATC”). In the phases which have an effect on the
vehicle acceleration, it transpires that a throt-
The tasks of activating the transmission and tle action is needed because excessively quick
coordinating the gearshift sequence (clutch, changes in vehicle acceleration are felt to be
engine, transmission) are the responsibility of unpleasant. The optimum interaction of en-
the OEM or the system integrator. gine, clutch, and transmission intervention
results in the best possible performance.
This also applies to clutch control, significant
parts of which can be taken over from the The synchromesh can be supported for
ECM system. Each vehicle manufacturer example by double-declutching. In the
brings its marque-specific philosophy response times, however, the maximum speed
regarding shifting time, shifting points, of the actuators is demanded. It is important
and shifting sequences to bear. here that the synchromesh does not experience
too hard a shock after the gear has been disen-
gaged and the following rapid phase.
4 Phases of AST shifting operation
t1
t2
Fig. 4 1 t0
Acceleration a
1 Current gear
2 Next gear
ΔM1 Torque reduction ΔM1 2

ΔM2 Torque increase
t0 Tractive-force t3 t4 t5
interruption ΔM2
t1 Shifting operation
t2 Acceleration
æ STS0262E
t3 Disengage and
select gear
t4 Synchronization Time t
t5 Shift through gear
Electronic Transmission Control Control of Automated Shift Transmission AST 321
Figure 5 shows a comparison of the shifting trol strategy such as torque correction or by
times that can be achieved with a hydraulic the interaction of the engine and the clutch.
system and an electric-motor system and the
shifting time necessary for a comfortable It is important to highlight here that there
shifting operation. The bar lengths equate to is hardly any difference between the two
the times required for the individual phases systems in the phases for response time, gear
and the same shading schemes are used. disengagement, and gear engagement. The
response times are not extended practically,
When the capacity of the actuators is even in the case of comfortable shifting.
exploited to maximum effect, the electric- However, the phases relevant to acceleration
motor system only demonstrates a time must be two to four times as long as in the
disadvantage compared with its hydraulic extreme case, both for the hydraulic and the
counterpart in the clutch-operation phase. electric-motor actuator systems.
This could be reduced in particular in the
torque-reduction phase by an intelligent con-
5 Comparison of achievable shifting times
100 ms
Maximum
hydraulic force
Maximum
electronic force
Comfortable
æ STS0263E
Torque Gear Synchro- Gear Torque

reduction out nization in increase Fig. 6
1 Engine electronics
(EDC)
2 Transmission
6 Automated shift transmission (AST) in a diesel vehicle (example: system diagram)
electronics
3 Transmission
1 8 2 9 3 11 10 actuator
4 Diesel engine
5 Dry interrupting
clutch
CAN 6 Clutch servo unit
7 Intarder electronics
8 Display
9 Driving switch
(selector lever)
10 ABS/TCS
11 Transmission
æ UTS0207-1Y
12 Air supply
___ Electrics
4 5 6 12 7
---- Pneumatics
___ CAN communication
322 Electronic Transmission Control Control of Automatic Transmissions
Control of Automatic Hydraulic Control

The main function of the hydraulic-control
Transmissions system (Figures 1 and 2) is to regulate, boost,
Requirements and distribute hydraulic pressures and volu-
The control system for an automatic trans- metric flows. This includes generating the
mission must fulfill the following essential clutch pressures, supplying the converter, and
requirements or functions: providing the lubricating pressure. The hous-
쐌 always shifting the correct gear or setting ings of the hydraulic-control system are made
the correct gear ratio as a function of from diecast aluminum and contain several
assorted influencing variables, precision-machined slide valves and electro-
쐌 executing the shifting operation through hydraulic actuators.
adapted pressure characteristics as com-
fortably as possible, 2 Main control with hydraulic valves
쐌 implementing additional manual interven-

tions on the part of the driver,
쐌 detecting maloperations, e.g. by preventing
non-permitted gearshifts,
쐌 providing ATF oil for cooling, lubrication
and for the converter.
Current control systems are exclusively

electrohydraulic in nature.
æ UTS0265Y
1 Exploded view of a hydraulic-control system (example: GM HYDRA-MATIC 4L60-E)
æ UTS0264Y
Electronic Transmission Control Control of Automatic Transmissions 323
Electrohydraulic Control Direct Control

Due to their extensive range of functions, all With direct control, the required pressure and
modern automatic transmissions with four to throughflow for rapid clutch filling are pro-
six gears and continuously variable transmis- vided directly by the actuator (Figure 4).
sions are exclusively controlled by electrohy- This results in a compact clutch-control sys-
draulic means. In contrast to earlier, purely tem with reduced hydraulic sophistication.
hydraulic control systems with mechanical
regulators, the clutches are activated individ- Shifting-Sequence Control
ually by pressure regulators, which facilitate Conventional Shifting-Sequence control
precise modulation and regulated overlap- The following two shifting scenarios are
ping gearshifts (without a one-way clutch). examples of conventional control of a simple
4-speed automatic transmission with one-
Clutch Control way clutches (Figure 5).
Clutch control is always performed with
either pilot-controlled or directly controlled Upshift Under Load
pressure. Unlike in a manual transmission, throttling
upshifts in an automatic transmission take
Pilot Control place without an interruption of tractive
With pilot control, the required pressure force. The graphic in Figure 6 shows the time
and throughflow for rapid clutch filling are curve of the characteristic variables during an
provided via a slide valve in the control hous- upshift into the direct gear (ratio 1). Shifting
ing. Pressure regulation is effected by pilot begins at time t0: The clutch is filled with
pressure acting on the sensing surface of the fluid and the friction elements are pressed
slide valve. An actuator generates this pilot against each other. The clutch transfers a
pressure (Figure 3). torque from time t1 onwards. As the clutch
This results in greater degrees of freedom torque increases, so the torque supported at
in the packaging and the use of standardized the one-way clutch decreases. The one-way
actuators, high dynamics, and small electro- clutch is released at time t2. Now the engine
magnets. speed begins to change. The clutch torque in-
creases to t3. The clutch slips up to t4, after Fig. 3
1 Supply to actuator
3 Pilot control (pictorial diagram) 5 Load transfer with one-way clutch US 2 Oil pan
3 Supply to slide valve
4 Actuator
5 Slide valve in control
1 2 3
Torque M
housing
M 6 Clutch
æ STS0266Y
4 5 6
p1 Fig. 4
1 Supply to actuator
Pressure p
2 Actuator
4 Direct control (pictorial diagram) 3 Clutch
p2
Fig. 5
1 nE p1 Pressure,
cutting-in clutch
Speed n
æ STS0268E
æ STS0267Y
p2 Pressure,
2 3 cutting-out clutch
Time t nE Engine speed
M Torque
which it sticks. After the end of the shifting Downshift Under Load
operation, the clutch pressure is controlled In contrast to upshifts, downshifts take place
upward to a level of safety. with an interruption of tractive force.
The speed difference between the engine Figure 7 shows the time progression.
and the transmission output speed remain- At t0 shifting begins with the clutch being
ing after t4 is caused by the converter, which drained. Engine torque is no longer trans-
Fig. 6
always operates with slip when locked up. ferred from t1 onwards and the engine revs
pC Clutch pressure up. At t2 the synchronization speed of the
pF Filling pressure The progression of the output torque in the new gear is reached and the one-way clutch
pS Safety pressure phase t1...t4 determines the gear-shift so- is engaged; up to t3 the converter slip is set
t0 Start of shift phistication (ease of shifting). To ensure according to the engine torque. The shifting
t1 Start of torque good shift quality, the clutch pressure must operation is completed at t3. The gear-shift
transfer, clutch
be set so that the output torque is between sophistication (ease of shifting) is deter-
torque rises,
one-way clutch
the level at t < t1 and the level at t > t4. The mined by the torque drop in the phase t0...t1
torque drops torque jump at t4 should also be as low as and depends quite significantly on the
t2 One-way clutch possible. torque increase between t2 and t2.
released The load on the friction elements is de- All the shifting scenarios that occur are
t3 Clutch slips, termined by the clutch torque and the slip primarily controlled by the electronic sys-
clutch torque
time (t4 – t1). It is clear here that controlling tem; the hydraulic system is left above all
remains constant
t4 Clutch sticks, clutch
the shifting sequence always involves a com- with the function of clutch power control.
torque decreases, promise. In all newer transmissions (5-speed
converter operates and 6-speed types), one-way clutches are
with slip replaced by regular clutches for weight
reasons. However, during the shifting
1 Output operations, they require overlap control
2 One-way clutch
3 Clutch
4 Engine
6 Pressure, torque and speed curve during an upshift 7 Time curve of a downshift
5 Transmission output
1
1
Torque M
2
Torque M
Fig. 7 2
pC Clutch pressure
3
pF Filling pressure 3
t0 Start of shift,
clutch drains
t1 End of torque
transfer, engine
revs up pS
t2 Synchronization
Pressure p
speed of new gear pC

pF pC pF
Pressure p
reached, one-way
clutch engaged,
converter operates
with slip
t3 Shifting operation
completed
4
4
Speed n
Speed n
1 Output 5 5
æ STS0269E
æ STS0270E
2 One-way clutch
3 Clutch t0 t1 t2 t3
t0 t1 t2 t3 t4
4 Engine Time t
Time t
5 Transmission output
for the clutches (Figure 8). This means that 8 Overlap control US
while clutch 1 opens for gear x, clutch 2 must
close for gear y. Since this type of control is
very elaborate and time-critical, it is necessary M
Torque M
to provide considerably higher computing
power in the ECU than for the simple shifting
sequences with one-way clutch shifts (see also
chapter entitled “ECUs”). p1
Pressure p
The most important features of overlap
control are: p2
쐌 low mechanical complexity,
쐌 minimal space requirement,
쐌 multiple use for different gear steps Fig. 8
nE p1 Pressure,
possible,
Speed n
cutting-in clutch
쐌 high control precision for load transfer
æ STS0271E
p2 Pressure,
required, cutting-out clutch
쐌 high software complexity for torque Time t nE Engine speed
control, M Torque
쐌 in event of incorrect control: excessive
speed (engine races) or onset of a braking 9 Fill time
torque (extreme case: transmission block-

ing).
Adaptive Pressure Control Δntu

Turbine speed ntu
The function of adaptive pressure control is

to achieve a consistently good shift quality
over the entire working life of the transmis-
ntu
sion and the accompanying changes in the
friction coefficients at the clutch surfaces. tfill
It also compensates for any potential devia-
æ STS0272E
tshift tvertex
tion of the calculated torque or the torque
Time t
transferred by the engine which can occur
on account of changes to the engine or
manufacturing tolerances.
In this case, an important role is played by Fill-Time Measurement
pressure adaptation with the aid of the shift- The fill time tfill (Figure 9) is the time from
ing times applied by the manufacturer. To this the start of the gearshift tshift to the start of
end, the applied shifting times are compared synchronization (a drop in speed is identified
with the real shifting times that occur. If the during the upshift [US]):
measurements are repeatedly outside a pre-
specified tolerance range, the pressure para- tfill = tvertex – tshift
meters pertaining to the shifting operation
are incrementally adapted. A distinction is Slip-Time Measurement
made here between the fill time and the slip The slip time tslip (Figure 10) of the clutch is
time of the clutch. the time from recognition of the speed vertex
(start of synchronization) to complete syn-
chronization of the speed in the new gear.
10 Slip time Shifting-Point Selection

Conventional Shifting-Point Selection
In the majority of automatic transmissions
tslip
currently available, the driving program is
selected using a selector switch or a button.
Turbine speed ntu
The following driving programs are generally

available in this respect:
ntu 쐌 Economy (very economical),
쐌 Sport, or
Δntu 쐌 Winter.
(sync)
æ STS0273E
tvertex tsync The individual programs differ in the posi-
Time t tion of the shifting points in relation to the
position of the accelerator pedal and the dri-
tslip = tsync – tvertex ving speed. The Economy and Sport shifting
maps of a 5-speed transmission are used here
The speed thresholds used for measuring the as examples (Figure 11).
slip time tslip (Figure 10) are calculated in
advance of the start of shifting, where the If the current driving speed or the accelera-
following relationship applies to upshifts: tor-pedal position corresponding to the
Start of fill-time measurement = start of driver command (accelerator-pedal value) in-
shifting tersects the shift curve, a gearshift is triggered.
A requested gearshift can be either canceled
Vertex: A decrease in the turbine speed ntu by or converted into a double shift within a spe-
at least ntu (vertex) revolutions is detected. cific period of time (which depends on the
ntu (t – 1) – ntu (t) > ntu (diff) hydraulic system of the automatic transmis-
sion)
Synchronization speed: An increase in the tur- For example, the driver is driving in fifth
bine speed ntu by at least ntu (sync) revolu- gear on an interstate highway and would like
tions is detected. to overtake. To do so, he presses the accelera-
ntu (t) – ntu (t – 1) > Δntu (sync) tor pedal to the floor, whereupon a downshift
is requested.
Pressure Correction
Pressure adaptation is only permitted within 11 US and DS characteristics in Economy mode (XE)
and Sport mode (XS)
specific limits on account of operational reli-
ability. The typical adaptation width lies in 2-1 RS 1-2 HS
the range of ±10% of the modulation pres- % XE XS XE XS
sure calculated for the shift. The correction 100
values are also still distinguished according

to speed bands.
The adaptation values are stored in a
non-volatile memory so that the optimum 50 1
modulation pressure can be reapplied when
the vehicle is restarted. The overall pattern
of pressure adaptation can also be evaluated
as a sign of changes in the transmission.
æ STS0274E
Fig. 11 0
0 50 km/h
1 Upshift
XE Economy mode Vehicle speed υF
XS Sport mode
When the accelerator pedal is firmly pressed Uphill Driving

to the floor, the 4-3 DS shift curve is inter- Recognition of uphill driving by comparing
sected directly after the 5-4 DS shift curve, the current acceleration with the requested
and a 5-3 double downshift is performed acceleration by way of the engine torque,
instead of a sequential downshift. Special results in upshifts and downshifts at higher
shifting points for kickdown (forced down- engine speeds and thus prevents gearshift
shift) allow the maximum possible engine hunting.
power to be utilized at this point.
Cornering
Adaptive Transmission Control (ATC) This facility uses the difference in wheel
All newer transmission-control systems have – speeds to calculate whether the vehicle is in a
instead of active driving-program selection by curve or bend. With active cornering recogni-
the driver – software which enables the driver tion, requested shifts are delayed or prohibited
to adapt to the special ambient conditions in order to increase vehicle stability.
while driving. This includes first and foremost
driver-type recognition and driving-situation Winter Recognition
recognition. Examples which are currently in Winter operation is recognized on the basis of
use are adaptive transmission control (ATC) slip detection from analysis of the wheel
from BMW and the dynamic shift program speeds. This serves primarily to
(DSP) from Audi. 쐌 prevent the wheels from spinning and
쐌 select a higher gear during starting so that
Driver-Type Recognition less torque is transferred to the drive
A driver type can be identified by means of wheels, thereby preventing premature
an evaluation of the actions he or she wheel spin.
performs. This includes:
쐌 kickdown operation, ASC Operation
쐌 brake operation, and If the system detects while driving that the
쐌 restriction via selector lever. ASC ECU (anti-slipping control or traction
control system, TCS) is in control mode,
For example, the kickdown evaluator counts requested gearshifts are suppressed in order
the number of times the driver engages kick- to support the ASC function.
down during a presettable period of time. If
the counter exceeds a specific threshold, the
driver-type recognition facility selects the next,
more sporty driving program. It automatically
switches back to a more economical driving
program once this time has elapsed.
Driving-Situation Recognition
For driving-situation recognition, different
transmission-control input variables are
linked to conclusions about the present
driving condition. The following situations
can generally be recognized:
쐌 uphill driving,
쐌 cornering,
쐌 winter operation, and
쐌 ASC operation.
Engine Intervention Requirements

Application The ever-increasing demand for more eco-
A precisely controlled time characteristic of nomical fuel consumption in motor vehicles
engine torque during the shifting operations dictates to a significant degree the develop-
of an automatic transmission offers the possi- ment objectives in the field of automatic
bility of optimizing transmission control with transmissions as well. In addition to measures
regard to gear-shift sophistication (conve- for improving the efficiency of the transmis-
nience), clutch service life, and transferrable sion itself (such as, for instance, the torque
power. The engine management system im- converter lockup clutch), these objectives
plements the torque command (reduction) include introducing transmissions with more
of the transmission control by retarding the gears. However, additional gear steps in-
moment of ignition. evitably call for increased shift frequency.
The theoretical principles, processes, and This in turn results in increased demands
measurement results are presented using the placed on gear-shift sophistication (conve-
example of engine intervention in ignition. nience) and the load capacity of the friction
elements.
Symbols and Abbreviations Engine intervention takes into account
C Spring stiffness of drivetrain both requirements and institutes an addi-
i Gear ratio tional degree of freedom for controlling an
J Mass moment of inertia automatic transmission. “Engine interven-
k Constant tion” covers all those measures which allow
M Engine torque the engine torque generated by the combus-
n Rotational speed tion process during the shifting operation
q Specific lost work in the transmission to be specifically influ-
Q Lost work enced and in particular reduced. Engine
t Time intervention can be used in both upshifts and
W Running resistance downshifts.
x Spatial coordinate The primary aim of engine intervention
δ Temperature in upshifts is to reduce the lost energy that oc-
ω Angular velocity curs in the friction elements during the shift-
Φ Angle of rotation ing operation. This is done by reducing the
φ Angle of rotation, linearized engine torque during synchronization without
interrupting the tractive force. The margin
Indices acquired in this process can be used to:
O Output 쐌 Increase the service life by shortening the
V Vehicle slip time (if all other operating parameters
limit Permitted limit value in the transmission, such as clutch pressure
C Clutch (friction element) and number of plates, remain unchanged).
kin Kinetic share 쐌 Improve the convenience by reducing the
E Engine (transmission input) clutch torque, brought about by lowering
red Reduced value the clutch pressure during the slip phase.
s Slip time 쐌 Transmit higher power, provided the
com Share of combustion energy mechanical strength of the transmission
(engine torque) permits this; in most cases, however, the
쐌 Reference variable power loss in the clutches is the limiting
1 Clutch drive side factor.
2 Clutch output side
Naturally, it is also possible to adopt a sensible Δω = ω1 – ω2 (1)

combination of these measures within the
framework of the specified margin. In relation to the lost energy which must
be absorbed or forwarded by the friction
The aim of engine intervention in downshifts elements during the shifting operation, the
is to reduce the jolt which occurs when the following equation applies:
one-way clutch or a friction element engages ts
at the end of the synchronization processes. Q = 冮 MC (t) · Δω(t) · dt (2)
This results in 0
쐌 improved convenience and
쐌 supported and improved synchronization Furthermore, the angular-momentum princi-
in transmissions without one-way clutches. ple applies to the drive and output sides of
the clutch. For the rotational masses of the
Interventions in the Mechanical Shifting drive side:
Sequence
1–i M – MC
The following explanations illustrate which Δω = ωO · + E ·t (3)
i JO
possibilities present themselves for interven-
tion in the mechanical shifting sequence. The
individual phases of upshifts and downshifts Under the above-mentioned preconditions,
are described in the section entitled “Shifting- this produces:
Sequence Control”.
Δω = ⱍωE – ωOⱍ = ωE (t = 0) – ωO + ωE · t
Upshifts
Engine intervention is discussed using the ex- or
ample of an upshift from the direct gear (i = 2
1) to overdrive (i < 1). The following simplifi-
cations serve to illustrate the physical rela-
Q = MC ωO ·冤 1–i
i
M – MC t S
· ts + E
JE
·
2 冥
tionships more clearly: From (1), (2), and (3), this produces for a
쐌 The influence of the torque converter is time-constant clutch torque the lost energy as
disregarded. a function of the shifting-sequence parame-
쐌 There is no overlap of friction elements, ters.
i.e., only one friction element participates
.
in the gearshift. J1 · ω 1 = ME – MC
쐌 The engine torque remains constant during
the gearshift, thereby providing linear The slip time itself is dependent on the clutch
speed characteristics. and engine parameters, where
쐌 The vehicle speed during the gearshift is
ωO 1 – i 1–i JE
taken to be constant. ts = . · = ωO · · (4)
ⱍω Eⱍ i i ⱍME – MCⱍ
쐌 The heating of the friction linings by
briefly successive shifting operations is dis- This produces the lost energy to be absorbed
regarded. by the friction element
冢冣
2
1 MC · JE 2 1–i
Upshifts take place without an interruption of Q= · · ωO (5)
2 M E – MC i
the tractive force. Synchronization of engine
and transmission takes place via a friction i.e., the lost energy is dependent only on the
element in slipping-intervention operation. clutch and engine torques, the driving speed,
The following relative speed ensues between and the gear ratios.
the drive and output side of the clutch: When the clutch torque determined by (4) is
applied in (5), this produces the lost energy as
the sum of a share of the kinetic energy which 12 Lost energy Qcom (a) and clutch torque MC (b)
is released when the rotational masses are of a clutch as a function of slip time tS and engine
torque ME
braked to the synchronization speed and a
share of the engine combustion energy: Nm
a Qlimit
冢冣
2
ω2 1–i ω 1–i
Q = Qkin + Qcom = JE · O · + ME · tS · O · (6) +2,000 S
2 i 2 i
Nm
+1,600
Lost energy Qcom

Both these shares are roughly of the same or- 7
200
0
der of magnitude. At speeds of n = 3000 rpm 50
10
+1,200
ME =
and typical values for the gear step and the 5
engine-drag torque (i = 0.8, +800
JE = 0.3 kg · m2, ME = 100 Nm, tS = 500 ms), 3
this produces: +400
Qcom/Qkin ≈ 1...4 Qkin=1,645
0 200 400 600 ms

This clearly shows the possibilities of engine Slip time ts
intervention for reducing the power loss in
the friction elements. b Nm
500 M E=
A further significant aspect is derived from 200
Clutch torque MC
Fig. 12 Nm
(6): Only the share of lost energy stemming 400
a Lost energy Qcom 8 ts limit
b Clutch torque MC from the combustion energy is dependent on
300
the slip time tS. The decisive factor is the
10
0
MClimit Maximum product of the engine torque and the slip

50
200 1
clutch torque time. This means, however, that the slip time MC limit
2
MCmin Minimum MC min
can be extended accordingly when the engine 100 4
clutch torque 6
torque is reduced without an increase in the
æ UTS0275E
ME Engine torque
Qlimit Maximum total lost energy. In actual fact, the wear of the 0
0 200 400 600 ms
permitted lost friction elements even decreases with con- Slip time ts
energy stant total lost energy when the slip time is
extended. The temperature of the friction lin-
ings corresponds to the load on the friction a reduction in gear-shift sophistication (con-
elements. venience). A reduction of the clutch pressure
Figure 12a shows the lost energy absorbed is not permitted in this case, as otherwise
by the friction element as a function of the Qlimit will be exceeded.
engine torque and the slip time. The maxi-
mum permitted lost energy Qlimit and the It is now easy to tell from Figure 12 which
engine torque to be transferred during this possibilities are offered by engine interven-
gearshift determine the maximum slip time, tion. It is taken that the engine torque to be
for instance in accordance with point S. The transferred ME = 100 Nm can be reduced
maximum permitted energy Qlimit corre- during the slip phase to an average of 50%.
sponds in accordance with (5) to the clutch When first the case of constant clutch torque
torque determined by the slip time MClimit (shift quality) is considered, reducing the en-
(point 1 in Figure 12b). gine torque to 50 Nm results in a shortening
To reduce the lost energy, the clutch torque of the slip time from 400 ms to 245 ms (point
in relation to point S would have to be in- 1 씮 point 2) with a simultaneous reduction
creased and thereby the slip time shortened. in lost energy to 61% (point 3). If, on the
However, this would lead in equal measure to other hand, the slip time is kept constant, the
clutch torque can be reduced from 179 Nm to 1 Numerical values for text examples and Figure 12
128.5 Nm (point 1 씮 point 4) with a reduc-
ME MredE MC ME t0 Q/Q100
tion in lost energy to 72% (point 5). Nm Nm Nm Nm ms %
100 100 179 400 3740 100
The maximum sensible slip time is then ob- 100 50 179 245 2285 61
tained if the minimum clutch torque MCmin 100 50 128.5 400 2693 72
during the gearshift does not drop below the 100 50 100 628 3290 88
value after the end of the shift. On the one 200 200 360 200 3740 100
200 100 179 400 3740 100
hand, a smaller engine torque as a result of Table 1
engine intervention would result in a deterio-
ration in convenience; on the other hand, the determine the gear-shift sophistication
clutch torque for safety reasons should at any (convenience) to a substantial extent.
rate be so large that the non-reduced engine For a better understanding of the character-
torque can be transferred by the friction ele- istic relationships, damping in the drivetrain is
ment after the end of the shift. disregarded in the following analysis. It also ap-
In this example, it is taken that the torque plies on the assumption that the overall vehicle
to be transferred at least by the clutch is dynamics can be reduced to engine mass, in-
100 Nm in accordance with the engine torque put-shaft rigidity, and vehicle inertia.
(direct gear). This means that the slip time
can be stretched from 400 ms to max. 625 ms In the case of all the moments of inertia relat-
(point 6), again with a simultaneous reducing to the transmission output, the engine and
tion in lost energy to 88% (point 7). the vehicle are governed by the following:
Finally, it can be gleaned from Figure 12 that JE · Φ̈E = ME, JV · Φ̈V = –W (9)
even an engine torque of 200 Nm, which
without intervention would require a maxi- At the point when the one-way clutch
mum slip time of 200 ms with a minimum engages, the drivetrain resembles a torsion
clutch torque of 360 Nm (point 8), can be damper (Figure 13, next page), and the
traced back to the example with a torque of motion equations are as follows:
100 Nm (point 1).
JE · Φ̈E = c (Φ̈V – Φ̈E) + ME (10a)
The results of this analysis are on the safe side
in this respect because extending the slip time JV · Φ̈V = c · (Φ̈V – Φ̈E) –W (10b)
with constant lost energy results in reduced
friction-lining temperature and thus protects Since in this case it is not the absolute angle
the friction linings. Table 1 lists the numerical of rotation but rather only the deviations
values for these examples. from the basic rotation (i.e., the rotation of
the input shaft) which are significant, these
Downshifts two equations can combined. If the driving
In contrast to upshifts, downshifts take place speed υ for the short time segments to be
in the throttling mode with an interruption considered is taken as constant, this produces
of load. The engine is decoupled from the with
drivetrain and runs as a result of the torque it
ΦE = ΦEo + φE, ΦV = ΦVo + φF,
Φ·Eo = Φ·Vo = υ = const.
has generated up to the synchronization
speed. Only after the one-way clutch or the
friction element has engaged is the frictional and
connection re-established. The torque ratios
ψ = φV – φE,
when the synchronization speed is reached
the motion equation (12) produces the relative acceleration
ψ̈ + c · 冢 J1
E
+
1
JV冣· ψ = –ME (11) ψ = ω0 ·
2 M
c
cos (ω0 · t)
The natural frequency ω0 for this system as and (10b) the vehicle acceleration
follows
M
冑苴苴
Φ̈V = · [1 + cos (ω0 · t)] (14)
ω0 = 冢冣 c·
1
+
1
JE JV
JV
This means that, at time t0 when the one-
The general solution produces the accelera- way clutch engages, an acceleration jump
tion acting on the driver: takes place, namely
from Φ̈V = 0 for t < t0 (15)
ψ = A · (sin ω0 · t) + B · cos (ω0 · t) (12)
M
to Φ̈V = 2 · for t = t0
JV
(9) produces as the end condition of the
engine run-up phase: followed by a drivetrain vibration damped
in the real vehicle.
ME
ψ̈E = for t < t0 (13)
JE
Similar conditions are present in a gearshift
The energy to be applied here simply to from friction element to friction element
accelerate the engine is transformed when (overlapping gearshift), only this involves
the one-way clutch engages (at time t0) the additional problem of cutting in the
abruptly into a torque, which causes the friction element of the new gear exactly
input shaft to rotate: when the synchronization speed is reached.
As the damping effects of the torque con-
φEo = M verter and the rest of the drivetrain have
c
been disregarded in this analysis, the possi-
13 Diagram of drivetrain during load interruption and bility which engine intervention offers is all
establishment of frictional connection
the more clear:
a W
According to (13), the initial acceleration
acting on the driver at time t0 is directly pro-
JV
portional to the engine torque and thus the
engine acceleration during the run-up
ME
phase. With precisely timed control of the
engine torque in the time segment t ≤ t0 to
t >> t0, it is possible to create an almost
JE
Fig. 13 continuous transition from the range of
a Load interruption tractive-force interruption to the range of
b Frictional connection
tractive-force transfer.
JV Mass moment of
Implementation takes the form of a
b
inertia of vehicle W marked reduction in engine torque at time
drivetrain JV t0 followed by renewed control-up in accor-
JE Mass moment of dance with a time function. The conve-
inertia of engine ΦV nience can be varied within broad limits
ME Engine torque ME
with this control-up.
ΦV Angle of rotation
æ STS0276E
of vehicle drivetrain
JE ΦE There is an equally clear possibility in
ΦE Angle of rotation
of engine gearshifts without a one-way clutch of influ-
W Running resistance encing the engine acceleration by control-
ling the engine torque in the time range The overrunning point is identified through
t ≤ t0 and thereby reducing the time de- continual monitoring of the transmission
mands on the cut-in precision of the friction input speed in the time phase after t0. For
element at the synchronization point. this purpose, the maximum speed is calcu-
lated in the time phase t0...t3 as well as the
Sequence Control speed gradient.
The process of reducing the engine torque The overrunning point is identified in the
is essentially very simple. However, effective event of a reduction in the gradient by more
control requires precise coordination as the than a pre-specified threshold value, and
entire process only lasts approximately engine-torque control begins with control-
500 ms. down to a pre-specified value in accordance
with a pre-specified time function.
Pure time control of engine intervention is
not practicable because different variables In order to determine the speed n3 at the
which determine the sequence (such as start of control-up, the synchronization speed
clutch fill times, plate friction coefficients, n4 = n1/i in the new gear is calculated from
and similar) fluctuate within broad limits the maximum speed n1 at the overrunning
depending on the temperature and the point and the ratio jump i of the gear change
service life. to be carried out. A speed-dependent share
Δn is added to this synchronization speed
As engine intervention is directly linked to in order to obtain a derivative action for
the shifting sequence, a speed sequence control-up. When the speed n3 = n4 + Δn is
control system suggests itself. The character- reached, torque control-up begins in accor-
istic variable that characterizes the shifting dance with a pre-specified time function. The
sequence exactly is the transmission input end of the gearshift is identified as soon as the
sequence. The engine speed is also suited with value of the non-corrected torque is achieved.
limitations to transmissions with hydrody-
namic converters as the controlled variable. 14 Sequence control, upshifts with engine
intervention
This is therefore important because it re-
quires a separate sensor to record the trans-
mission input speed, and not every transmis- S2
signal S
sion is fitted with this sensor for cost reasons. S1

Shift
In the interests of clarity, the following text

describes the control system with the trans-
mission input speed as the characteristic a
Accele-
ration a
variable and reference is made, where neces-

sary, to the limitations or changes when the
engine speed is used. nE1 nE2
nE3
nE nE4
Upshifts
Speed n
The time curve of the characteristic vari- nT Fig. 14

ables for an upshift is depicted in Figure 14. 1 Control-down phase
The ratio of the old gear is retained up to 2 Control-up phase
the overrunning point t2; only then does ME
a Acceleration
Torque M
only a slipping clutch intervene. For this 1 2

nT Transmission input
reason, the engine torque cannot be reduced
æ STS0277E
speed
before the overrunning point is reached, t0 t1 t2 t3 t4 t5 t6 nE Engine speed
otherwise this would entail an intensified Time t ME Engine torque
dip in the output torque in the phase t1...t2. S Shift signal
For upshifts in the upper load range (greater 쐌 retarding the ignition angle too early ex-
than half load), the engine speed is used as the tends the engine revving phase and thus
controlled variable instead of the transmission the time of tractive-force interruption and
input speed because here the shifting points 쐌 engine intervention after the one-way
are at such great engine speeds that the con- clutch has engaged does not bring about
verter operates in the clutch range and is thus any improvement in convenience, but
subject to roughly constant slip. rather a deterioration as this causes a dip in
Gearshifts at part load on the other hand torque for the duration of the engine inter-
take place in the converter range. This means vention.
that the slip can change very substantially
during a gearshift. Here the engine speed is The synchronization speed is calculated via
no longer suitable for determining the syn- the gear step from the transmission input
chronization speed. In this case, a speed at the start of the gearshift. The engine
superimposed form of time control, which torque is abruptly reduced approximately
terminates engine intervention after a pre- 200 rpm before the synchronization speed is
specified time, is suitable for the part-load reached until this speed is reached or slightly
range t3...t4. exceeded. Then the engine torque is slowly
controlled up again.
Downshifts The synchronization speed cannot be
The time curve of the characteristic variables calculated directly by means of the engine
for a downshift is depicted in Figure 15. speed as a result of the slip at the hydrody-
Precise determination and recording of the namic torque converter. A consideration of
synchronization speed are crucial to engine the converter program map with the requisite
intervention in the case of downshifts because accuracy requires too much computation
effort in the microcontroller.
It is possible however to calculate the syn-

15 Sequence control, downshifts with engine chronization speed from the transmission
intervention
output speed by multiplying it by the corre-
sponding gear step. It is now possible to iden-
S1 tify by means of the engine speed when the
signal S
S2 synchronization point is reached, since the

Shift-
speed difference between the engine and the

turbine is approximately zero when the en-
gine revs up freely (interruption of tractive
a force) to the synchronization point.
Accele-
ration a
nE3
nE2 The different possibilities of torque reduction
nE1 are now discussed in the following text.
nE
Speed n
nT
Fig. 15
2 Control-up phase ME 2
Torque M
a Acceleration
nT Transmission input
æ STS0278E
speed
nE Engine speed t0 t1 t2 t3 t4 t5
ME Engine torque
Time t
S Shift signal
Shifting the Ignition Angle Engine-Torque Specification

The oldest version of engine intervention In appropriately equipped vehicles with
involves intervention through shifting of the their CAN network of all the ECUs in the
ignition angle. This form of intervention drivetrain (Figure 17), torque reduction is
offers the following advantages: performed on the basis of a torque interface
쐌 continuous regulation of the engine between engine management (TI-Motronic)
torque within broad limits, and electronic transmission control (ETC).
쐌 short response time, and The torque reductions of the ABS and TCS
쐌 availability in all vehicles with gasoline- ECUs must also be taken into consideration.
engines.
Figure 18 shows how a current transmis-
Figure 16 shows in schematic form the de- sion-control system calculates the desired
pendence of the engine torque on the igni- engine torque intervention (ETI_Etc).
tion angle for different load conditions and The next torque intervention is determined
engine speeds. It is clear from this figure that as a function of the available torque (actual
adjusting a pre-specified engine torque gen- torque). The torque M is the engine torque of
erally requires an ignition map as a function the engine-management system without
of engine load and engine speed. intervention by transmission control.
The response time τ between the initiation

of engine intervention and the start of
reduction in engine torque is specified by
the ignition angle, therefore
1
τ≈·
(z/2) · n E
where z is the number of cylinders and n E
the engine speed. In the effective speed
range n ≥ 2000 rpm, the maximum delay
for a 6-cylinder engine is 10 ms for initiation
and 30 ms for a complete reduction in
engine torque.
Fig. 16
16 Engine torque as a function of ignition angle 17 ECUs in the drivetrain
a Full load (index F)
αF1, MF1 Ignition angle
or engine torque
without engine
a
MF1 αF1 intervention
TI- αF2, MF2 Reduced
ETC
Motronic ignition angle or
Engine torque ME
High-Speed engine torque with

MF2 25 °KW αF2 CAN engine intervention
MP1 b αP1 b Part load (index P)
αP1, MP1 Ignition angle
MP2 20 °KW αP2 or engine torque
ABS TCS without engine
intervention
æ STS0282E
æ STS0281E
50 40 30 20 10 0 10 °KW αP2, MP2 Reduced

Ignition angle α ignition angle or
advance retard engine torque with
engine intervention
18 Calculation of engine-torque intervention (ETI) in current transmission control
Torque without Limitation of Calculation of torque

transmission torque gradient intervention from
intervention program maps and functions
Parameter
of gradient
limitation
Engine
torque M Gradient ETI_Dyn_Lim
limitation
Parameter Calculate No
of shift status reduced intervention
torque
S_Phase Status TI_State Torque- Torque

check request limitation
calculation
Shift-sequence Torque- Determination of Limitation of

phase intervention torque intervention torque intervention
status as a function of
æ UTS0279E
phase and status Desired torque
of shift sequence intervention
Key to Figure 18: Torque Converter Lockup Clutch

ETI_Etc = f(ETI_Dyn_Lim, TI_State) Application and Operating Concept
The hydrodynamic converter is (necessitated
where
by its operating principle) subject to a level of
ETI_ETC: Engine-torque request of transmission
control
slip which is required particularly for conve-
ETI_Dyn_Lim: Dynamically limited engine torque (reduction nience reasons during startup and in certain
of torque change in the engine to ensure driving situations to increase torque. Since
convenience) this slip also involves a simultaneous loss of
TI_State: Current status of torque intervention power, the torque converter lockup clutch
(TCLC) was developed with this in mind (see
also section entitled “Torque Converter”).
It is only advisable to lock up the converter

from a specific speed because at low speeds
the irregular rotation of the engine would
cause uncomfortable vibrations in the
drivetrain. The controlled torque converter
lockup clutch (CTCC) was developed in order
to be able to utilize these ranges for lockup
as well.
Controlled Torque Converter Lockup Clutch point value and corrects the pressure contin-
The controlled torque converter lockup ually. Special functions perform changeovers
clutch (CTCC) sets a very low level of slip between the individual states and provide for
(40...50 rpm) and thus an almost stationary a comfortable shift performance.
state. In this way, it keeps unwanted vibra-
tions away from the drivetrain. The converter
lockup clutch therefore has three different
states:
쐌 open, 19 Characteristic (pictorial diagram) of controlled
torque converter lockup clutch (CTCC)
쐌 controlled, and
%
쐌 closed.
100
CTCC CTCC
These states are defined by means of charac- open closed
Throttle-valve opening
teristic curves, which are plotted like shift

curves for each gear against throttle-valve
opening and driving speed (Figure 19). As is
50
the case with shift curves, fuel consumption
and tractive force are crucial criteria for the
torque converter lockup clutch.
In slipping controlled operation, the speed CTCC CTCC CTCC
open controlled closed
differential between the converter impeller
æ STS0280E
and turbine is constantly set to a low value. A 0
0 1,000 2,000 rpm 3,000
closed control loop constantly compares the Engine speed nE
speed differential with a pre-specified set-
20 Tiptronic® shifting strategies (flowchart)
Measured Throttle valve Engine speed Vehicle speed Lateral Linear

4 160 acceleration acceleration
variables 3 5 120 200
2 6 80 240
1 7 40 280
Level
Adaptation of shift characteristics (transmission and torque converter lockup clutch)
Measured-value summation Modulation factor Shift/TCLC characteristics
1 Filtering MF1... MF5 SK5
SK3 SK4
Averaging MFA N SK1 SK2
Weighting % V
2
Short-time influencing Special function
Prevention of Gear holding Active shifting Upshift during
overrun upshifts in curves Jump to SC5 brake overrun
ahead of curves to low μ
3
Manual tip (nudge) shifting P + Upshift

D Computer-aided adherence to
- Downshift permitted speed limits
æ UTS0203-1E
338 Electronic Transmission Control Control of Continuously Variable Transmission
Control of Continuously Open and Closed-Loop Control

Functions
Variable Transmission The aforementioned transmission equipment
Requirements requires the following open and closed-loop
Continuously variable transmissions that op- control functions:
erate according to the wrap principle have a 쐌 contact-pressure control,
whole variety of different equipment 쐌 ratio control,
specifications (Table 1). The following equip- 쐌 driving program,
ment packages are widely used in compact 쐌 clutch activation,
and mid-range class vehicles: 쐌 activation for converter and torque
쐌 When the master/slave concept is used, the converter lockup clutch,
primary pulley (transmission input side) 쐌 pump activation,
has double the surface of the secondary 쐌 reverse-gear lock, and
pulley (transmission output side). The 쐌 deactivation of limp-home function.
pressure in the primary chamber can thus
always be below the secondary pressure. Contact-Pressure Control
쐌 The converter with torque converter The belt contact pressure is adjusted in
lockup clutch as power take-up element of- accordance with the current load situation
fers very good starting convenience and fa- with the aid of the measured secondary pres-
cilitates a good starting response through sure. To achieve a high level of efficiency, the
torque increase so that the large ratio span secondary pressure is reduced to such an ex-
of the CVT is completely beneficial to the tent that the current engine torque can still be
overdrive range. transferred to a specific degree of safety and
쐌 Two wet clutches for the forward and reliability without the belt slipping.
reverse gears.
쐌 Variable-capacity pump.
쐌 Convenient fail-safe and
limp-home strategies. 1 Variations of CVT based on wrap principle
Assembly, Variants
In the event of a control-elec- function
tronics failure, the fail-safe and Wrap Band Chain Belt
limp-home requirements par- element
tially determine the hydraulic Variator principle Master/slave Partner principle Partner principle
Converter: Fitted Not fitted Not fitted
concept. Engine overspeeding
– Converter clutch Yes No No
and associated high slip at the
– Slip duration Brief Permanent Permanent
driven wheels must be avoided Clutch:
at all costs. An adjustment in – Type Friction surfaces Magnetic-powder Magnetic-powder
the overdrive direction would – Pressures Low High High
satisfy this requirement, but – Slip duration Brief Permanent Permanent
starting from a standing stop Pump adjustment Constant 2-stage Continuous
Limp home Not possible Limited Unlimited
would no longer be possible.
(loss of (increased
convenience) fuel
consumption)
Vehicle:
– Class Small to Compact class Mid-range and
mid-range class Luxury class
– Engine size <3 l <2 l >2 l
– Type of drive Front transversal Front longitudinal Rear
Table 1
Electronic Transmission Control Control of Continuously Variable Transmission 339
Ratio Control Pump Activation

The gear ratio can be changed by means A variable-capacity pump must be used to
of the primary pulley. The enclosed fluid ensure high transmission efficiency. This
volume determines the axial position of the pump enables the delivery flow to be limited
moving part of the primary pulley and thus at high speeds.
the radius on which the belt circulates on the
pulley. The primary pressure adjusts itself in Suitable suction-throttled pumps which
response to the secondary pressure. operate without additional activation have
been in development for years, but have failed
The requirements of driveability determine to make a breakthrough as yet. An initial step
the necessary adjustment speed. For example, towards a “variable-capacity pump” has been
in the case of kickdown, it is necessary to taken in the form of a two-stage version, in
switch from overdrive to low within 1.5 s. On which the more favorable delivery flow can be
the other hand, the pump delivery limits the selected as a function of the current demand.
adjustment speed.
Further concepts are feasible with continu-
Driving Program ously variable pumps, in which secondary-
A driving program ascertains the desired gear pressure control and pump adjustment can
ratio. In addition to different program maps be combined.
for normal operation, in which there is the
option of choosing between economical and Reverse-Gear Lock
sporty operation (see also section entitled Engagement of the reverse gear is disabled
“Adaptive Transmission Control, ATC”), it is during forward driving at speeds above a
also possible to implement special functions limit to be defined (e.g., 7 km/h).
such as kickdown, downhill driving, etc.
It is also possible to simulate range trans- Deactivation of Limp-Home Function
missions, where any intermediate variations Limp-home is an emergency function which
between copying a manually shifted transmis- is shut down during normal control opera-
sion and an automatic transmission are real- tion.
izable (see also chapter entitled “Transmis-
sion for Motor Vehicles”). The fail-safe function remains permanently
activated so that engine overrevving is
Clutch Activation avoided even in the event of a partial failure
The interrupting clutch between engine and or late identification of partial failures.
drivetrain is designed as a function of the po-
sition of the selector lever (P-R-N-D), the en-
gine speed, and the engine load.
Activation of Converter and

Torque Converter Lockup Clutch
In order to achieve the greatest possible
efficiency, it is essential for the converter to be
locked up as early as possible. Depending on
the power requirement, the torque increase is
used up to different speeds for acceleration.
340 Electronic Control Units (ECUs) ECUs for Electronic Transmission Control
ECUs for Electronic TC (Figure 1a). In the USA, on the other

hand, it is predominantly the combined dri-
Transmission Control vetrain ECU (MEG) that is used for 4-speed
Application and 5-speed automatic transmissions (Figure
In implementing electronic transmission 1b), the reason for this being that automatic
control, engineers can choose to locate the transmissions dominate the market in Amer-
ECUs in different positions in the vehicle. ica with a market share of over 85 %.
There are, for example, separate, combined, The newer 6-speed or 7-speed transmission
mounted, or integrated ECUs (Figure 1). types and the increasing demands on engine-
The way in which these ECUs are distributed management systems (emission-control legis-
in the vehicle is essentially determined by lation, CARB requirements) have now started
쐌 the ratio of vehicles with automatic trans- a trend in the USA moving away from com-
missions to vehicles with manually shifted bined ECUs in favor of separate ECUs. This
transmissions and trend is consolidated still further by the latest
쐌 the demands made by the transmission on generation of 6-speed transmissions. These
the control system (performance of the mi- transmissions are already equipped with
crocontroller used). electronic modules with integrated electronic
In Europe, the market is still dictated by the circuitry.
separate printed-circuit-board ECUs ME and
1 Possible distribution of ECUs in the drivetrain
a ABS control unit TCS EC
Wheel brakes CAN
Separate ECUs Engine control unit (ME) Transmission control unit (TC)
or alternatively:
ME
Add-on ECU (ME) TC
or integrated ECU (TC)
Engine Transmission
b
CAN
Combination ECU Drivetrain ECU (MEG)
or alternatively: MEG
Combination ECU
and intelligent power
driver stages
Fig. 1
a Layout with separate
ME and TC printed
æ UAE0954E
circuit board ECUs

b Layout with
combined drive- Engine Transmission
train ECU
Electronic Control Units (ECUs) ECUs for Electronic Transmission Control 341
Design and Operating Concept The ECU acquires the transmission-fluid

The various ECUs and their technical and temperature as an analog input signal because
functional details will be discussed below. the properties of the fluid have a significant
effect on shift quality, especially when the en-
Printed Circuit Board ECUs gine is cold. The ECU acquires the position of
The most widely used ECUs currently are the selector lever in the form of a digital sig-
printed circuit board (PCB) units. nal. The following information can also be
Figure 2 shows an ECU with a 32-bit obtained and evaluated via the CAN inter-
microcontroller (Motorola 683xx) for a face:
ZF 5-speed transmission. This transmission 쐌 accelerator-pedal position
has been in mass production at BMW for (driver command),
some years now. The figure depicts the 쐌 kickdown switch,
essential layout and the data flow of the ECU 쐌 engine temperature, and
in a block diagram. The ECU itself can be 쐌 engine torque.
roughly divided into three sections:
2. Computer Core
1. Input side The computer core comprises microcon-
The input side comprises the power supply troller, flash, RAM, EEPROM, analog-digital
(terminals 15 and 30), signal acquisition, and converter, and CAN bus system.
the communication interface.
The input signals include the signals for 3. Output Side
engine speed, turbine speed, output speed, The output side features the driver stages for
and wheel speeds. The transmission control the on/off valve, ASICs, current control
unit usually receives the engine-speed and (CG205), and low-level signal driver stages.
wheel-speed signals via the CAN interface
from the acquiring ECUs (engine and ABS
control units).
2 Printed circuit board ECU for electronic transmission control (block diagram)
Ignition
Internal power supply On/Off
Battery voltage solenoid valves
Micro- ASIC for
controller current-controlled
high-power
Interface engine μC driver stages
management Actuators
CAN interface
Vehicle-speed
signals FEPROM
Rotational frequency
Analog RAM
voltage signals Interface for
analog signals ADC
Diagnosis EEPROM
Serial interface
for diagnosis
CAN
Digital Interface
æ UAE0953E
input signals Interface for Low-power to vehicle

digital signals driver stages
Drivetrain ECUs 4 BMW drivetrain ECU from 1983

MEG drivetrain ECUs (MEG = engine
ETC/EGAS transmission) are based on the
standard printed circuit board ECUs for
engine and transmission, and are commonly
used in the USA. As the block diagram in
Figure 3 shows, the main advantage of this
ECU lies in the fact that specific electronic
components only have to be fitted once,
thereby reducing costs.
An MG ECU (MG = engine/transmis-
sion) was the first example of elec-
tronic transmission control to be
mass-produced.
This ECU was devel-
oped back in 1983
for BMW for use
with a ZF 4HP22
automatic trans-
mission (Figure 4).
æ UAE0946Y
3 MEG drivetrain ECU for electronic transmission control (block diagram)
Voltage ME 7 MEG 7
regulator Memory Memory
Motronic
CAN Voltage Motronic
μC Motronic
Motronic regulator μC
output
Diagnosis output driver stages
EGAS driver stages
Motronic Controller CAN
inputs μC
Diagnosis EGAS
Voltage EGS Controller
regulator Memory
EGS Motronic
CAN + EGS ETC
μC output
ETC ETC μC
inputs driver stages
Diagnosis output
driver stages
ETC Watchdog Memory
æ UAE0952E
inputs
Not necessary in MEG 7 integrated ECU

5 MG7.9 combined drivetrain ECU
æ UAE0950Y
Figure 5 shows the current configuration of 1 Development of computer performance
a drivetrain ECU for the block diagram in Year Computer Memory RAM
Figure 3. Year Computer Memory RAM
1983 Cosmac 8 k ROM 128 byte
Since then, the demands on ECU computing 1988 80515 32 k ROM 256 byte
power and memory capacity have changed 1992 80517 64 k ROM 256 byte
1996 80509 128 k flash 2k
dramatically (see Figure 6 and Table 1).
199x C167 256 k flash 4k
As the figures in Table 1 demonstrate, these 1996 683xx 256 k flash 8k
demands are constantly rising and there is no 2001 MPC555 448 k flash 28 k
end in sight to their development. 2003 MPC555 1 MB flash 28 k
2005 ? 1.5 MB flash 66 k Table 1
6 Increasing demands on ECU computing power and memory capacity
8-Bit (80519)
Computing power/memory capacity
64k ROM
32-Bit (MPC555)
448k Flash
8-Bit (8051)
8k ROM
8-Bit (80515) 32-Bit (683xx)

32k ROM 256k Flash
æ UAE0951E
1983 1987 1992 1996 2001 Year

Microhybrid ECUs Passive components are electrically contacted

The introduction of new transmissions (such by means of conductive bonding agents.
as the ZF 6HP26) has seen a transformation In contrast to the currently mass-produced
in the type of ECU from the PCB to the mi- circuits with LTCC (low-temperature cofired
crohybrid. This development has been influ- ceramics), finer layout structures are used in
enced by the changing demands, mainly be- the new systems with 32-bit processors. This
cause of the environmental conditions under relates in particular to the via density and the
which the ECU is used (Table 2). bondland size.
The microhybrid ECU basically contains
the same circuitry as the PCB, but now with The previous bondland grid of 450 μm would
unencapsulated semiconductor components require four bond rows and at least three
being used, i.e. as “bare” silicon chips. wiring layers to route the computer core.
Electrical contacting is provided by wire With the via grid of 260 μm used, two bond
bonding (in the PCB ECU with soldering). rows are sufficient, and this even with a re-
duced space requirement and only two wiring
layers.
2 Technical limits for PCB and microhybrid
Type PCB Microhybrid Figure 7a shows the ABS computer bond

zones (44 bonds) on an LTCC standard sub-
strate compared with the 32-bit controller
in the diesel-control system (240 bonds) on
LTCC-fine-line in Figure 7b.
Installation Interior or In transmission
location engine compartment In addition, Figure 8 shows a comparison of
Temperature –40...+85/+105°C –40...+140°C the wiring density of the inner layers (Figure
Vibration ...5 g ...30 g 8a) and the reverse side of the hybrid with
Protection IP 40 / IP 69 IP6K9K in ATF
the integrated resistors (Figure 8b).
7 Bondzone microcontroller LTCC compared 8 Hybrid wiring densities

with LTCC-fine-line
a a
Fig. 7 b b
a On LTCC standard
substrate
b On LTCC-fine-line
substrate
æ UTS0318Y
æ UTS0319Y
Fig. 8
a Inner layers
b Reverse side with
resistors
The following significant measures were Figure 9 shows the complete microhybrid in
taken to bring about an improvement in the its housing. The following procedures are
process for microhybrid ECUs: used for the assembly process:
쐌 use of finer punch needles, 쐌 All the components are bonded with
쐌 finer screens, conductive bonding agent.
쐌 adaptation of the pastes used, and 쐌 Bonding is carried out using a 32 μm gold
쐌 tolerance optimization through adapted wire and a 200 μm aluminum wire.
process management. 쐌 The hybrid is bonded to the steel plate with
heat-conducting bonding agent.
This compression of the layout makes it pos- 쐌 The connection to the glass bushing is
sible to produce the circuit for transmission made by a 200 μm aluminum bond.
control on an area measuring 2 x 1.2≤. In 쐌 The housing is hermetically sealed tight.
other words, a substrate with the working for-
mat of 8 x 6≤ alone can process 20 circuits in
parallel.
For optimum cooling of ICs subject to high

heat loss, thermal vias with a diameter of 300
μm are filled in parallel to the function vias.
This increases the thermal conductivity of the
substrate from approx. 3 W/mK to effectively
20 W/mK.
9 Microhybrid ECU in steel housing
æ UAE0948Y
ASIC Chips Figure 10a shows the encapsulated ASIC, as is

In addition to computer and memory chips, used in PCB ECUs.
application-specific integrated circuits make There is also the option of adjusting the
up a significant proportion of the electronic current range and the PWM output fre-
components in the ECUs. quency with the aid of an external circuit.
Different functions have been combined
into ASIC chips for the purpose of reducing Watchdog ASIC CG 120
costs and standardizing the electronic design Due to the fact that the 6-speed transmission
of transmission-control systems. These ASICs with integrated electronic module no longer
are available in encapsulated and has a mechanical connection between the se-
unencapsulated form and used for both lector lever and the transmission, the control
microhybrid and PCB ECUs. Transmission system requires specific safety mechanisms.
control features various ASICs which are cur- This function is performed by the ASIC
rently in volume production; the three ASICs CG120 (Figure 10b), which monitors the
used in the microhybrid ECU are discussed in function of the microcontroller (see also the
the following. chapter entitled “Diagnostic functions”).
The ASIC CG120 performs the following
Current-Regulator ASIC CG205 functions:
The current-regulator ASIC CG205 with
integrated shunt was developed for high-pre- 쐌 power supply with 5 or 3.3 V,
cision pressure control in the transmission. 쐌 sensor supply,
It achieves a control precision of 1% over the 쐌 watchdog,
entire temperature range. 쐌 serial interface,
쐌 CAN interface,
쐌 ISO 9141 interface, and
쐌 programmable via SPI interface.
10 ASIC chips
I/O ASIC CG115

In order to achieve the high level of integra-
tion in the microhybrid, it is necessary to
combine as many functions as possible in
a
CG one ASIC. Achieving this with individual
205 components would therefore take up too
much space on the substrate.
b The following functions are integrated in the

I/O ASIC CG115 (Figure 10c):
CG
120 쐌 voltage monitoring,
쐌 inputs and outputs for digital signal trans-
mission,
쐌 2 inputs for inductive Hall-sensor signals,
c
CG
115 쐌 8-channel analog multiplexer,
쐌 serial interface, and
Fig. 10
쐌 programmable via SPI interface.
a Current-regulator
æ UAE0955Y
ASIC CG205
b Watchdog ASIC
CG120
c I/O ASIC CG115
Electronic Control Units (ECUs) Thermo-Management 347
Thermo-Management 1 Heat-dissipation model in the housing of a

microhybrid ECU
Dissipating the heat losses generated in the 1 2 1 3 4
ECUs represents a prime consideration in the
TJu
design of mechatronic modules, particularly RCh
in those cases where “hotspots” cause a highly RKl
uneven distribution of the heat losses. Figure RSu Fig. 1
RKl 1 Bonding agent
1 shows the heat-dissipation model in an RBo 2 Substrate
ECU to the point of a heat sink which is situ- RKo
3 Si chip
æ UTS0320Y
ated in the valve housing in this example. RVe
4 Thermal vias
This is an LTCC microhybrid in a welded TUm
5 Al valve housing
steel housing which is mounted on the alu- 5 6 7 6 ATF
minum housing of the hydraulic main con- 7 Steel housing base
trol stage.
2 Cross-section through a 4-layer microhybrid with
thermal vias
Effective heat management of the IC neces-
sitates a close contact between the chips and
the housing. Materials with high thermal
conductivity are suitable for use in this re-
spect.
As other tests on the various substrates for
high-temperature applications have shown,
these materials demonstrate very different
thermal-conductivity properties. LTCC glass
ceramic shows itself to be less effective than
æ UTS0321Y
aluminum-oxide ceramic (Al2O3) by a factor 1 mm
of almost 10. However, this drawback is com-
pensated by thermal vias in the microhybrid
to such an extent that LTCC technology
demonstrates equally good levels of thermal Pl = (Tj – Ta) / Rjth
conductivity as aluminum-oxide technology.
where
Figure 2 shows an area of thermal vias.
The production process creates these heat- Tj Junction temperature
dissipating thermal vias (heat spreaders) in Rjth Thermal internal resistance
parallel to the electrical connections. Ta Ambient temperature
Essentially, the terms “power loss”, “depletion- The thermal resistance Rjth (auxiliary quan-
layer/junction temperature”, and “heat dissi- tity) is dependent on the geometric parame-
pation” define the limits for a microhybrid ters and the specific thermal conductivity
ECU. of the material and is determined from
measurements. The maximum permitted
The power loss Pl can be described in junction temperature Tj determines the maxi-
simplified form for stationary operation mum permitted power loss Plmax, where
as follows: Tj is dependent on the material (for silicon
Tjmax = 150...200°C). Current specifications
laid down by microcontroller manufacturers
establish an upper limit of Tjmax = 150°C.
Since Tj is also dependent on the design, a
348 Electronic Control Units (ECUs) Thermo-Management
3 Comparison of thermal resistance of aluminum-oxide consideration of design rules is important

and glass ceramics (LTCC) for optimization of the junction tempera-
ture, e.g.:
a 쐌 Circuit sections subjected to thermal load
1 2 4
should not be designed in accordance
with the otherwise applicable minimum
criteria, rather affected driver-stage areas
(transistors/pn junctions), for example,
should be geometrically enlarged.
쐌 “Hotspots” should not be positioned in
the corners of the ICs, this enables the IC
substrate material to act in all directions
as a heat spreader.
The design of the ASICs for transmission

control described in the “ASIC Chips”
section is suitable for maximum junction
temperatures of Tjmax = 175°C.
b
1 3 4
In relation to the complete microhybrid
system, designers are working towards opti-
mizing the level of sophistication between
increasing IC size, using thermal vias, and
mounting on special substrates such as DBC
or PCs.
Explanation of abbreviations:
DBC (direct bonded copper): copper-coated
ceramic
and
PCs (power chips): chips soldered onto
copper plates.
c Figure 3 shows a comparison of the thermal

resistance of aluminum oxide (Figure 3a)
and glass ceramic, i.e., LTCC (Figure 3b), and
also specifically in chart form (Figure 3c).
Fig. 3
a Aluminum-oxide
th
ceramic
Thermal resistance
Rth = 12 K/W (CS 200

on aluminum)
b LTCC
Rth = 10...11 K/W (CS
200 in ABS 5.3)
c Comparison
æ UTS0322E
1 Chip
2 Al2O3
3 LTCC (thermal vias) Thermal vias. surface
4 Aluminum
Electronic Control Units (ECUs) Processes and Tools Used in ECU Development 349
Processes and Tools Used in Circuit Simulation with SABER

Before a circuit is built with real components,
ECU Development it is possible to test its function using the
Simulation Tools SABER® simulation tool. Many component
Simulation of individual components and of manufacturers already offer the data on their
the complete system is playing an increasingly products in a SABER library for this purpose
important role in improving and stepping up (Figure 2).
the development process. The advantages of
mathematical modeling over actual physical These data can be used to test the circuit with
models (prototypes) are as follows: regard to its robustness, thermal properties,
쐌 frequent reproducibility as desired, worst-case performance, and EMC behavior
쐌 deeper understanding of system behavior, and thereby facilitate any necessary circuit
쐌 individual parameter analyses possible, optimization at a very early stage in its devel-
쐌 lower costs, opment.
쐌 less time required for model modifications,
and This approach can reduce the number of
쐌 flexible application in all technical fields. redesigns.
The following text deals briefly with some

tools for simulation that are used.
Figure 1 shows by way of example the
thermo-simulation of a microhybrid ECU.
It clearly shows the high temperature of the
hotspots in the electronic circuitry which in
this case occurs at a voltage regulator. This
simulation is used to optimize the positioning
of the thermal vias and the distribution of the
components before the real ECU is built.
1 Thermo-simulation of a microhybrid layout 2 Simulation with SABER

æ UTS0323Y
æ UTS0324Y
350 Modules for Transmission Control Application
Modules for Transmission Control

Modules are compact function and Modules exhibit the following advantages
constructional units that enable different over individual components:
standardized components to be integrated 쐌 reduced space requirement,
with the minimum component and space 쐌 reduced mass,
requirements and simplified interfaces. 쐌 fewer individual parts,
Hydraulic, electronic, and electrohydraulic 쐌 increased reliability,
modules are available to suit the level of 쐌 standardization of the components
integration required. integrated in the module, and thus
쐌 reduced costs.
Application
These advantages are essentially provided by
Mechatronic Modules simplifications to the mechanical and electri-
Since the number of sensors and actuators cal interfaces between the individual compo-
deployed in the automatic transmission is in- nents and the transmission. There is in future
creasing while the amount of space available still further potential for reducing costs in
to accommodate them is diminishing, the this field.
drive towards ever-increasing integration However, any further increase in module
suggests itself not just for cost reasons. complexity could also lead to drawbacks; this
Mechatronic modules can be used to com- is because a manufacturer of modules must
bine actuators and sensors, their contacting demonstrate competence over the entire
and if necessary even an ECU into a single spectrum of the components contained
control system. The level of integration, or therein. Since it is more difficult to separate
the scope, of a mechatronic module is dic- the components in the event of a fault, there
tated by the requirements of the relevant is the danger that increased follow-up costs
vehicle manufacturer (Figure 1). may be incurred in the event of both assem-
bly errors and repairs. Mechatronic systems
Improvement Potential therefore place strict demands on quality and
Modular technology offers comprehensive reliability in the manufacturing process and
potential for improvement in comparison necessitate precautionary measures to ensure
with individual components (Figure 2). that subsequent repairs can be carried out.
1 Variable scope of a mechatronic module
Electronic
Sensors Actuators
control unit
• Single unit • Temperature • on /off
• Mounting • Pressure • PWM
• Integrated • Rotational speed • Pressure regulator
• Combined • Position • (Stepping motor)
Electronic module (EM)
Hydraulic module (HM)

æ STS0291E
Electrohydraulic module (EHM)

Modules for Transmission Control Module Types 351
2 Improvement potential through modular system
Transmission manufacturer Vehicle manufacturer

Development Purchasing Assembly Final inspection Purchasing/Assembly
Individual components Individual components
ECU ECU ECU
Wiring Wiring Wiring

harness harness harness
4 to 8 4 to 8 4 to 8
actuators actuators actuators Transmission Transmission
4 to 8 4 to 8 4 to 8
sensors sensors sensors
Module Module
Electrohydraulic Electrohydraulic Electrohydraulic Transmission Transmission
module module module system system
Improvement potential Improvement potential

• Reduced expenditure • Reduced logistics • Simple assembly • Better adjustment • Reduced logistics
on electronics • Reduced costs • Lower wiring-harness • Increased reliability • No ECU mounting
• Optimized development scope • Lower wiring-harness
æ STS0315E
• Only one partner for • Reduced costs scope
development and • Space saving
supply
Module Types 3 HM 5R55 hydraulic module (example: Ford)
Hydraulic Modules (HM)

The hydraulic module (HM) (example, 1 2 3 4 4 5
Figure 3) represents the first step towards the
simplified assembly of modules. This mod-
ule type comprises the following
components (primarily sensors and
actuators):
쐌 pressure regulator (1),
쐌 PWM valve (2),
쐌 integrated switching valves (4),
쐌 transmission connector
matched to the vehicle type,
쐌 temperature sensor,
쐌 electrical connections, and Fig. 3
쐌 joint filter seal between 1 Pressure regulator
module housing and 2 PWM valve
3 Transmission
adapter plate with
connector
쐌 hydraulic channels (5).
æ UTS0283Y
4 Integrated switching
valves
5 Adapter plate with
hydraulic channels
352 Modules for Transmission Control Module Types
Electronic Modules (EM) be routed, depending on the application.

Design As well as sensitive signal conditioning, the
The next stage of development involves the electronic transmission-control system requires
fusion of mechanics and electronics into the 쐌 power electronics with a high-precision
mechatronic electronic module (EM), con- current regulator (1 A with 1% accuracy)
sisting of sensors and ECU. In many cases, for pressure control and
it offers the possibility for system simplifica- 쐌 semiconductor relays for currents of
tion and cost reduction. up to 8 A.
Electronic modules already come in differ-
ent types and versions. What they all have in Further functions are:
common is their optimum adaptation to their 쐌 power supply,
environment, as the examples in Figures 4 to 쐌 safety monitoring, and
6 demonstrate. 쐌 CAN bus system for data transfer.
Bosch module for ZF transmission The system also includes:

The highly developed transmission functions 쐌 electrically actuated pressure-control valves
with real-time calculations require a 32-bit and solenoid valves,
microcontroller (Motorola MPC555). 쐌 position sensor,
This means that up to 250 connections have to 쐌 rotational-speed sensors,
쐌 temperature sensor, and
4 Siemens-VDO electronic module 쐌 transmission connector.
The environmental conditions which

impact on the transmission pose a particular
challenge.
The temperature temporarily reaches
140°C and acceleration values of up to 30 g
occur. In addition, the electronic circuitry is
completely immersed in transmission fluid,
which contains contaminants, abrasion parti-
cles, and chemical additives (see also chapter
æ UTS0316Y
entitled “Automatic transmissions AT” sec-

tion “Automatic transmission luid/ATF”).
5 Conti-TEMIC electronic module 6 Bosch electronic module for ZF 6HP26

æ UTS0284Y
æ UTS0317Y
Modules for Transmission Control Module Types 353
Electrohydraulic Modules (EHM) Figure 7 shows an electrohydraulic module:

Expanding the components of the electronic A three-piece component carrier made of
module with the additional integration of ac- plastic holds the components integrated in
tuators results in the electrohydraulic module the module. A pressed screen, located be-
(EHM). tween the plastic components to provide pro-
tection against dirt and metal shavings, con-
An electrohydraulic module therefore tains the electrical wiring. The components in
comprises the pressed screen are contacted by laser
쐌 sensors, welding or special insulation-displacement
쐌 actuators, and connection technology.
쐌 ECU.
7 Electrohydraulic module
3 Fig. 7
1 Transmission
connector
2 Pressure-control
valves
3 Rotational-speed
æ UTS0223Y
sensor
4 Microhybrid-
technology ECU
5 Position sensor
354 Antilock braking system System overview
Antilock Braking System (ABS)

In hazardous driving conditions, it is possible System overview
for the wheels of a vehicle to lock up under
braking. The possible causes include wet or The ABS braking system is based on the com-
slippery road surfaces, and abrupt reaction ponents of the conventional system. Those
on the part of the driver (unexpected hazard). are
The vehicle can become uncontrollable as a 쐌 the brake pedal (Fig. 1, 1),
result, and may go into a slip and/or leave the 쐌 the brake booster (2),
road. The antilock braking system (ABS) de- 쐌 the master cylinder (3),
tects if one or more wheels are about to lock 쐌 the reservoir (4),
up under braking and if so makes sure that 쐌 the brake lines (5) and hoses (6), and
the brake pressure remains constant or is 쐌 the brakes and wheel-brake cylinders (7).
reduced. By so doing, it prevents the wheels
from locking up and the vehicle remains In addition there are also the following
steerable. As a consequence the vehicle can components:
be braked or stopped quickly and safely. 쐌 the wheel-speed sensors (8),
쐌 the hydraulic modulator (9), and
쐌 the ABS control unit (10).
The warning lamp (11) lights up if the ABS is

switched off.
1 Braking system with ABS
7 6 5 4 11 7 8 8 7
Fig. 1
11 Brake pedal
12 Brake booster
13 Master cylinder SB
A
14 Reservoir
15 Brake line
16 Brake hose
17 Wheel brake
with wheel-brake
cylinder
18 Wheel-speed
sensor
19 Hydraulic modulator
10 ABS control unit
(in this case,
æ UFB0697Y
attached unit fixed

8 10 9 8 7 3 2 1
onto hydraulic
modulator)
11 ABS warning lamp

Antilock braking system System overview 355
Wheel-speed sensors As the degree of brake slip increases due

The speed of rotation of the wheels is an im- to braking on a slippery surface or heavy
portant input variable for the ABS control braking, the risk of the wheels locking up
system. Wheel-speed sensors detect the speed also increases. The solenoid valves are then
of rotation of the wheels and pass the electri- switched to the “maintain pressure” setting.
cal signals to the control unit. The connection between the master cylinder
A car may have three or four wheel-speed and the brakes is shut off (inlet valve is
sensors depending on which version of the closed) so that any increase of pressure in the
ABS system is fitted (ABS system versions). master cylinder does not lead to a pressure
The speed signals are used to calculate the de- increase at the brakes.
gree of slip between the wheels and the road If the degree of slip of any of the wheels in-
surface and therefore detect whether any of creases further despite this action, the pres-
the wheels is about to lock up. sure in the brake(s) concerned must be re-
duced. To achieve this, the solenoid valves are
Electronic control unit (ECU) switched to the “pressure release” setting. The
The ECU processes the information received inlet valve is still closed, and in addition, the
from the sensors according to defined mathe- outlet valve opens to allow the return pump
matical procedures (control algorithms). The integrated in the hydraulic modulator to
results of those calculations form the basis for draw brake fluid from the brake(s) concerned
the control signals sent to the hydraulic mod- in a controlled manner. The pressure in the
ulator. relevant brake(s) is thus reduced so that
wheel lock-up does not occur.
Hydraulic modulator
The hydraulic modulator incorporates a se-
ries of solenoid valves that can open or close
the hydraulic circuits between the master
cylinder (Fig. 2, 1) and the brakes (4). In
addition, it can connect the brakes to the 2 Principle of hydraulic modulator with 2/2 solenoid
valves (schematic)
return pump (6). Solenoid valves with two
hydraulic connections and two valve posi-
1 2
tions are used (2/2 solenoid valves). The inlet Fig. 2
1 Master cylinder
valve (7) between the master cylinder and the
with reservoir
brake controls pressure application, while the 5
3 2 Brake booster
outlet valve (8) between the brake and the re- 3 Brake pedal
turn pump controls pressure release. There is 4 Wheel brake
one such pair of solenoid valves for each with wheel-brake
brake. cylinder
6 Hydraulic modulator with
7 5 Damping chamber
Under normal conditions, the solenoid valves
6 Return pump
in the hydraulic modulator are at the “pres- 7 Inlet valve
sure application” setting. That means the 8 Outlet valve
8
inlet valve is open. The hydraulic modulator 9 Brake-fluid
then forms a straight-through connection accumulator
between the master cylinder and the brakes.
9 4 Inlet valve:
Consequently, the brake pressure generated
shown in open
in the master cylinder when the brakes are
æ UFB0701Y
setting
applied is transmitted directly to the brakes at Outlet valve:
each wheel. shown in closed
setting
356 Antilock braking system Requirements placed on ABS
Requirements placed on ABS limit speed (that is the absolute maximum

speed at which a vehicle can successfully
An ABS system must meet a comprehensive negotiate a bend of a given radius with the
range of requirements, in particular all the drive disengaged).
safety requirements associated with dynamic 쐌 The system should also be able to ensure
braking response and the braking-system that handling stability and steerability are
technology. maintained and the best possible braking
effect obtained on a bumpy or uneven road
Handling stability and steerability surface, regardless of the force with which
쐌 The braking control system should be the driver applies the brakes.
capable of ensuring that the car retains 쐌 Finally, the braking control system should
its handling stability and steerability on be able to detect aquaplaning (when the
all types of road surface (from dry roads wheels “float” on a film of water) and re-
with good adhesion to black ice). spond appropriately to it. In so doing it
쐌 An ABS system should utilize the available must be able to maintain vehicle controlla-
adhesion between the tires and the road bility and course.
surface under braking to the maximum
possible degree, giving handling stability Effective range
and steerability precedence over minimiz- 쐌 The braking control system must be
ing braking distance. It should not make effective across the entire speed range of
any difference to the system whether the the vehicle right down to crawling speed
driver applies the brakes violently or grad- (minimum speed limit approximately
ually increases the braking force to the 2.5 km/h). If the wheels lock up below
point at which the wheels would lock. the minimum limit, the distance traveled
쐌 The control system should be capable before the vehicle comes to a halt is not
of adapting rapidly to changes in road-sur- critical.
face grip, e.g. on a dry road with occasional
patches of ice, wheel lock-up on the ice Timing characteristics
must be restricted to such short time spans 쐌 Adjustments to take account of braking-
that handling stability and steerability are system hysteresis (delayed reaction to
not impaired. At the same time, it should release of brakes) and the effects of the
allow maximum utilization of adhesion engine (when braking with the drive en-
where the road is dry. gaged) must take as little time as possible.
쐌 When braking under conditions where the 쐌 Pitching of the vehicle in response to sus-
amount of available grip differs between pension vibration must be prevented.
wheels (e.g. wheels on one side on ice, on
dry tarmac on the other – referred to as Reliability
“μ-split” conditions), the unavoidable yaw 쐌 There must be a monitoring circuit which
forces (turning forces around the vehicle’s continuously checks that the ABS is func-
vertical axis which attempt to turn the tioning correctly. If it detects a fault which
vehicle sideways) should only be allowed could impair braking characteristics, the
to develop at a rate slow enough for the ABS should be switched off. A warning
“average driver” to easily counteract them lamp must then indicate to the driver that
by steering in the opposite direction. the ABS is not functioning and only the
쐌 When cornering, the vehicle should retain basic braking system is available.
its handling stability and steerability under
braking and be capable of being braked to
a standstill as quickly as possible provided
its speed is sufficiently below the corner’s
Antilock braking system Dynamics of a braked wheel 357
Dynamics of a braked wheel The degree of brake slip,

(υF – υR)
Figs. 1 and 2 show the physical interdepen- λ= υF · 100%
dencies during a braking sequence with ABS.
The areas in which the ABS is active are indicates the degree to which the wheel’s
shaded blue. circumferential speed, υR, lags behind the
Curves 1, 2 and 4 in Fig. 1 relate to road vehicle’s linear speed (road speed), υF.
conditions in which the level of adhesion and From the progression shown in Fig. 1 of
therefore the braking force increases up to a curves 1 (dry conditions), 2 (wet conditions)
maximum limit as the brake pressure rises. and 4 (black ice), it is evident that shorter
Increasing the brake pressure above that braking distances are achieved with ABS than
maximum adhesion limit on a car without if the wheels are overbraked and lock up
ABS would constitute overbraking. When that (brake slip λ = 100%). Curve 3 (snow) shows
happens, the “sliding” proportion of the tire a different pattern whereby a wedge of snow
footprint (the area of the tire in contact with forms in front of the wheels when they lock
the road) increases so much as the tire de- up and helps to slow the vehicle down; in
forms that the static friction diminishes and this scenario, the advantage of ABS is in its
the kinetic friction increases. ability to maintain handling stability and
The brake slip λ is a measure of the steerability.
proportion of kinetic friction, whereby
λ = 100% is the level at which the wheel locks As the two curves for coefficient of friction,
and only kinetic friction is present. μHF, and lateral-force coefficient, μS, in Fig. 2
illustrate, the active range of the ABS has to
be extended for the large lateral slip angle of
α = 10° (i.e. high lateral force due to rapid
1 Coefficient of friction, μHF, relative 2 Coefficient of friction and lateral-force coefficient, μS,
to brake slip, λ versus brake slip, λ, and slip angle
1.2
Coefficient of friction μ HF, Lateral-force coefficient μ S
1.2
1.0 α = 2° Fig. 1
1.0 1 1 Radial tire on
μ HF dry concrete
α = 10°
Coefficient of friction μ HF
0.8 2 Cross-ply tire

0.8 on wet tarmac
2 3 Radial tire on
0.6 loose snow
0.6
4 Radial tire on
wet black ice
0.4 Blue shaded areas:
0.4
ABS active zones
α = 10°
3 Fig. 2
0.2 0.2
α = 2° μHF Coefficient
μS
4 of friction
æ UFB0360-1E
æ UFB0362-2E
0 0 μS Lateral-force
0 20 40 60 80 % 0 20 40 60 80 %
coefficient
Brake slip λ Brake slip λ
α Slip angle
Blue shaded areas:
ABS active zones
358 Antilock braking system Dynamics of a braked wheel, ABS control loop
lateral acceleration of the vehicle) as com- ABS control loop

pared with the smaller lateral slip angle of
α = 2°, that is to say, if the vehicle is braked Overview
hard in a corner when the lateral acceleration The ABS control loop (Fig. 1) consists of the
is high, the ABS cuts in sooner and allows an following:
initial level of brake slip of, say, 10%. At α =
10°, an initial coefficient of friction of only The controlled system
μHF = 0.35 is obtained, while the lateral-force 쐌 The vehicle and its brakes
coefficient is almost at its maximum at μS = 쐌 The wheels and the friction pairing of tire
0.80. and road surface
As the vehicle’s speed and therefore its
lateral acceleration are reduced by braking The external variables affecting the
through the corner, the ABS is able to allow control loop:
increasingly higher levels of brake slip so that 쐌 Changes in the adhesion between the tires
the deceleration increases, while the lateral- and the road surface caused by different
force coefficient diminishes as the lateral types of road surface and changes in the
acceleration reduces. wheel loadings, e.g. when cornering
쐌 Irregularities in the road surface causing
Braking while cornering causes the braking the wheels and suspension to vibrate
forces to rise so quickly that the overall brak- 쐌 Lack of circularity of the tires, low tire
ing distance is only marginally longer than pressure, worn tire tread, differences in
when braking in a straight line under identi- circumference between wheels, (e.g. spare
cal conditions. wheel)
쐌 Brake hysteresis and fade
쐌 Differences in master-cylinder pressure
between the two brake circuits
1 ABS control loop The controllers

쐌 The wheel-speed sensors
쐌 The ABS control unit
ABS ECU
The controlled variables
Calculate 쐌 The wheel speed and, derived from it, the
Con- (actual slip)
trolled wheel deceleration,
variable
Regulate 쐌 The wheel acceleration and the brake slip
(required slip) Monitor 6
The reference variable:

쐌 The foot pressure applied to the brake
Hydraulic modulator pedal by the driver, amplified by the brake
with solenoid valves
Act on brak- booster, and generating the brake pressure
5
ing pressure in the braking system
Fig. 1 4
1 Brake pedal
Manipulated variable Build up brak-
2 Brake booster Controlled system ing pressure The correcting variable:
3 Master cylinder 쐌 Braking pressure in the wheel-brake
with reservoir 3
Disturbance values Pedal cylinder.
4 Wheel-brake (e.g. road condition) force
æ UFB0689E
cylinder
5 Wheel-speed 2 1
sensor
6 Warning lamp
Antilock braking system ABS control loop 359
Controlled system zone, the wheel deceleration is limited to a

The data-processing operations performed by small rate, whereas in the unstable zone it
the ABS control unit are based on the increases rapidly. As a consequence, the curve
following simplified controlled system: for friction coefficient versus brake slip re-
쐌 a non-driven wheel, veals opposite characteristics in the stable
쐌 a quarter of the vehicle’s mass apportioned and unstable zones. The ABS exploits those
to that wheel, opposing characteristics.
쐌 the brake on that wheel and, representing
the friction pairing of tire and road surface,
쐌 a theoretical curve for coefficient of fric-
tion versus brake slip (Fig. 2).
2 Ideal curve for friction coefficient versus slip
That curve is divided into a stable zone with
a linear gradient and an unstable zone with
a constant progression (μHFmax). As an addi-
tional simplification, there is also an assumed
Coefficient of friction μ HF
initial straight-line braking response that is μ HFmax

equivalent to a panic braking reaction.
Fig. 3 shows the relationships between Fig. 2

braking torque, MB (the torque that can be a Stable zone
a b
b Unstable zone
generated by the brake through the tire), or
æ UFB0361-1E
λK Optimum
road-surface frictional torque, MR (torque
0 λK 100 % brake slip
that acts against the wheel through the fric-
Brake slip λ μHFmax Maximum
tion pairing of tire and road surface), and coefficient
time, t, as well as the relationships between of friction
the wheel deceleration (a) and time, t,
whereby the braking torque increases in lin- 3 Initial braking phase, simplified
ear fashion over time. The road-surface fric-
Road-surface frictional torque M R
tional torque lags slightly behind the braking MB

torque by the time delay, T, as long as the
braking sequence is within the stable zone of
the curve for friction coefficient versus brake
Braking torque M B
slip. After about 130 ms, the maximum level

MB – MR
(μHFmax) – and therefore the unstable zone – M Rmax
of the curve for friction coefficient versus
brake slip is reached. From that point on, T
the curve for friction coefficient versus brake
slip states that while the braking torque, MB, MR Fig. 3
continues to rise at an undiminished rate, (–a) Wheel
the road-surface frictional torque, MR, cannot deceleration
increase any further and remains constant. (–amax) Maximum wheel
deceleration –a
In the period between 130 and 240 ms (this deceleration

is when the wheel locks up), the minimal MB Braking torque
MR Road-surface
torque difference, MB – MR, that was present
Wheel
æ UFB0363-1E
– a max frictional torque

in the stable zone rises rapidly to a high fig-
MRmax Maximum road-
ure. That torque difference is a precise mea- surface frictional
0 100 200 ms
sure of the wheel deceleration (–a) of the torque
Time t
braked wheel (Fig. 3, bottom). In the stable T Time lag
360 Antilock braking system ABS control loop
Controlled variables Controlled variables for non-driven wheels

An essential factor in determining the effec- The wheel acceleration and deceleration are
tiveness of an ABS control system is the generally suitable as controlled variables for
choice of controlled variables. The basis driven and non-driven wheels provided the
for that choice are the wheel-speed sensor sig- driver brakes with the clutch disengaged. The
nals from which the ECU calculates the decel- reason can be found in the opposing charac-
eration/acceleration of the wheel, brake slip, teristics of the controlled system in the stable
the reference speed and the vehicle decelera- and unstable zones of the curve for friction
tion. On their own, neither the wheel deceler- coefficient versus brake slip.
ation/acceleration nor the brake slip are suit- In the stable zone, the wheel deceleration is
able as controlled variables because, under limited to relatively low rates so that when the
braking, a driven wheel behaves entirely dif- driver presses harder on the brake pedal, the
ferently to a non-driven wheel. However, by car brakes harder without the wheels locking
combining those variables on the basis of ap- up.
propriate logical relationships, good results In the unstable zone, on the other hand, the
can be obtained. driver only needs to apply slightly more pres-
As brake slip is not directly measurable, the sure to the brake pedal to induce instanta-
ECU calculates a quantity that approximates neous wheel lock-up. This characteristic
to it. The basis for the calculation is the refer- means that very often the wheel deceleration
ence speed, which represents the speed under and acceleration can be used to determine the
ideal braking conditions (optimum degree of degree of brake slip for optimum braking.
brake slip). So that speed can be determined, A fixed wheel deceleration threshold
the wheel-speed sensors continuously trans- for initiation of the ABS control sequence
mit signals to the ECU for calculating the should only be fractionally above the maxi-
speed of the wheels. The ECU takes the signals mum possible vehicle linear deceleration.
from a pair of diagonally opposed wheels (e. g. This is particularly important if the driver
right front and left rear) and calculates the ref- initially only applies the brakes lightly but
erence speed from them. Under partial brak- then increasingly applies more pressure to
ing, the faster of the two diagonally opposite the pedal. If the threshold is set too high,
wheels generally determines the reference the wheels could then progress too far along
speed. If the ABS cuts in under emergency the curve for friction coefficient versus slip
braking, the wheel speeds will be different into the unstable zone before the ABS detects
from the vehicle’s linear speed and can thus the imminent loss of control.
not be used for calculating the reference speed When the fixed threshold is initially
without adjustment. During the ABS control reached under heavy braking, the brake
sequence, the ECU provides the reference pressure at the wheel(s) concerned should
speed based on the speed at the start of the not automatically be reduced, because with
control sequence and reduces it at a linear modern tires, valuable braking distance
rate. The gradient of the reference-speed would be lost on a surface with good grip,
graph is determined by analyzing logical especially in cases where the initial speed
signals and relationships. is high.
If, in addition to the wheel acceleration/
deceleration and the brake slip, the vehicle’s
linear deceleration is brought into the equa-
tion as an additional quantity, and if the logi-
cal circuit in the ECU is modulated by com-
putation results, then ideal brake control can
be achieved. That concept has been realized in
the Bosch Antilock Braking System (ABS).
Antilock braking system ABS control loop 361
Controlled variables for driven wheels Effectiveness of control

If first or second gear is engaged when the An efficient antilock braking system must
brakes are applied, the engine acts on the dri- meet the criteria listed below for the standard
ven wheels and substantially increases their of control.
effective mass moment of inertia ΘR, i. e. the 쐌 Maintenance of handling stability by
wheels behave as if they were considerably provision of adequate lateral forces at the
heavier. The sensitivity with which the wheel rear wheels
deceleration responds to changes in the 쐌 Maintenance of steerability by provision of
braking torque in the unstable zone of the adequate lateral forces at the front wheels
curve for friction coefficient versus brake slip 쐌 Reduction of braking distances as com-
diminishes to an equal extent. pared with locked-wheel braking by opti-
The starkly opposing characteristics dis- mum utilization of the available adhesion
played by non-driven wheels in the stable and between tires and road surface
unstable zone of the curve for friction coeffi- 쐌 Rapid adjustment of braking force to dif-
cient versus brake slip are evened out to a ferent friction coefficients, for instance
substantial degree, so that in this situation the when driving through puddles or over
wheel deceleration is often insufficient as a patches of ice or compacted snow.
controlled variable for identifying the degree 쐌 Ensuring application of low braking-
of brake slip offering the greatest possible torque control amplitudes to prevent
friction. It is necessary instead to introduce as suspension vibration
an additional controlled variable a quantity 쐌 High degree of user-friendliness due to
that approximates to brake slip and to com- minimal pedal feedback (“pedal judder”)
bine it in a suitable manner with the wheel and low levels of noise from the actuators
deceleration. (solenoid valves and return pump in the
hydraulic modulator).
Fig. 4 compares an initial braking sequence
on a non-driven wheel and on a driving The criteria listed can only be optimized
wheel that is connected to the drivetrain. collectively rather than individually. Never-
In this example, the engine’s inertia increases theless, vehicle handling stability and steer-
the effective wheel inertia by a factor of four. ability are always among the top priorities.
On the non-driven wheel, a specific threshold
for deceleration (–a)1 is exceeded very soon
after leaving the stable zone of the curve for 4 Initial braking phase for a non-driven wheel and
a driving wheel connected to the drivetrain
friction coefficient versus brake slip. Because
it has a moment of inertia that is four times
greater, the driven wheel requires a torque
Road-surface frictional torque MR,
difference four times as big (– a) 2

MB
ΔM2 = 4 · ΔM1
(– a)1 ΔM2 = 4 • ΔM1 Fig. 4
ΔM1 Index 1: non-driven
Braking torque MB
to exceed the threshold (–a)2. Consequently, wheel

the driven wheel may by then have progressed Index 2: driven wheel
a long way into the unstable zone of the curve (in this example, the
for friction coefficient versus brake slip, re- wheel moment of
MR1 inertia is increased
sulting in impaired vehicle handling stability.
æ UFB0364-1E
MR2 by a factor of 4)
(–a) Threshold for wheel
t1 t2
deceleration
Time t
M Torque difference
MB – MR
362 Antilock braking system Typical control cycles
Typical control cycles As a result, the relevant solenoid valve

switches to the “maintain pressure” setting.
Control cycle on surfaces with good grip At this point the brake pressure must not be
(High coefficient of friction) reduced, because the threshold (–a) might be
If the ABS sequence is activated on a road exceeded within the stable zone of the curve
surface with good grip (high coefficient of for friction coefficient versus brake slip and
friction), the subsequent pressure rise must then potential braking distance would be
be 5 to 10 times slower than in the initial “wasted”. At the same time, the reference
braking phase in order to prevent undesirable speed, υRef, reduces according to a defined
suspension vibration. That requirement pro- linear gradient. The reference speed is used as
duces the control-cycle progression for high the basis for determining the slip switching
coefficients of friction illustrated in Fig. 1. threshold, λ1.
During the initial phase of braking, the brake At the end of phase 2, the wheel speed, υR,
pressure at the wheel and the rate of wheel drops below the λ1 threshold. At that point,
deceleration (negative acceleration) rise. At the solenoid valves switch to the “reduce
the end of phase 1, the wheel deceleration pressure” setting so that the pressure drops,
passes the set threshold level (–a). and they remain at that setting as long as the
1 Braking control cycle for high-adhesion conditions

Vehicle speed υ
υF
υ Ref
λ1
υR
+A
Wheel acceleration
+a
0
–a
Fig. 1
υF Vehicle speed
υRef Reference speed Phase 1 2 3 4 5 6 7 8
υR Wheel speed
Brake pressure p
λ1 Slip switching
threshold –Δp ab
Switching signals:
+A, + a Thresholds for
æ UFB0365-1E
wheel acceleration
–a Threshold for
wheel deceleration Time t
–Δ␳ab Brake-pressure
drop
Antilock braking system Typical control cycles 363
2 Braking sequence without ABS

Vehicle speed υF
Wheel speed υ R
Wheel locks
Vehicle no longer steerable
Brake pressure p
æ UFB0699E
Time t
wheel deceleration is above the threshold In phase 6, the brake pressure is once again
(–a). held constant because the acceleration is
above the threshold (+a). At the end of this
At the end of phase 3, the deceleration rate phase, the wheel acceleration falls below the
falls below the threshold (–a) again and a threshold (+a). This is an indication that the
pressure-maintenance phase of a certain wheel has returned to the stable zone of the
length follows. During that period, the wheel curve for friction coefficient versus brake slip
acceleration has increased so much that the and is now slightly underbraked.
threshold (+a) is passed. The pressure contin-
ues to be maintained at a constant level. The brake pressure is now increased in
stages (phase 7) until the wheel deceleration
At the end of phase 4, the wheel acceleration passes the threshold (–a) (end of phase 7).
exceeds the relatively high threshold level This time, the brake pressure is reduced
(+A). The brake pressure then increases for as immediately without a λ1 signal being
long as the acceleration remains above the generated.
threshold (+A).
By comparison, Fig. 2 shows the progressions
for an emergency braking sequence without
ABS.
Control cycle on slippery surfaces threshold, the brake pressure is reduced over
(low coefficient of friction) a short, fixed period.
In contrast with good grip conditions, on A further short pressure-maintenance
a slippery road surface even very light pres- phase follows. Then, once again, the wheel
sure on the brake pedal is frequently enough speed is compared to the slip switching
to make the wheels lock up. They then re- threshold λ1 and, as a consequence, the
quire much longer to emerge from a high- pressure reduced over a short, fixed period.
slip phase and accelerate again. The process- In the following pressure-maintenance
ing logic of the ECU detects the prevailing phase, the wheel starts to accelerate again
road conditions and adapts the ABS re- and its acceleration exceeds the threshold
sponse characteristics to suit. Fig. 3 shows (+a). This results in another pressure-main-
a typical control cycle for road conditions tenance phase which lasts until the accelera-
with low levels of adhesion. tion drops below the threshold (+a) again
In phases 1 to 3, the control sequence is (end of phase 5). In phase 6, the incremental
the same as for high-adhesion conditions. pressure-increase pattern seen in the preced-
Phase 4 starts with a short pressure-maining section takes place again until, in phase 7,
tenance phase. Then, within a very short pressure is released and a new control cycle
space of time, the wheel speed is compared starts.
with the slip switching threshold λ1. As the
wheel speed is lower than the slip switching
3 Braking control cycle for low-adhesion conditions
υF
Vehicle speed υ
υ Ref
λ1
υR
acceleration signals
+a
Slip/wheel
0
-a
-λ
Fig. 3
υF Vehicle speed
Phase 1 2 3 4 5 6 7 8 9
υRef Reference speed
υR Wheel speed
Brake pressure p
λ1 Slip switching
threshold
Switching signals:
+a Threshold for -Δpab
æ UFB0366-2E
wheel acceleration
–a Threshold for
wheel deceleration Time t
–␳ab Brake-pressure
drop
In the cycle described above, the control logic Control cycle with yaw-moment
detected that following pressure release – buildup delay
triggered by the signal (–a) – two more When the brakes are applied in situations
pressure-reduction stages were necessary to where the grip conditions differ significantly
induce the wheel to accelerate again. between individual wheels (“μ-split” condi-
The wheel remains in the higher-slip zone for tions) – for example, if the wheels on one side
a relatively long period, which is not ideal for of the car are on dry tarmac while those on
handling stability and steerability. the other side are on ice – vastly different
In order to improve those two characteris- braking forces will be produced at the front
tics, this next control cycle and those that wheels (Fig. 4). That difference in braking
follow incorporate continual comparison of force creates a turning force (yaw moment)
wheel speed with the slip switching threshold around the vehicle’s vertical axis. It also gener-
λ1. As a consequence, the brake pressure is ates steering feedback effects of varying types
continuously reduced in phase 6 until, in dependent on the vehicle’s kingpin offset.
phase 7, the wheel acceleration exceeds the With a positive kingpin offset, corrective
threshold (+a). Because of that continuous steering is made more difficult, while a nega-
pressure release, the wheel retains a high level tive kingpin offset has a stabilizing effect.
of slip for only a short period so that vehicle
handling and steerability are improved in Heavy cars tend to have a relatively long
comparison with the first control cycle. wheelbase and a high level of inertia around
the vertical axis. With vehicles of this type,
the yaw effect develops slowly enough for the
driver to react and take corrective steering
action during ABS braking. Smaller cars with
4 Yaw-moment build-up induced by areas of short wheelbases and lower levels of inertia,
widely differing adhesion
on the other hand, require an ABS system
supplemented by a yaw-moment buildup
delay (GMA system) to make them equally
controllable under emergency braking in
conditions where there are wide differences
in grip between individual wheels. Develop-
1 2 ment of the yaw moment can be inhibited by
delayed pressure increase in the brake on the
FB2 front wheel that is on the part of the road
FB1
offering the higher level of adhesion (the
“high” wheel).
MYaw
Fig. 5 (overleaf) illustrates the principle of
the yaw-moment buildup delay:
Curve 1 shows the brake pressure, p, in the
master cylinder. Without yaw-moment
buildup delay, the wheel on the tarmac
quickly reaches the pressure phigh (curve 2)
and the wheel on the ice, the pressure plow
(curve 5). Each wheel is braked with the
æ UFB0290-1E
Fig. 4
specific maximum possible deceleration
μ HF1 = 0.8 μ HF2 = 0.1 Myaw Yaw moment
(individually controlled). FB Braking force
1 “High” wheel
2 “Low” wheel
GMA 1 system GMA 2 system

On vehicles with less extreme handling charac- The GMA 2 system is used on vehicles with
teristics, the GMA 1 system is used. With this more extreme handling characteristics. With
system, during the initial phases of braking this system, as soon as the brake pressure at
(curve 3), as soon as the pressure is reduced for the “low” wheel is reduced, the ABS solenoid
the first time at the “low” wheel because this valves for the “high” wheel are directed to
shows a tendency to lock up, the brake pres- maintain and then reduce the pressure for
sure at the “high” wheel is increased in stages. a specific period (Fig. 5, curve 4). Renewed
Once the brake pressure at the “high” wheel pressure increase at the “low” wheel then trig-
reaches its lock-up point, it is no longer af- gers incremental increase of pressure at the
fected by the signals from the “low” wheel, and “high” wheel, though with pressure-increase
is individually controlled so that it is able to periods a certain amount longer than for the
utilize the maximum possible braking force. “low” wheel. This pressure metering takes
This method gives the type of vehicle referred place not only in the first control cycle but
to satisfactory steering characteristics under throughout the braking sequence.
emergency braking on surfaces offering un-
equal grip to individual wheels. The effect of the yaw moment on steering
As the maximum brake pressure at the “high” characteristics is all the more exaggerated the
wheel is reached within a relatively short greater the speed of the vehicle when the
time (750 ms), the braking distance is only brakes are first applied. The GMA 2 system
marginally longer than for vehicles without a divides the vehicle speed into four ranges.
yaw-moment buildup delay facility. In each of those ranges, the yaw moment
is inhibited to differing degrees. In the high
speed ranges, the pressure-increase periods at
5 Progression of brake pressure/steering angle with the “high” wheel are progressively shortened,
yaw moment build-up delay
while the pressure-decrease periods at the
“low” wheel are progressively lengthened in
order to achieve effective inhibition of yaw-
moment generation at high vehicle speeds.
1 Fig. 5 below compares the steering angle pro-
gression necessary for maintaining a straight
Brake pressure p
course under braking without GMA (curve 6)

2 and with GMA (curve 7).
Another important aspect with regard
3 to GMA usage is the vehicle’s response to
braking in a corner. If the driver brakes when
4
Fig. 5 cornering at a high speed, the GMA increases
1 Pressure, pHz, in 5 the dynamic load on the front wheels while
the master cylinder
reducing it at the rear. As a result, the lateral
2 Brake pressure,
phigh, without GMA
forces acting on the front wheels become
stronger while those acting on the rear wheels
Steering angle α
3 Brake pressure,
phigh, with GMA 1 diminish. This generates a turning moment
4 Brake pressure, 6 towards the inside of the bend as a result of
phigh, with GMA 2 which the vehicle slews off its line towards the
5 Brake pressure, 7 inside of its intended course and is very diffi-
æ UFB0291-1E
plow
cult to bring back under control (Fig. 6a).
6 Steering angle, α,
without GMA Time t
7 Steering angle, α,
with GMA
In order to avoid this critical response to The ideal method of inhibiting yaw-moment
braking, the GMA also takes the lateral accel- buildup involves a compromise between good
eration into account. The GMA is deactivated steering characteristics and suitably short
at high lateral acceleration rates. As a result, a braking distance and is developed by Bosch
high braking force is generated at the outside individually for a specific vehicle model
front wheel during the initial phase of brak- through consultation with the manufacturer.
ing in a corner and creates a turning moment
towards the outside of the bend. That turning
moment balances out the turning moment
acting in the opposite direction that is pro-
duced by the lateral forces, so that the vehicle
slightly understeers and thus remains easily
controllable (Fig. 6).
6 Response to braking when cornering at critical speeds with/without GMA
a
FB1
FS1
M1
b
FB2
FS2
Fig. 6
a GMA activated (no
M2
individual control):
vehicle oversteers
b GMA deactivated
(individual control):
æ UFB0367-1Y
vehicle slightly
understeers
FB Braking force
FS Lateral force
M Torque
Control cycle for four-wheel-drive (whereby the wheel with the lower degree of
vehicles adhesion, μHF, determines the brake pressure
The most important criteria for assessing the for both rear wheels) is thus canceled, and
various types of four-wheel-drive configura- both rear wheels utilize the maximum brak-
tion (Fig. 7) are traction, dynamic handling, ing force. As soon as the inter-axle lock is en-
and braking characteristics. As soon as differ- gaged, the system forces the front wheels to
ential locks are engaged, conditions are cre- assume the same average speed as the rear
ated that demand a different response from wheels. All four wheels are then dynamically
the ABS system. interlinked and the engine drag (engine brak-
ing effect when backing off the throttle) and
When a rear-axle differential is locked, the inertia act on all four wheels.
rear wheels are rigidly interconnected, i.e.
they always rotate at the same speed and re- In order to ensure optimum ABS effectiveness
spond to the braking forces (at each wheel) under those conditions, additional features
and friction levels (between each tire and the have to be incorporated according to the type
road surface) as if they were a single rigid of four-wheel-drive system (Fig. 7) in use.
body. The “select low” mode that would
otherwise be adopted for the rear wheels
7 Four-wheel-drive configurations
a b c
2 1
1 1
2 2
Fig. 7
a Four-wheel-drive 3
6
system 1
b Four-wheel-drive 4
system 2
c Four-wheel-drive
system 3
1 Engine
2 Transmission
3 Freewheel and
viscous clutch
Differential with
4 Manual lock or
4 5 7
æ UFB0368-1Y
viscous lock
5 Proportional lock
6 Automatic clutch
and lock
7 Automatic lock
Four-wheel-drive system 1 When analyzing braking on such surfaces,

On four-wheel-drive system 1 with manual the response threshold (–a) for wheel deceler-
locks or permanently active viscous locks in ation is halved and the diminishing rise
the propeller shaft and the rear-axle differen- in the reference speed is limited to specific,
tial, the rear wheels are rigidly interconnected relatively low levels. As a result, imminent
and the average speed of the front wheels is wheel lock-up can be detected early and
the same as that of the rear wheels. As already “sensitively”.
mentioned, the rear differential lock results in On four-wheel-drive vehicles, heavy
deactivation of the “select low” mode for the application of the throttle on slippery road
rear wheels and utilization of the maximum surfaces can cause all four wheels to spin.
braking force at each rear wheel. When brak- In such situations, special signal-processing
ing on road surfaces with unequal levels of methods ensure that the reference speed can
grip at the two rear wheels, this can generate a only increase in response to the spinning
yaw moment with a potentially critical effect wheels within the limits of the maximum
on vehicle handling stability. If the maximum possible vehicle acceleration. In a subsequent
braking force difference were also applied braking sequence, the initial ABS pressure
very quickly at the front wheels, it would not reduction is triggered by a signal (–a) and a
be possible to keep the vehicle on a stable specific, minimal difference in wheel speed.
course.
This type of four-wheel-drive arrangement Four-wheel-drive system 2
therefore requires a GMA system for the front Because of the possibility of all four wheels
wheels in order to ensure handling stability spinning with four-wheel-drive system 2
and steerability are maintained in road condi- (viscous lock with freewheel in the propeller
tions where there is a significant divergence shaft, proportional rear-axle differential
between the levels of grip at the left and right lock), the same special procedures must
wheels. In order to maintain ABS effective- be adopted for signal processing.
ness on slippery surfaces, the engine drag Other modifications for ensuring ABS
effect – which of course acts on all four effectiveness are not necessary as the free-
wheels on a four-wheel-drive vehicle – has to wheel disengages the wheels under braking.
be reduced. This is done by using an engine However, the system can be improved by the
drag-torque control system that applies just use of an engine drag-torque control system.
enough throttle to counteract the unwanted
engine braking effect. Four-wheel-drive system 3
Another factor that demands refinement of As with the first two systems, four-wheel-
the ABS control cycle in order to prevent drive system 3 (automatic differential/inter-
wheel lock-up is the reduced sensitivity of the axle locks) requires adoption of the signal
wheels to changes in road-surface adhesion processing procedures described above in the
on slippery surfaces that is caused by the event of wheel-spin at all four wheels.
effect of the engine inertia. The fact that all In addition, automatic release of the differen-
four wheels are dynamically linked to the ential locks whenever the brakes are applied is
gine’s inertial mass therefore requires addi- necessary. Other modifications for ensuring
tional analysis and processing operations on ABS effectiveness are not necessary.
the part of the ABS control unit. The vehicle
longitudinal deceleration is calculated to
detect smooth or slippery road surfaces where
μHF is less than 0.3.
370 Traction control system Tasks, function description
Traction Control System (TCS)

Critical driving situations can occur not only lating the optimum slip (see μ-slip curve in
while braking, but also whenever strong lon- “Basic principles of vehicle dynamics”).
gitudinal forces should be transferred at the The upper limit here is, of course, set by the
contact area between the tire and the ground. traction requirement stipulated by the driver.
This is because the transferable lateral forces
are reduced by this process. Critical situa-
tions can also occur when starting off and Function description
accelerating, particularly on a slippery road
surface, on hills, and when cornering. These Unless otherwise stated, all the following de-
kinds of situations can overtax the driver not scriptions refer to single-axle driven vehicles
only causing him/her to react incorrectly but (Fig. 1). It makes no difference whether the
also causing the vehicle to become unstable. vehicle is rear-wheel or front-wheel drive.
The traction control system (TCS) solves
these problems, providing the vehicle Drive slip and what causes it
remains within the physical limits. If the driver presses the accelerator when the
clutch is engaged, the engine torque will rise.
Tasks The drive axle torque MKar also increases. This
torque is distributed to both driven wheels in
The antilock braking system (ABS) prevents a ratio of 50 : 50 via the transversal differential
the wheels from locking up when the brakes (Fig. 1). If this increased torque can be trans-
are applied by lowering the wheel brake pres- ferred completely to the road surface, the
sures. The traction control system (TCS) pre- vehicle will accelerate unhindered. However, if
vents the wheels from spinning by reducing the drive torque MKar/2 at one driven wheel
the drive torque at each driven wheel. TCS exceeds the maximum drive torque that can be
therefore provides a logical extension of ABS transferred, the wheel will spin. The transfer-
during acceleration. able motive force is therefore reduced and the
In addition to this safety-relevant task vehicle becomes unstable due to the loss of
Fig. 1 of ensuring the stability and steerability lateral stability.
1 Engine with of the vehicle when accelerating, TCS also
transmission improves the traction of the vehicle by regu-
2 Wheel
3 Wheel brake 1 Drive concept of a single-axle driven vehicle with TCS
4 Transversal
differential
5 Control unit with
TCS functionality
VR HR
2 MStr,R
Engine, transmission,
gear ratio of differential MBr,R 3
υR
and losses are combined 1
in one unit MR = MKar/2
MKar , υKar
4
MKar Drive axle torque
υKar Drive axle speed 5 ML = MKar/2
MBr Braking torque
υL
MStr Torque transferred
to the road MBr,L MStr,L
Engine intervention
υ Wheel speed
æ UFB0757E
VL HL
V Front
H Rear
R Right
L Left

Traction control system Function description 371
The TCS regulates the slip of the driven The last two interventions are rapid
wheels as quickly as possible to the optimum interventions, the first a slower means
level. To do this the system first determines of intervention (Fig. 2). The availability of
a reference value for the slip. This value these interventions depends on the vehicle
depends on a number of factors which are in- manufacturer and engine version.
tended to represent the current driving situa-
tion as closely as possible. These factors in- In diesel-engine vehicles, the drive torque
clude: MKar is influenced by the electronic diesel
쐌 the basic characteristic for TCS reference control system (EDC) (reduction of the
slip (based on the slip requirement of a tire quantity of fuel injected).
during acceleration),
쐌 effective coefficient of friction, The braking torque MBr can be regulated
쐌 external tractive resistance (deep snow, for each wheel via the braking system.
rough road, etc.), The TCS function requires the original ABS
쐌 yaw velocity, lateral acceleration, and steer- hydraulic system to be expanded because
ing angle of the vehicle. of the need for active pressure build-up
(see also “Hydraulic modulator”).
TCS interventions
The measured wheel speeds and the respective Fig. 2 compares the response times with
drive slip can be influenced by changing the various TCS interventions. The figure
torque balance MGes at each driven wheel. shows that exclusive drive torque regulation
The torque balance MGes at each wheel results by means of the throttle valve can be unsatis-
from the drive torque MKar/2 at this wheel, the factory due to the relatively long response
respective braking torque MBr and the road time.
torque MStr (Fig. 1).
MGes = MKar/2 + MBr + MStr

(MBr and MStr are negative here.)
This balance can obviously by influenced by

the drive torque MKar provided by the engine
as well as by the braking torque MBr. Both
these parameters are therefore correcting
variables of the TCS which can be used to
regulate the slip at each wheel to the reference 2 Comparison of response times with various TCS
interventions
slip level.
In gasoline-engine vehicles, the drive
torque MKar can be controlled using the
following engine interventions:
쐌 Throttle valve (throttle valve adjustment),
쐌 Ignition system (ignition-timing advance),
Control variance
쐌 Fuel-injection system (phasing out

individual injection pulses). Fig. 2
1 Throttle-valve/wheel
1 2 3 brake intervention
æ UFB0382-1E
2 Throttle-valve
intervention/
TCS response time ignition adjustment
3 Throttle-valve
intervention
372 Traction control system Structure of traction control system
Structure of traction control driven wheel is only noticeable at first in the

torque balance of this wheel. The brake appli-
system (TCS) cation has basically the same effect as an
The expanded ABS hydraulic system allows asymmetric distribution ratio of the transver-
both symmetrical brake application (i. e. sal differential (but applied to a drive torque
brake application at both driven wheels) MKar that is reduced by the asymmetric brak-
as well as individual brake application. This is ing torque). The differential speed controller
the key to further structuring of the TCS, i.e. is also referred to as the transversal differential
structuring according to controlled variable lock controller because it can be used to influ-
rather than according to the actuator (en- ence to a certain extent the distribution ratio
gine/brake). of the transversal differential, i. e. to mimic the
effect of a differential lock.
Drive axle speed controller
The drive axle speed υKar or the drive torque Together the drive axle speed controller and
MKar can be influenced by means of engine transversal differential lock controller form
interventions. Symmetric brake applications the TCS system (Fig. 3). The drive axle speed
also influence the drive axle speed υKar and controller uses the drive axle speed υKar to
effect torque balance between the individual regulate the drive torque MKar provided by
wheels in the same way as reducing the drive the engine. The transversal differential lock
torque MKar. The drive axle speed controller controller functions primarily like a con-
is used to regulate the drive axle speed in this troller that uses the differential speed υDif
way. to regulate the distribution ratio ML to MR of
the transversal differential and therefore the
Transversal differential lock controller distribution of the drive torque MKar to the
Asymmetric brake application (brake driven wheels.
application at just one driven wheel) is used
primarily to regulate the differential speed
at the driven axle υDif = υL – υR. This task is
carried out by the differential speed controller.
Asymmetric brake application at just one
3 TCS controller concept for a single-axle-driven vehicle (rear-wheel drive)
VR HR
1
υR, MR
2
Transversal
Engine MKar , υKar differential
Fig. 3
1 Wheel Drive axle Transversal
speed differential
2 Wheel brake controller lock controller
υR, υL Wheel speeds

υKar Drive axle speed υL, ML
æ UFB0758E
V Front VL HL
H Rear
R Right
L Left
Traction control system Typical control situations 373
Typical control situations FBr + Fl is also transferred (characteristic

of the differential). Since the brakes are not
Split: Transversal differential lock applied on this side, the entire force can be
controller used as drive force FBr* + Fl (FBr* results
Fig. 4 shows a typical situation (“μ-split”) from FBr taking the different effective radii
whereby the transversal differential lock con- into account). Overall the drive force trans-
troller of the TCS becomes active when the ferred is increased by FBr* (the drive torque
vehicle pulls away after being stationary. The MKar must of course also be increased ac-
left side of the vehicle is on a slippery road cordingly). This ability of the transversal
surface with a low coefficient of friction μl differential lock controller to increase the
(“l” for low) and the right side of the vehicle traction is part of the traction control sys-
is on dry asphalt with a considerably higher tem (TCS).
coefficient of friction μh (“h” for high). The drive torque can be regulated to a
Without brake application by the trans- maximum possible drive force. The value
versal differential lock controller only the of μh represents a physical upper limit.
drive force Fl could be transferred on both When both driven wheels run synchro-
sides since the differential distributes the nously again (υDif = 0), the single-sided
drive torque equally on both sides. A greater braking force FBr or the corresponding brak-
drive torque MKar causes the wheel on the ing torque MBr is reduced again (Fig. 5).
side with μl to spin and leads to a differential
speed of υDif > 0 (see also Fig. 5). In this The exact buildup and reduction of MBr de-
case, the excess drive torque is lost as lost pends on the internal implementation of the
torque in the differential, engine, and trans- transversal differential lock controller (the
mission. controller functions like a PI-controller).
To prevent the wheel on the side of the
vehicle with μl from spinning if the drive Low: Drive axle speed controller
torque is too high, the braking force FBr If both driven wheels are on a slippery road
is applied (Fig. 4, see also Fig. 5). The differ- surface with a low coefficient of friction
ential can then transfer the force FBr + Fl to (e.g. the vehicle is standing on ice) when
this side (or a torque corresponding to this the vehicle pulls away, the drive axle speed
force), whereby FBr is diffused by the brake controller of the TCS becomes active. Fig. 4
MKar Drive torque
action. The drive force Fl remains as before.
FBr Braking force
On the side of the vehicle with μh the force FBr* Braking force, based
on effective radii
4 Differential lock effect due to asymmetric 5 Typical wheel speed and braking torque curves μl Low coefficient of
brake application when setting off on μ-split
friction
μh High coefficient of
friction
FBr υl υh = υFz Fl Transferable motive
υDif force on μl
Fh Transferable motive
Speed
FL force on μh
μl
M Kar
Fig. 5
υ Wheel speed
Fl μh MBr,l
æ UFB0383-1Y
MBr Braking torque

Braking
FBr*
æ UFB0765E
torque
MBr,h l Low-μ wheel

Fh h High-μ wheel
Time t
υFz Vehicle speed
υDif Differential speed
374 Traction control system Typical control situations, traction control system for four wheel drive vehicles
If the driver increases the driver-specified Traction control system (TCS)

torque MFahVorga, the drive torque MKar in-
creases almost simultaneously. Both driven
for four wheel drive vehicles
wheels then spin at almost the same speed. In recent years four wheel drive vehicles have
The differential speed υDif = υL – υR is ap- continued to increase in popularity. Amongst
proximately 0, while the drive axle speed these types of vehicles, Sport Utility Vehicles,
υKar = (υL + υR)/2 = υL = υR is considerably or SUVs, are the most popular of all. These
greater than a reasonable reference value are road vehicles with off-road characteristics.
υSoKar determined by the TCS due to the If all four wheels of a vehicle are to be
spinning driven wheels. The drive axle driven, both a second transversal differential
speed controller reacts by reducing the drive and an additional longitudinal differential
torque MKar to a level below the torque speci- (also called a central differential) are required
fied by the driver MFahVorga and by initiating (Fig. 7). The first task of the longitudinal dif-
brief, symmetric brake application MBr, Sym ferential is to compensate for the differences
(Fig. 6). As a result the drive axle speed υKar is between the drive axle speed of the front and
reduced and thus the speed of the spinning rear axle υKar, VA and υKar, HA, respectively.
wheels. The vehicle begins to accelerate. Since A rigid connection would result in tension
the optimum point of the μ-slip curve (see between the front and rear axle. The second
also “Basic principles of vehicle dynamics”) task is to achieve optimum distribution of the
would not be achieved without these TCS in- drive torque MKar to the two axles MKar, VA and
terventions, the acceleration would be slower MKar, HA, respectively.
while the wheels were spinning and consider- Less expensive SUVs are often equipped with
ably less lateral stability would be present. a longitudinal differential with a preset distrib-
The exact characteristic of MKar and ution ratio. Unlike with a transversal differen-
MBr, Sym depends in turn on the internal tial, fixed distribution ratios other than 50:50
implementation of the drive axle speed are useful – e.g. 60:40 for a design with em-
controller (the controller functions like phasis on the rear-wheel drive of the vehicle.
a PID-controller). Brake application by the traction control sys-
tem (TCS) can be used to mimic the behavior
of a longitudinal differential lock.
By applying the brakes to eliminate part
6 Typical wheel speed, engine and braking torque of MKar, VA, the distribution ratio MKar, HA to
curves when setting off on low-μ
MKar, VA can be increased, or can be decreased
by applying the brakes to eliminate part of
MKar, HA. The principle is the same as previ-
υSoKar
ously described for the transversal differential
TCS
υKar = υL= υR lock or transversal differential. The only dif-
Fig. 6 ference is that the braking torque of the TCS
ΔυKar
Speed
υ Wheel speed υFz

does not have to be asymmetric, (i. e. at one
υFz Vehicle speed
υKar Drive axle speed
wheel of the driven axle) but can occur sym-
υSoKar Drive axle speed MFahVorga metrically at both wheels of a driven axle.
Engine
torque
reference value Moreover, the longitudinal differential lock

MKar
MBr, Sym Symmetric controller regards the two drive axle speeds
braking torque υKar, VA and υKar, HA as input parameters rather
braking torque
MFahVorga Driver-specified
than the speed difference of the left and right
Symmetr.
drive torque
æ UFB0759E
(via accelerator
MBr, Sym wheel of the driven axle (transversal differen-
pedal position) Time t tial lock controller, see above).
L Left
R Right
Traction control system Traction control system for four wheel drive vehicles 375
Fig. 8 shows the expansion of the TCS Electronic differential locks designed as part
concept from Fig. 3 for an all-wheel drive of the TCS software have the advantage that
vehicle. As with a single-axle driven vehicle, they do not require additional hardware.
the drive axle speed controller uses the drive There are therefore very cost-efficient. They
axle speed υKar to regulate the drive torque are used for road vehicles, which is usually
MKar provided by the engine. As already de- the intended application of SUVs. When
scribed, the longitudinal differential lock used in true off-road cross-country vehicles,
controller distributes this torque to the front electronic differential locks reach their limit
and rear axle (MKar, VA and MKar, HA, respec- in tough off-road conditions, at the latest
tively). The transversal
differential lock con- 8 TCS controller concept for a four-wheel-drive vehicle
troller uses the differ-
ential speed υDif, XA to
regulate the distribu- Fig. 8
tion of the drive 1 Wheel
torque MKar, XA per axle VR HR 2 Wheel brake
to the driven wheels. 1 Drive axle
Engine υ Wheel speed
This must now be car- 2 speed controller
MVR,υVR MHR,υHR υKar Drive axle speed
ried out for both the MKar , υKar
Transversal differential Transversal differential MKar Drive axle torque
front and rear axle
Transversal differential Longitudinal Longitudinal Transversal differential
A Axle
(“X” = “V” (front) or differential differential V Front
lock controller lock controller lock controller
“X” = “H” (rear)). M ,υ M ,υ H Rear
VL VL HL HL
R Right
L Left
VL MKar, VA , υKar, VA MKar, HA , υKar, HA HL æ UFB0761E
Fig. 7
1 Engine with
transmission
2 Wheel
3 Wheel brake
7 Drive concept of a four-wheel-drive vehicle with TCS 4 Transversal
differential
5 Longitudinal
differential
6 Control unit with
VR 6 HR TCS functionality
7 Transversal
2
differential
3 M Br, VR M Br, HR
Engine, transmission,
Engine intervention gear ratios of differentials
1
and losses are combined
M Kar , υKar M Kar, HA , υKar, HA
4 5 7 into one unit
υ Wheel speed
υKar Drive axle speed
M Kar, VA , υKar, VA MBr Braking torque
M Br, VL M Br, HL R Right
æ UFB0760E
L Left
V Front
VL HL
H Rear
A Axle
376 Traction control system Traction control system for four wheel drive vehicles
when the brakes overheat. Vehicles for these Summary: Advantages of TCS
conditions are therefore often fitted with me- Below is a summary of the advantages of
chanical locks (examples can be seen in Figs. using TCS to prevent the driven wheels from
9 and 10). The lock controllers of the TCS spinning when starting off or accelerating
software are then only used as a backup sys- on slippery road surfaces under one or both
tem and they do not intervene during normal sides of the vehicle, when accelerating when
operation. cornering, and when starting off on an
incline:
쐌 Unstable vehicle conditions are avoided
and therefore the driving safety is en-
hanced.
쐌 Increased traction due to regulating the
optimum slip.
쐌 Mimicking the function of a transversal
differential lock.
쐌 Mimicking the function of a longitudinal
9 Classic solution of a differential lock differential lock with four-wheel-drive
vehicles.
쐌 Automatic control of the engine output.
쐌 No “grinding” of the tires when driving
around tight corners (unlike with mechani-
cal differential locks).
쐌 Reduction of tire wear.
쐌 Reduction of wear to drive mechanism
(transmission, differential, etc.) especially
on μ-split or if a wheel suddenly starts to
spin on a road surface providing good grip.
쐌 Warning lamp informs the driver during
æ UFB0762Y
situations close to the physical critical lim-

its.
쐌 Efficient double-use of existing ABS
hydraulic components.
쐌 Adoption of tasks of ESP vehicle dynamics
10 Electronically controllable differential lock control as subordinate wheel controller
(Haldex coupling)
(see “Closed-loop control system”).
2 4
3 5
1 6
Fig. 10
1 Output shaft
2 Working piston
æ UFB0763Y
3 Lamella
4 Axial-piston pump
5 Control valve
6 Input shaft
Control Unit Basic principles of automotive control engineering 377
왘 Basic principles of automotive control engineering
Many subsystems of a driving safety system control variance is decreased, i. e., yactual = yref
(e. g. ESP) influence the driving dynamics of at least approximately.
a vehicle by means of a controller i. e., they This task may be made more difficult by
form a control loop together with the relevant unknown natural dynamics of the controlled
components of the vehicle. system and other external factors z, which also
influence the controlled system.
Control loop
A simple standard control loop consists Example: TCS transversal differential
of controllers and a controlled system. The lock controller
objective is to influence the characteristic of The principle of a control loop can be ex-
the parameter yactual (controlled variable) of plained using the transversal differential lock
the controlled system using the controller controller of the TCS system as an example.
such that the parameter follows a reference The controlled variable yactual = υDif is the differ-
characteristic yref as closely as possible. To do ential speed of the two wheels of a driven axle.
this the controlled variable is measured and The reference value υSoDif is determined by the
passed to the controller. The actual value of TCS and adapted to the current driving situa-
the controlled variable is constantly compared tion. When driving straight ahead, this value is
with the current reference value by generating typically 0. The asymmetric braking torque is
the control variance e = yref – yactual. used as a correcting variable to influence the
The main task of the controller is to deter- controlled variable. The controlled system is
mine a suitable value for the correcting vari- the vehicle itself which is affected by external
able u for every control variance e so that the influences such as changing road surfaces.
왘 Standard control loop using TCS transversal differential lock controller as an example
z
Reference value for Asymmetr. Differential
differential speed ˘SoDif Transversal braking torque speed ˘Dif yactual Controlled variable
differential Vehicle yref Reference variable
yref + e lock controller u
y actual e Control variance
(Controller) (Control loop)
æ UFB0764E
yref – yactual
u Correcting variable
z External disturbance
values
Standard controller The degree of counter-reaction of these

Proportional, Integral and Differential elements controllers increases the greater the control
are often used as controllers. The correcting variance (P-controller), the longer the control
variable u is determined by the current control variance lasts (I-controller) or the greater the
variance e as shown below: tendency of the control variance to change
P-con- Multiplication u(t) = KP · e(t) (D-controller). Combining these basic con-
troller trollers gives PI, PD and PID-controllers.
I-con- Time integration u(t) = KI · ∫e(t)dt The transversal differential lock controller of
troller the traction control system (TCS) is designed
D-con- Time derivative u(t) = KD · de(t)/dt as a PI-controller that contains additional non-
troller linear elements.
378 Electronic stability program Requirements
Electronic Stability Program (ESP)

Human error is the cause for a large portion Requirements
of road accidents. Due to external circum-
stances, such as an obstacle suddenly appear- ESP enhances driving safety by providing the
ing on the road or driving at inappropriately following assets:
high speeds, the vehicle can reach its critical 쐌 Enhanced vehicle stability; the system
limits and it becomes uncontrollable. The lat- keeps the vehicle on track and improves
eral acceleration forces acting on the vehicle directional stability under all operating
reach values that overtax the driver. Elec- conditions, including emergency stops,
tronic systems can make a major contribu- standard braking maneuvers, coasting,
tion towards increasing driving safety. acceleration, trailing throttle (overrun),
and load shift
The Electronic Stability Program (ESP) is 쐌 Increased vehicle stability at the limits
a closed-loop system designed to improve ve- of traction, such as during sharp steering
hicle handling and braking response through maneuvers (panic response), to reduce the
programmed intervention in the braking sys- danger of skidding or breakaway.
tem and/or drivetrain. The integrated func- 쐌 In a variety of different situations, further
tionality of the ABS prevents the wheels from improvements in the exploitation of trac-
locking when the brakes are applied, while tion potential when ABS and TCS come
TCS inhibits wheel spin during acceleration. into action, and when engine drag torque
In its role as an overall system, ESP applies a control is active, by automatically increas-
unified, synergistic concept to control the ve- ing engine speed to inhibit excessive engine
hicle’s tendency to “plow” instead of obeying braking. The ultimate effects are shorter
the helm during attempted steering correc- braking distances and greater traction
tions; and at the same time it maintains sta- along with enhanced stability and higher
bility to prevent the vehicle breaking away to levels of steering response.
the side, provided the vehicle remains within
its physical limits.
1 Lateral dynamic response on passenger car without ESP
β
Fig. 1
1 Driver steers, lateral-
force buildup.
2 Incipient instability
because side-slip 4
angle is too large.
3 Countersteer,
driver loses control
of vehicle.
4 Vehicle becomes
uncontrollable. 3
β
MG Yaw moment
1 2
FR Wheel forces MG
æ UAF0085Y
β Directional deviation β
β
from vehicle’s FR
longitudinal axis
(side-slip angle)

Electronic stability program Tasks and method of operation 379
Tasks and method Because this “discriminatory” control concept

relies on two individual intervention strate-
of operation gies, the system has two options for steering
The electronic stability program is a system the vehicle: it can brake selected wheels
that relies on the vehicle’s braking system as a (selective braking) or accelerate the driven
tool for “steering” the vehicle. When the sta- wheels. Within the invariable limits imposed
bility-control function assumes operation it by the laws of physics, ESP keeps the vehicle
shifts the priorities that govern the brake sys- on the road and reduces the risk of accident
tem. The basic function of the wheel brakes – and overturning. The system enhances road
to decelerate and/or stop the vehicle – as- safety by furnishing the driver with effective
sumes secondary importance as ESP inter- support.
venes to keep the vehicle stable and on
course, regardless of the conditions. Below are four examples comparing vehicles
Specific braking intervention is directed with and without ESP during operation “on
at individual wheels, such as the inner rear the limit”. Each of the portrayed driving ma-
wheel to counter understeer, or the outer neuvers reflects actual operating conditions,
front wheel during oversteer, as shown in and is based on simulation programs de-
Fig. 2. For optimal implementation of stabil- signed using data from vehicle testing. The
ity objectives, ESP not only initiates braking results have been confirmed in subsequent
intervention, but it can also intervene on the road tests.
engine side to accelerate the driven wheels.
2 Lateral dynamic response on passenger car with ESP
Fig. 2
force buildup.
β 2 Incipient instability,
ESP intervention at
4
right front.
3 Vehicle remains
under control.
4 Incipient instability,
ESP intervention at
left front, complete
stabilization.
3
β
MG Yaw moment
FR Wheel forces
1 2 β Directional deviation
MG
β
æ UAF0021-1Y
from vehicle’s
FR β longitudinal axis
(side-slip angle)
Increased braking
force
380 Electronic stability program Maneuvers
Maneuvers 쐌 on a high-traction road-surface

(coefficient of friction μHF = 1),
Rapid steering and countersteering 쐌 without the driver braking,
This maneuver is similar to lane changes 쐌 with an initial speed of 144 km/h.
or abrupt steering inputs such as might be
expected for instance Initially, as they approach the S-bend, the
쐌 when a vehicle is moving too fast when conditions for both vehicles, and their reac-
it enters a series of consecutive S-bends, tions, are identical. Then come the first
쐌 or which have to be initiated when, with steering inputs from the drivers (phase 1).
oncoming traffic, an obstacle suddenly
appears on a country road, or Vehicle without ESP
쐌 which are necessary when an overtaking As can be seen, in the period following the
maneuver on the highway or freeway initial, abrupt steering input the vehicle
suddenly has to be aborted. without ESP is already threatening to be-
come unstable (Fig. 4 on left, phase 2).
Figs. 3 and 4 demonstrate the handling re- Whereas the steering input has quickly gen-
sponse of two vehicles (with and without erated substantial lateral forces at the front
ESP) negotiating a series of S-bends with wheels, there is a delay before the rear
rapid steering and countersteering inputs wheels start to generate similar forces. The
vehicle reacts with a clockwise movement
3 Curves for dynamic response parameters during around its vertical axis (inward yaw). The
a right-left cornering sequence
next stage is phase 3 with the second steer-
ing input. The vehicle without ESP fails to
+180° respond to the driver’s attempt to counter-
1+2
Steering-wheel
steer and goes out of control. The yaw veloc-

0 ity and the side-slip angle rise radically, and
the vehicle breaks into a skid (phase 4).
angle
-180° Vehicle with ESP

+10 On this vehicle ESP brakes the left front
m/s2
1 wheel to counter the threat of instability
acceleration
0
(Fig. 4 on right, phase 2) that follows the ini-
Lateral
tial steering input. Within the ESP context

this is referred to as active braking, and it
2
-10 proceeds without any intervention on the
+ 45 driver’s part. This action reduces the ten-
°/s
1
dency to yaw toward the inside of the corner
(inward yaw). The yaw velocity is reduced
0
velocity
and the side-slip angle is limited. Following

Yaw
2 the countersteer input, first the yaw moment

-45 and then the yaw velocity reverse their direc-
+22° tions (phase 3). In phase 4 a second brief
1 brake application – this time at the right
front wheel – restores complete stability.
Side-slip
0 The vehicle remains on the course defined

æ UAF0061-1E
angle
Phase 1 2 3 4 2 by the steering-wheel angle.

Fig. 3 -22°
1 Vehicle without ESP 0 Time s 11
2 Vehicle with ESP
Electronic stability program Maneuvers 381
4 Vehicle tracking during right-left cornering sequence
Vehicle without ESP Vehicle with ESP
Fig. 4
Increased braking
force
force buildup.
2 Incipient instability
Right: ESP interven-
tion at left front.
3 Countersteer
Phase 4 Left: Driver loses
control of vehicle;
Phase 4
Right: Vehicle remains
Phase 3
under control.
4 Left: Vehicle becomes
Phase 3
uncontrollable,
Right: ESP interven-
tion at right front,
Phase 2 Phase 2
complete stabilization.
Fig. 5
æ UAF0060-1E
a Oversteering
Phase 1 behavior.
Phase 1
1 The rear end of the
vehicle breaks away.
2 ESP applies the
brake at the outer
5 Over and understeering behavior when cornering
front wheel and this
a b reduces the risk of
skidding.
3 The vehicle without
3 ESP breaks into
3 a slide.
b Understeering
2 behavior
2
1 The front of the
1 vehicle breaks away.
1 2 ESP applies the
brake at the inner
rear wheel and this
with ESP with ESP reduces the risk of
understeering.
without ESP without ESP 3 The vehicle without
æ UAF0086E ESP is understeered

and leaves the road.
Lane change with emergency braking 쐌 one equipped solely with the Antilock
When the last vehicle in a line of stopped Braking System (ABS) and
traffic is hidden by a rise in the road, and thus 쐌 one that is also fitted with ESP.
invisible to traffic approaching from behind, Both vehicles
drivers closing on the traffic jam cannot reg- 쐌 are traveling at 50 km/h and
ister the dangerous situation until the very 쐌 on a slippery road surface (μHF = 0.15)
last second. Stopping the vehicle without col- at the moment when the driver registers
liding with the stationary traffic will entail a the dangerous situation.
lane change as well as braking.
7 Curves for dynamic response parameters for lane
change during an emergency stop at υ0 = 50 km/h
Figs. 6 and 7 show the results of this evasive and μHF = 0.15
action as taken by two different vehicles:
50
6 Lane change during emergency braking
km/h
Vehicle without Vehicle with ESP
Vehicle speed
ESP, with ABS 1
2
0
4 4
+300°
1
Steering-wheel angle
0
2
-300°
+40
°/s 1
Yaw velocity
3 3 2
0
Fig. 6
+16°
υ0 = 50 km/h
μHF = 0.15 1
Side-slip angle
2 2
Increased brake slip
0
2
Fig. 7
æ UAF0062-1Y
υ0 = 50 km/h
æ UAF0063-1E
1 2 3 4
μHF = 0.15 1 1
-16°
0 Time s 15
1 Vehicle without ESP
2 Vehicle with ESP
Vehicle with ABS but without ESP the complexity of the steering process and
Immediately after the initial steering input lowers the demands placed on the driver.
both the side-slip angle and the yaw velocity Yet another asset is that the ESP vehicle stops
have increased to the point where driver in- in less distance than the vehicle equipped
tervention – in the form of countersteer – has with ABS alone.
become imperative (Fig. 6, on
left). This driver action then 9 Oversteering and understeering when cornering
generates a side-slip angle in the 1
opposite direction (technically:
with the opposite operational
sign). This side-slip angle in-
creases rapidly, and the driver 2
must countersteer for a second
time. Here the driver is able –
but only just – to restabilize the
vehicle and bring it to a safe halt.
Vehicle with ESP

Because ESP reduces yaw veloc- 1
ity and side-slip angle to easily Fig. 9
controllable levels, this vehicle 1 Vehicle with ESP
remains stable at all times. The 3 2 Oversteered vehicle
driver is not confronted with without ESP
unanticipated instability and can
æ UAF0088Y
3 Understeered vehicle
thus continue to devote full at- without ESP
tention to keeping the vehicle on

course. ESP substantially reduces Fig. 8
Vehicle without ESP
8 Critical obstacle-avoidance maneuver with and without ESP 1 Vehicle approaches
an obstacle.
2 Vehicle breaks away
2 and does not follow
3 the driver’s steering
movements.
3 Vehicle slides
uncontrolled off
4
the road.
2
3 Vehicle with ESP
1 Vehicle approaches
an obstacle.
2 Vehicle almost
breaks away 씮 ESP
!
intervention, vehicle
follows driver’s
with ESP steering movements.
1 3 Vehicle almost
without ESP breaks away again
æ UAF0087E
when recentering
the steering wheel
씮 ESP intervention.
4 Vehicle is stabilized.
Extended steering and countersteering clearly seen during this kind of dynamic ma-
sequence with progressively greater neuver when the steering wheel has to
input angles be turned to progressively greater angles to
A vehicle traversing a series of S-curves (for negotiate each turn.
instance, on a snaking secondary road) is in Figs. 10 and 11 illustrate the handling
a situation similar to that encountered on a response of two vehicles (one with and
slalom course. The way the ESP works can be one without ESP) under these conditions
쐌 on a snow-covered road (μHF = 0.45),
10 Curves for dynamic response parameters for rapid 쐌 without the driver braking, and
steering and countersteering inputs with increasing
steering-wheel angles 쐌 at a constant velocity of 72 km/h.
+180° Vehicle without ESP

Engine output will have to be increased con-
wheel angle
1+2
tinually in order to maintain a constant road
Steering-
0 speed. This, in turn, will generate progressively

greater slip at the driven wheels. A sequence of
steering and countersteering maneuvers with a
-180°
40° steering-wheel angle can quickly increase
+5
m . s-2 1
drive slip to such levels that a vehicle without
ESP becomes unstable. At some point in this
Lateral acceleration
alternating sequence the vehicle suddenly

ceases to respond to steering inputs and breaks
0 into a slide. While lateral acceleration remains
virtually constant, both side-slip angle and yaw
velocity rise radically.
2
-5 Vehicle with ESP
The Electronic Stability Program (ESP) inter-
+30
°/s venes at an early stage in this sequence of
1 steering and countersteering maneuvers to
counter the instability that threatens right
Yaw velocity
from the outset. ESP relies on engine inter-

0 vention as well as individually controlled
2 braking of all four wheels to maintain the
vehicle’s stability and steering response. Side-
slip angle and yaw velocity are controlled so
-30 that the driver’s steering demands can be
+10°
complied with as far as possible considering
1 the prevailing physical conditions.
Side-slip angle
0
2
æ UAF0065-1E
1 2 3 4 5 6
Fig. 10 -10°
0 Time s 14
1 Vehicle without ESP
2 Vehicle with ESP
11 Vehicle tracking for rapid steering and countersteering inputs with increasing steering-wheel angles.
6
4 4
3 3
1 1
æ UAF0064-1Y
Fig. 11
Increased braking
force
Acceleration/deceleration during cornering for steering input starts to rise rapidly, while at
A decreasing-radius curve becomes progres- the same time the side-slip angle increases dra-
sively tighter as one proceeds. If a vehicle matically. This is the upper limit at which the
maintains a constant velocity through such driver can still manage to keep the car inside
a curve (as frequently encountered on free- the skid pad. A vehicle without ESP enters its
way/highway exit ramps) the outward, or cen- unstable range at roughly 98 km/h. The rear
trifugal force, will increase at the same time end breaks away and the driver must counter-
(Fig. 12). This also applies when the driver steer and leave the skid pad.
accelerates too soon while exiting a curve.
In terms of the physics of vehicle operation, Vehicle with ESP
it produces the same effects as the situation de- Up to a speed of roughly 95 km/h, vehicles
scribed above (Fig. 13). Excessive braking in with and without ESP display identical re-
such a curve is yet another potential source of sponse patterns. Because this speed coincides
radial and tangential forces capable of induc- with the vehicle’s stability limit, the ESP refuses
ing instability during cornering. to implement continued driver demand for
A vehicle’s handling response during further acceleration to beyond this point. ESP
simultaneous acceleration and cornering uses engine intervention to limit the drive
is determined by testing on the skid pad torque. Active engine and braking intervention
(semi-steady-state circulation). In this test the work against the tendency of the vehicle to un-
driver gradually accelerates dersteer. This results in minor deviations from
쐌 on a high-traction surface (μHF = 1.0) and the initially projected course, which the driver
쐌 attempts to stay on a skid pad with a can then correct with appropriate steering ac-
100 meter radius while circulating at pro- tion. At this point, the driver has assumed an
gressively faster rates; this continues until active role as an element within the control
the vehicle reaches its cornering limits. loop. Subsequent fluctuations in steering-
wheel and side-slip angle will now be a func-
Vehicle without ESP tion of the driver’s reactions, as will the final
During testing on the skid pad at approxi- speed of between 95 and 98 km/h as stipulated
mately 95 km/h the vehicle responds to the im- for the test. The vital factor is that ESP consis-
pending approach of its physical operating tently maintains these fluctuations within the
limits by starting to understeer. The demand stable range.
12 Vehicle tracking when cornering while braking with a constant steering-wheel angle
Fig. 12
æ UAF0066-1Y
Increased braking
force
Decreased braking
force
13 Vehicle tracking when cornering while accelerating.
æ UAF0067-1Y
Fig. 13
Increased braking
force
14 Comparison of cornering with vehicles with and without ESP
æ UAF0089Y
388 Electronic stability program Closed-loop control system and controlled variables
Closed-loop control system System and control structure

The Electronic Stability Program (ESP)
and controlled variables embraces capabilities extending far beyond
Electronic stability program concept those of either ABS or ABS and TCS com-
Application of the ESP closed-loop stability bined. Based on advanced versions of ABS
control in the vehicle’s limit situation as and ABS/TCS system components, it allows
defined by the dynamics of vehicle motion active braking at all four wheels with a high
is intended to prevent the level of dynamic sensitivity. Vehicle response
쐌 linear (longitudinal) velocity, the is adopted as an element within the control
쐌 lateral velocity and the loop. The system controls braking, propulsive
쐌 rotational speed around the vertical axis and lateral forces so that the actual response
(yaw velocity), converges with the ideal response under the
given circumstances.
exceeding the ultimate control limits. Assum- An engine-management system with CAN
ing appropriate operator inputs, driver de- interface can vary engine output torque in or-
mand is translated into dynamic vehicular re- der to adjust the driven-wheel slip rates. The
sponse that is adapted to the characteristics of advanced ESP system provides highly precise
the road in an optimization process designed performance for selective adjustment of the
to ensure maximum safety. As shown in Fig. dynamic longitudinal and lateral forces acting
1, the first step is to determine how the vehi- on each individual wheel.
cle should respond to driver demand during
operation in the limit range (ideal response), Fig. 2 shows ESP control in a schematic
and also how it actually does respond (actual diagram with
response). Actuators are then applied to mini- 쐌 the sensors that determine the controller
mize the difference between the ideal and input parameters,
the actual response (control deviation) by 쐌 the ESP control unit with its hierarchically-
indirectly influencing the forces acting at the structured controller, featuring a higher-
tires. level ESP controller and the subordinate
slip controllers,
쐌 the actuators used for ultimate control
of braking, drive and lateral forces.
1 Block diagram of electronic stability program (ESP)
Vehicle (control loop)

Actuators for brake and engine interventions
Measured/estimated motion variables
Driver command Vehicle response

Specified behavior Actual behavior
Control variance
æ UAF0083E
Controller for calculating correcting variables

Electronic stability program Closed-loop control system and controlled variables 389
Controller hierarchy and the vehicle speed are also included in

Level 1 ESP controller the processing calculations as supplementary
Task parameters. The “monitor” estimates these
The ESP controller is responsible for factors based on signals transmitted by the
쐌 determining the current vehicle status sensors for
based on the yaw velocity signal and the 쐌 wheel speed,
side-slip angle estimated by the “monitor” 쐌 lateral acceleration,
and then 쐌 braking pressures, and
쐌 achieving maximum possible convergence 쐌 yaw velocity.
between vehicle response in the limit range
and its characteristics in the normal oper- The desired vehicle response is brought about
ating range (ideal response). by generating a yaw moment acting on the
vehicle. In order to generate the desired yaw
The following components register driver de- moment, the controller intervenes in the
mand and the system processes their signals tire-slip rates to indirectly influence the
as the basis for defining ideal response: longitudinal and lateral forces. The system
쐌 engine-management system (e. g. pressure influences the tire slip by varying the specifi-
on accelerator pedal), cations for slip rate, which must then be
쐌 primary-pressure sensor (e. g. activation of executed by the subordinated ABS and TCS
brakes), or controllers.
쐌 steering-wheel sensor (steering-wheel The intervention process is designed to
angle). maintain the handling characteristics that
the manufacturer intended the vehicle to
At this point the specified response is defined have and to serve as the basis for ensuring
as driver demand. The coefficient of friction consistently reliable control.
2 ESP control loop in vehicle
Sensors ESP control unit Actuators

5 Fig. 2
6 7
1 Yaw-rate sensor
1
with lateral-accelera-
tion sensor
2 Steering-wheel-
ESP controller 8
angle sensor
9
3 Primary-pressure
2 sensor
4 Wheel-speed
sensors
10 5 ESP control unit
3 6 Hydraulic modulator
7 Wheel brakes
ABS/TCS slip controller, 8 Engine management
EDTC engine drag 11 ECU
4 torque controller 9 Fuel injection
æ UAF0007-1E
only for gasoline-engines:

10 Ignition-timing
intervention
11 Throttle-valve
intervention (ETC)
The ESP controller generates the specified 쐌 yaw velocity (measured parameter),
yaw moment by relaying corresponding 쐌 steering-wheel angle (measured para-
slip-modulation commands to the selected meter),
wheels. 쐌 lateral acceleration (measured parameter),
The subordinate-level ABS and TCS con- 쐌 vehicle’s linear velocity (estimated para-
trollers trigger the actuators governing the meter), and
brake hydraulic system and the engine- 쐌 longitudinal tire forces and slip rates
management system using the data generated (estimated parameters)
in the ESP controller.
the monitor determines the following:
Design 쐌 lateral forces acting on the wheel,
Fig. 3 is a simplified block diagram showing 쐌 slip angle,
the design structure of the ESP controller. 쐌 side-slip angle, and
It portrays the signal paths for input and out- 쐌 vehicle lateral speed.
put parameters. Based on the
3 Simplified block diagram showing ESP controller with input and output variables
Vehicle
Monitored variables Engine management with

Steering-wheel angle electronic throttle control (ETC)
Yaw velocity
Wheel speed
Brake circuit pressure Hydraulic modulator
ESP controller
Monitor defines
estimated parameters
Reference value calculation

for yaw velocity/side-slip angle
Yaw moment status controller
Reference value calculation

for brake locking torque/tire slip
ABS controller
æ UAF0084E
TCS controller
EDTC engine drag torque controller
The specifications for side-slip angle and yaw 왘 Single-track model

velocity are determined on the basis of the
following parameters, which may be either Ranges of lateral acceleration
directly or indirectly defined by driver input: Passenger cars can achieve lateral accelera-
쐌 steering-wheel angle, tions up to 10 m/s2. Lateral acceleration in the
쐌 estimated vehicle speed, small signal range (0...0.5 m/s2) can be caused
쐌 coefficient of friction, which is determined by road conditions such as ruts or by cross-
on the basis of the longitudinal acceleration winds.
(estimated parameter) and lateral The linear range extends from 0.5...4 m/s2.
acceleration (measured parameter), and Typical lateral maneuvers include changing
쐌 accelerator-pedal travel (engine torque) lane or load-change reactions when cornering.
or brake-circuit pressure (force on brake The response of the vehicle in these situations
pedal). can be described using the linear single-track
model.
These processes also take into account the
In the transition range (4...6 m/s2) some vehi-
special characteristics related to vehicle dy-
cles still respond linearly while others do not.
namics, as well as unusual situations, such
as a crowned road or μ-split surface (e.g. high The critical range above 6 m/s2 is only
traction on left of lane with right side slip- reached in extreme situations such as situations
pery). that almost result in an accident. In this case
the vehicle response is highly non-linear.
Method of Operation
The ESP controller governs the two status pa- Assumptions with the single-track model
rameters “yaw velocity” and “side-slip angle”
The linear single-track model can be used to
while calculating the yaw moment required to
obtain important information about the lateral
make the actual and desired-state parameters
behavior of a vehicle. In the single-track model
converge. As the side-slip angle increases, so
the lateral properties of an axle and its wheels
does its significance for the controller.
are summarized into one effective wheel. In
The control program is based upon data
the simplest version, the properties taken into
concerning the maximum potential lateral
account are modeled using linear equations;
acceleration and other data selected to reflect
thus the model is referred to as a linear single-
the vehicle’s dynamic response patterns.
track model.
These are determined for each vehicle in
steady-state skid-pad testing. In subsequent The most important model assumptions are:
steady-state vehicle operation, as well as dur- 쐌 Kinematics and elastokinematics of the axle
ing braking and acceleration, this data – are only modeled linearly.
defining how the steering angle and vehicle 쐌 The lateral force buildup of the tire is linear
speed relate to the yaw velocity – serves as the and the aligning torque is ignored.
basis for defining the desired vehicle motion.
The required data (nominal yaw velocity) is 쐌 The center of gravity is assumed to be at
stored in the program in the form of a single- road level. The only rotational degree of
track model. freedom of the vehicle is therefore the yaw
motion. Rolling, pitching and lifting (transla-
The nominal yaw velocity must be limited tional movements in z-axis) are not taken
in line with the friction coefficient so as to into account.
keep the vehicle on the predefined physically
feasible track.
For example, if a vehicle breaks into over-
steer while coasting into a right-handed
curve, and the specified yaw velocity is ex- Level 2 ABS controller
ceeded (the vehicle evinces a tendency to Task
rotate too quickly around its vertical axis), The hierarchically subordinated ABS con-
ESP responds by braking the left front wheel troller goes into operation whenever the
to generate a defined brake slip which shifts desired slip rate is exceeded during braking,
the yaw moment toward greater counter- and it becomes necessary for ABS to inter-
clockwise rotation thus suppressing the vene. During both ABS and “active” braking,
vehicle’s tendency to break away. the closed-loop control of wheel-slip rates
If a vehicle breaks into understeer while as applied for various dynamic-intervention
coasting into a right-handed curve, and the functions must be as precise as possible.
yaw velocity is below the specified yaw rate The system needs precise data on slip as a
(the vehicle evinces a tendency to rotate too precondition for dialing in the specified slip
slowly around its vertical axis), ESP responds rates. It must be pointed out that the system
by braking the right rear wheel to generate does not measure the vehicle’s longitudinal
a defined brake slip which shifts the yaw speed directly. Instead, this parameter is de-
moment toward greater clockwise rotation rived from the rotation rates of the wheels.
thus suppressing the vehicle’s tendency to
push over the front axle. Design and method of operation
By briefly “underbraking” one wheel, the ABS
ESP controller functions during ABS and controller performs an indirect measurement
TCS operation of vehicle speed. It interrupts the slip control
The entire spectrum of monitored and to lower the current braking torque by a
estimated data is relayed to the subordinate defined increment. The torque is then main-
controllers for continuous processing. This tained at this level for a given period.
guarantees maximum exploitation of the Assuming that the wheel has stabilized and
traction available between tire and road sur- is turning freely with no slip at the end of
face for the basic ABS and TCS functions this period, it can serve as a suitable source
under all operating conditions. for determining (no-slip) wheel speed.
During active ABS operation (with wheels The calculated velocity at the center of
tending to lock) the ESP controller provides gravity can be used to determine the effective
the subordinate ABS controller with the (free-rolling) wheel speeds at all four wheels.
following data: These data, in turn, form the foundation for
쐌 lateral vehicle velocity, calculating the actual slip rates at the remain-
쐌 yaw velocity, ing three – controlled – wheels.
쐌 steering-wheel angle, and
쐌 wheel speeds as the foundation for Level 2 engine drag-torque controller
providing the desired ABS slip. (EDTC)
Task
When TCS is active (wheels threatening to Following downshifts and when the accelera-
break into uncontrolled spin when moving tor is suddenly released, inertia in the engine’s
off or during rolling acceleration) the ESP moving parts always exerts a degree of brak-
controller transmits the following offset ing force at the drive wheels. Once this force
values to the subordinate TCS controller: and the corresponding reactive torque rise
쐌 change in the specified value for the beyond a certain level, the tires will lose their
drive-slip, ability to transfer the resulting loads to the
쐌 change in the slip tolerance range, and road. Engine drag-torque control intervenes
쐌 change in a value to influence the torque under these conditions (by “gently” accelerat-
reduction. ing the engine).
Design and method of operation Design

Factors such as variations in the character of The specified propshaft speed and wheel-
the road surface can lead to conditions under speed differential data are calculated from
which the engine-braking torque is suddenly the specified slip values and speeds for the
too high. The result is a tendency for the “coasting” wheels. The propshaft speed and
wheels to lock. One available countermeasure wheel-speed differential controlled variables
is judicious throttle application. Here, the are derived from the driven wheel speeds.
ECU transmits signals to trigger the corre-
sponding actuators in the engine-manage- Method of Operation
ment system (with ETC function) for an in- The TCS module calculates the desired brak-
crease in drive torque. Intervention at the ing torque for both driven wheels and the
engine-management level is employed to reg- desired value for engine torque reduction to
ulate the driven wheel within the prescribed be implemented by the engine-management
limits. system.
Because the propshaft speed is affected
Level 2 TCS controller by inertial forces originating from the
Task drivetrain as a whole (engine, transmission,
The hierarchically subordinated TCS con- driven wheels, and the propshaft itself), a
troller is triggered in case of excessive slip relatively large time constant is employed
during (for example) standing-start and dur- to describe its corresponding leisurely rate
ing rolling acceleration when it becomes nec- of dynamic response. In contrast, the time
essary for TCS to intervene. TCS is intended constant for the wheel-speed differential is
to prevent the driven wheels from breaking relatively small, reflecting the fact that the
into a free spin. It functions by limiting en- wheels’ own inertial forces are virtually the
gine torque to a level corresponding to the sole determining factor for their dynamic
drive torque that the wheels can transfer to response. Another relevant factor is that the
the road surface. wheel-speed differential – unlike propshaft
Intervention at the driven wheels is carried speed – is not affected by the engine.
out by means of brake application or by in- The torques prescribed for the propshaft
cluding the engine management system in the and wheel-speed differential are taken as the
process. In diesel-engine vehicles, the elec- basis for defining the actuators’ positioning
tronic diesel control system (EDC) reduces forces. The system achieves the specified dif-
the engine torque by modifying the quantity ference in braking torque between the left and
of fuel injected. With gasoline-engines, the right-side driven wheels by transmitting the
engine torque can be reduced by varying the appropriate control signals to valves in the
throttle valve aperture (ETC), modifying the hydraulic modulator.
ignition timing, or suppressing individual Propshaft torques are regulated to the
injection pulses. desired level using symmetrical braking as
Active braking at non-driven wheels is gov- well as engine intervention. With a gasoline-
erned directly by the ABS controller. Unlike engine, adjustments undertaken through the
with ABS, TCS receives values from the ESP throttle valve are relatively slow to take effect
controller for changing the specified slip and (lag and the engine’s transition response).
permissible slip difference of the driven Retarding the ignition timing and, as a
axle(s). These changes take effect in the form further option, selective suppression of
of an offset applied to the basic values defined injection pulses are employed for rapid en-
in the TCS. gine-based intervention, while symmetrical
braking can be applied for brief transitional
support of engine torque reduction.
394 Automatic brake functions Overview
Automatic brake functions

The possibilities of today’s electronic brake Overview
systems go far beyond the tasks for which
they were originally designed. Originally The main function of the electronic brake
the antilock braking system (ABS) was system is the Electronic Braking-force
only used to prevent the wheels of a vehicle Distribution (EBD) function which replaces
from locking up and therefore to ensure the the mechanical components for braking-force
steerability of the vehicle even during emer- distribution between the front and rear axles.
gency braking. Today, the brake system also This function not only cuts costs, but also
controls the distribution of the braking- makes the electronic distribution of the
force. The electronic stability program (ESP), braking force extremely flexible.
with its ability to build up brake pressure Additional functions are gradually being
independently of the position of the brake integrated into electronic brake systems.
pedal, offers a whole series of possibilities The following additional functions are
for active brake intervention. The ESP is currently available:
intended to assist the driver by applying the 쐌 Hydraulic brake assist (HBA):
brakes automatically and to therefore pro- HBA detects emergency braking situations
vide the driver with a higher level of comfort and shortens the braking distance by
and convenience. Some functions, however, building up the brake pressure up to the
enhance the vehicle safety since automatic wheel-lock limit.
brake application during an emergency 쐌 Controlled deceleration for parking brake
results in shorter braking distances. (CDP):
CDP brakes the vehicle until it is stationary
when requested by the driver.
1 Block diagram
ESP
Deceleration
ACC HBA management/
CAN CDP standstill
HHC management
HDC
CDD Braking force
HFC
HRB
BDW Braking-force
distribution
Wheel braking torque
ABS Coordination
TCS of wheel
VDC braking torque
Hydraulic model
æ UFB0776E

Automatic brake functions Overview 395
쐌 Hill hold control (HHC): 쐌 Brake disk wiping (BDW):

HHC intervenes in the brake system when BDW removes splash water from the brake
pulling away on a hill and prevents the disks by briefly applying the brakes. This
vehicle from rolling backwards. brake application is not noticed by the
쐌 Hill descent control (HDC): driver.
HDC assists the driver when driving
downhill on steep terrain by automatically These functions work together with the
applying the brakes. Electronic Stability Program (ESP). Some
쐌 Controlled deceleration for driver assistance of these functions may also be available
systems (CDD): with the Antilock Braking System (ABS)
CDD brakes the vehicle if required in or the Traction Control System (TCS).
combination with automatic vehicle- Most of the additional functions operate
to-vehicle ranging. with the sensor technology of the existing
쐌 Hydraulic fading compensation (HFC): electronic brake systems. Some functions,
HFC intervenes if the maximum possible however, require additional sensors.
vehicle deceleration is not achieved even
though the driver is forcefully pressing the
brake pedal, e. g. due to high brake disk
temperatures.
쐌 Hydraulic rear wheel boost (HRB):
HRB also increases the brake pressure in
the rear wheels up to the wheel-lock limit
during ABS brake application.
2 System layout of additional functions
Sensors and setpoint generators ESP control unit Hydraulic diagram

CDP, HHC, HDC, CDD, HFC, HRB
Master cylinder HBA logic

Sensor technology for HBA,
pressure sensor CDP logic

Stop lamp switch HHC logic
HDC logic P
U
Wheel-speed sensor CDD logic
Ambient temperature HFC logic
sensor HRB logic
BDW logic HSV SV
Windshield wiper
signal
Rain sensor
Pedal-travel sensor ESP logic OV IV

technology for HHC-S
Parking brake switch

Additional sensor
Engine-speed sensor
Engine torque
Transmission
information
Safety logic
Lateral-acceleration
æ UFB0777E
sensor
Tilt sensor
396 Automatic brake functions Standard function
Standard function Electronic distribution

Electronic Braking-force Distribution (EBD)
Electronic braking-force distribution allows the distribution between front and
(EBD) rear brakes to be adjusted according to condi-
Requirements tions. Handling response is continuously
Legal requirements demand that the braking monitored and a greater proportion of the
systems of road vehicles are designed in such overall braking force can be applied to the
a way that they ensure a deceleration up to of rear brakes when conditions allow because
1) Gravitational 0.83g 1) and provide stable driving behavior the proportioning valve is dispensed with or
acceleration during all types of maneuvers so that the else larger dimensioned rear wheel brakes are
g = 9.81 m/s2
vehicle does not exhibit unstable handling fitted. This releases additional braking poten-
characteristics (i. e. a tendency to skid). tial for the front wheels which can be utilized
on vehicles with a high forward weight bias
Conventional braking-force distribution in particular.
On vehicles without ABS, this is achieved by a
fixed braking-force distribution between the Design
front and rear brakes or by the use of propor- The vehicle is designed in such a way that
tioning valves for the rear brakes (Fig. 1). without a proportioning valve, the fixed
Curve 2 shows the pattern for a vehicle braking-force distribution curve intersects
with a fixed braking-force distribution which, the ideal braking-force distribution curve
within the range 0...0.83g is below the ideal (curve 1) at a point P (Fig. 2) at a lower
braking-force distribution curve (1l) for an overall braking force, e.g. 0.5g. The use of the
unladen vehicle and fails to utilize the poten- existing ABS system’s hydraulics, sensors and
tial for higher rear-wheel braking forces. With electronics, but with modified valves and
a fully laden vehicle, (curve 1b), the utiliza- software, allows the braking force at the rear
tion of potential is even lower. Curve 3 illus- wheels to be reduced at higher overall braking
trates the characteristic obtained by using a levels.
proportioning valve, whereby a significant
gain in rear-wheel braking force is obtained
in an unladen vehicle. With the vehicle fully
laden, however, the im-
provement is relatively 1 Conventional braking-force distribution
small.
This latter character-
istic can be improved
Fig. 1 by the use of load-de-
1 Ideal braking-force pendent or decelera-
distribution of a tion-dependent propor- 1b
Braking force, rear wheels
vehicle:
tioning valves, but at
1l Unladen vehicle 4
1b Fully laden
the cost of complex me- 1l
vehicle chanics and hydraulics.
3
2 Fixed braking-force
distribution
2
3 Braking-force
distribution with
æ UFB0710-1E
proportioning valve
4 Straight line for
retardation of 0.83g
(g: gravitational Braking force, front wheels
acceleration)
Automatic brake functions Standard function 397
Method of operation the ABS system are activated, the return

The ECU continuously calculates the slip pump motor in the hydraulic modulator unit
difference between the front and rear wheels remains de-energized.
in all driving situations. If the ratio of front
to rear wheel slip exceeds a defined stable- Advantages
handling threshold when braking, the ABS The characteristics of the EBD system
pressure inlet valve for the appropriate rear outlined above provide the following
wheel is closed. This prevents further increase advantages:
of brake pressure at that wheel. 쐌 Optimized vehicle handling stability under
If the driver then further increases the force all payload conditions, in all cornering
applied to the brake pedal, and therefore the situations, on uphill or downhill gradients,
brake pressure, the degree of slip at the front and in any drivetrain status (clutch engaged/
wheels also increases. The difference between disengaged, automatic transmission),
front and rear wheel slip diminishes and the 쐌 No need for conventional proportioning
pressure inlet valve is opened again so that the valves or limiting valves,
pressure at the rear wheel rises once more. This 쐌 Reduced thermal stresses on the front
process may then be repeated a number of brakes,
times depending on the brake pedal force and 쐌 Even wear between front and rear brake
the maneuver being performed. The electronic pads,
braking-force distribution curve then takes on 쐌 Better vehicle deceleration with the same
a staircase appearance (curve 3) which approx- pedal force,
imates to the ideal braking-force distribution 쐌 Constant braking-force distribution
curve. patterns over the entire life of the vehicle,
쐌 Only minor modifications to existing ABS
For the Electronic Braking-force Distribution components are required.
(EBD) function, only the rear-brake valves of
2 Electronic braking-force distribution (EBD)
3 1
4
Braking force, rear wheels
Fig. 2
1 Ideal braking-force
distribution
2 Design braking-force
distribution
æ UFB0711E
3 Electronic braking-
Braking force, front wheels force distribution
4 Gain in rear-wheel
braking force
398 Automatic brake functions Additional functions
Additional functions Controlled Deceleration for Parking

Brake CDP
Hydraulic Brake Assist (HBA) The Electromechanical Parking Brake (EPB)
The main task of the hydraulic brake assist is an automated parking brake system. It
is to detect an emergency braking situation replaces the conventional hand brake lever
and as a result to automatically increase the or foot-operated parking brake lever with
vehicle deceleration. The vehicle deceleration an electric motor. The disadvantage of the
is only limited by the intervention of ABS parking brake is that it only affects the rear
control and is therefore close to the optimum axle and its braking force is limited in the
level possible within physical limits. A normal event of emergency braking. The CDP
driver is therefore able to achieve the short function increases braking deceleration and
braking distances that could previously only simultaneously enables an ESP system con-
be achieved by specially trained drivers. If the troller to be used to ensure vehicle stability.
driver reduces the desired level of braking,
the vehicle deceleration is reduced in accor- The CDP function is an additional function
dance with the force applied to the brake to actively increase the brake pressure in vehi-
pedal. The driver can therefore precisely cles fitted with an hydraulic braking system
modulate the vehicle deceleration when the and ESP system. On the driver’s request, the
emergency braking situation has passed. CDP function automatically decelerates the
The driver’s desired level of braking is vehicle until it is stationary. Once the vehicle
determined by the force or pressure he/she has come to a standstill, the ESP hydraulics
applies to the brake pedal. The pedal pressure system briefly adopts all static parking brake
is derived from the measured master cylinder processes.
pressure taking into account the current
hydraulic control.
The driver can intervene in the brake applica-

tion at any time and can therefore directly in-
fluence the vehicle’s response. The HBA can
only increase the brake pressure. The primary
pressure applied by the driver is therefore the 1 Comparison of braking with and without
brake assist function
minimum pressure used by the system. If a
system error occurs, the HBA is shutoff and
Average driver with brake assist
the driver is informed that an error has
occurred.
Braking pressure
Good driver
Average
driver
Delay
Extra braking distance

without brake assist
æ UFB0709-2E
Shortest braking distance

with brake assist
Automatic brake functions Additional functions 399
Hill Hold Control HHC as by the transmission outputting an engine

The Hill Hold Control (HHC) starting-off torque (e. g. automatic/continuously variable
assistant is a comfort and convenience func- transmission).
tion that prevents the vehicle from rolling
backwards when pulling away on hills and The brake pressure is not maintained in the
inclines. The gradient of the hill or slope brake system if there is already sufficient en-
is measured by a tilt sensor (longitudinal gine torque when the vehicle is stationary
acceleration sensor). To operate, the starting- (e.g. due to the accelerative force of the
off assistant requires the brake pressure automatic transmission).
available when the vehicle is stationary that If the driver presses the accelerator pedal
was built up when the driver pressed the foot during the vehicle-specific holding time,
brake. the holding time is extended until sufficient
engine torque is available for the vehicle to
The brake pressure specified by the driver pull away.
when stopping the vehicle is maintained in If the driver does not press the accelerator
the brake system when the system detects that or the brake pedal, the function is deactivated
the vehicle is stationary, even if the driver re- at the latest after two seconds. The vehicle
leases the brake pedal. The brake pressure is then starts rolling.
reduced after a pressure holding time of up to
a maximum of two seconds. During this time The HHC function is also designed as an
the driver can press the accelerator pedal and additional function for the ESP system and
pull away. The brake pressure is reduced it uses parts of this system. The function is
when the system detects the driver’s intention activated automatically.
to pull away.
The system detects that the driver wishes to

pull away if the engine torque is sufficient for
the vehicle to move in the desired direction.
This state can be triggered by the driver press-
ing the accelerator pedal and/or clutch as well
2 Function description of HHC
act
1 3 5
iva
ted
2 4 R
Activation dea
ctiv
• Stop on an incline (> approx. 2%) ate
d
• Press brake pedal
• Engage 1st gear
Deactivation
æ UFB0713E
(if front of vehicle is facing uphill)

• Engage reverse gear • Start off
(if rear of vehicle is facing uphill) • Shift to neutral position
Hill Descent Control HDC Another advantage of the HDC function is

The Hill Descent Control (HDC) is a comfort the variable distribution of the braking force
and convenience function that assists the which is coupled to the automatic driving di-
driver when driving down hill (on gradients rection recognition function. When reversing
of up to 50%) by automatically applying the the rear axle is braked more forcefully to
brakes. Once this function is activated a ensure optimum steerability even if the front
predetermined, low desired-speed is intro- axle is relieved of the load.
duced and regulated without any necessary
intervention from the driver. The level ground detection function inte-
The driver must activate and deactivate the grated in the HDC only permits brake appli-
HDC function by pressing the HDC push- cation by the HDC when driving downhill.
button. If the vehicle is on the flat or driving uphill,
If required the driver can vary the predeter- the HDC switches to a standby mode and
mined speed by pressing the brake and accel- is reactivated automatically as soon as the
erator pedal or using the control buttons of a system detects that the vehicle is traveling
speed control system. downhill.
To prevent the driver misusing this func-
If the brake slip of the wheels is too high dur- tion, the HDC also switches to its standby
ing HDC control, the ABS system intervenes mode if the accelerator pedal is pressed
automatically. If the wheels are on different passed a certain threshold or if the vehicle
road surfaces, the braking torque of the slip- exceeds a maximum control speed. The
ping wheels is automatically distributed to HDC function is deactivated if the vehicle
the wheels with a higher coefficient of fric- accelerates beyond a preset cutoff-speed.
tion.
If engine braking torque is available it is The status of the HDC function is indicated
used automatically. Compared to the exclu- by the HDC indicator lamp. The brake lamp
sive use of engine braking torque in order to indicates when the HDC function applies the
maintain the vehicle speed, the HDC func- brakes.
tion has the advantages that if the wheels
leave the surface of the road (loss of engine
braking torque) the vehicle speed is main-
tained and sudden acceleration phases do not
occur.
3 Function description of HDC

d
ate
activ
Functions de
• Default speed 8 km/h
Speed range 5…35 km/h
(can be adjusted by driver)
• Brake activation only when driving downhill
• Designed for off-road conditions
• Full functionality and power
only in 1st gear or reverse gear ted
tiva
Activation ac
• Switch
æ UFB0714E
Deactivation
• Switch
• High speed (> 60 km/h)
Automatic brake functions Additional functions 401
Controlled Deceleration for Driver Hydraulic Fading Compensation HFC

Assistance Systems CDD The Hydraulic Fading Compensation (HFC)
The basic CDD function is an additional function offers the driver additional brake
function for active brake application with the servo assistance. The function is activated if
Adaptive Cruise Control (ACC), i. e. for auto- the maximum possible vehicle deceleration
matic vehicle-to-vehicle ranging. The brakes is not achieved even if the driver forcefully
are applied automatically without the driver presses the brake pedal to the point that
pressing the brake pedal as soon as the dis- would normally cause the lockup pressure to
tance to the vehicle in front falls below a pre- be reached (primary pressure over approx.
determined distance. CDD is based on a hy- 80 bar). This is the case, for example at high
draulic braking system and an ESP system. brake disk temperatures or if the brake pads
have a considerably reduced coefficient of
The CDD function receives a request to friction.
decelerate the vehicle by a desired amount When the HFC is activated, the wheel
(input). CDD then sends the actual amount pressures are increased until all wheels have
of deceleration (output) which is achieved reached the lockup pressure level and ABS
by controlling the pressure using hydraulics. control is initiated. The brake application
The requested amount of deceleration is is therefore at the physical optimum. The
specified by the cruise control system which pressure in the wheel-brake cylinders can
is connected upstream. then exceed the pressure in the master
cylinder, also during ABS control.
If the driver reduces the desired level of brak-
ing to a value below a particular threshold
value, the vehicle deceleration is reduced in ac-
cordance with the force applied to the brake
pedal. The driver can therefore precisely modu-
late the vehicle deceleration when the braking
4 Function description of CDD situation has passed. The HFC cuts off if the
primary pressure or vehicle speed falls below
the respective cutoff threshold.
5 Function description of HFC

Typical signal curve
requirement
on
Braking
off Emergency braking range

with HFC HFC
inter-
Specified vention
Deceleration
deceleration
Emergency braking range
without HFC
Brake pressure
Current
deceleration
er
ow
Brake pressure
ce p
tan ing
Pressure in wheel-
sis rak
brake cylinders
as ith b
æ UFB0778E
æ UFB0779E
w
Time 0
Pedal force
Hydraulic Rear Wheel Boost HRB Brake Disk Wiping BDW

The Hydraulic Rear wheel Boost (HRB) pro- The Brake Disk Wiping (BDW) function de-
vides the driver with additional brake servo tects rain or a wet road by evaluating wind-
assistance for the rear wheels if the front shield wiper or rain sensor signals and then
wheels are controlled by the ABS system. actively increases the brake pressure in the
This function was introduced because many service brake. The brake pressure buildup is
drivers do not increase the pedal force at used to remove splash water from the disk
the start of ABS control even though the brake to ensure minimum brake response
situation would require this. When the HRB times when driving in wet conditions. The
function is active, the wheel pressures are pressure level when dry-braking is adjusted
increased at the rear wheels until they also so that the vehicle deceleration cannot be
reach the lockup pressure and ABS control is perceived by the driver.
initiated. The brake application is therefore Dry-braking is repeated at a defined inter-
at the physical optimum. The pressure in val for as long as the system detects rain or
the rear axle wheel-brake cylinders can then a wet road. If required just the disks at the
exceed the pressure in the master cylinder, front axle can be wiped.
also during ABS control. The BDW function interrupts the wiping
The HRB cuts off when the wheels at the procedure as soon as the driver applies the
front axle are no longer under ABS control or brakes.
when the primary pressure falls below the
cutoff threshold.
6 Function description of HRB
Emergency braking range HRB

for rear axle with inter-
HRB intervention vention
at rear
Emergency braking range axle
for front axle
Brake pressure
r
we
ce po
tan ing
sis ak
as th br
æ UFB0780E
wi
0
Pedal force
Electronic Control Unit Directional stability 403
왘 Directional stability
Satisfactory handling is defined according to face, the side-slip angle becomes excessive
whether a vehicle maintains a path that accu- (curve 3). Although with a controlled yaw ve-
rately reflects the steering angle while at the locity the vehicle will rotate just as far around
same time remaining stable. To meet this sta- its vertical axis as in curve 2, the larger side-
bility criterion, the vehicle must remain consis- slip angle is now a potential source of instabil-
tently secure and sure-footed, without “plow- ity. For this reason the electronic stability pro-
ing” or breaking away. gram controls the yaw velocity and limits the
Dynamic lateral response is a critical factor side-slip angle β (curve 4).
in the overall equation. The response pattern
is defined based on the vehicle’s lateral mo- 2 Dynamic lateral response
tion (characterized by the side-slip angle) and
its tendency to rotate around its vertical axis
3
(yaw velocity) (Fig. 1). β 4
Fig. 2 illustrates the dynamic lateral response β Fig. 2

of a vehicle being driven with a fixed steering 1 Steering input, fixed
steering-wheel angle
angle (skid-pad circulation). Position 1 repre-
2 Vehicle path on
sents the instant when steering input is ap- high-grip surface
plied. Curve 2 is the vehicle’s subsequent 2 3 Vehicle path on
course on a high-grip road surface; this track low-grip surface with
is an accurate reflection of the steering angle. yaw velocity control
This is the case when the coefficient of friction
æ UFA0005-1Y 4 Vehicle path on
is sufficient to transfer the lateral acceleration low-grip surface with
1 additional control of
forces to the road surface. If the coefficient of
the side-slip angle β
friction is lower e.g. due to a slippery road sur-
(ESP)
1 Vehicle travel directions
Vehicle vertical axis

Yaw velocity
Vehicle
longitudinal axis
Vehicle horizontal axis
Linear acceleration
æ UFA0013-1E
Rolling
Steering movement
404 Hydraulic modulator Development history
Hydraulic modulator
The hydraulic modulator forms the hy- Development history
draulic connection between the master cylin-
der and the wheel-brake cylinders and is The transition from 3/3 to 2/2 solenoid valve
therefore the central component of elec- was a milestone in the development of ABS.
tronic brake systems. It converts the control With 3/3 valves, which were used in genera-
commands of the electronic control unit and tion 2, the control functions for building up,
uses solenoid valves to control the pressures maintaining and reducing pressure could be
in the wheel brakes. carried out using just one valve. The valves
had three hydraulic connections to carry out
A distinction is made between systems that these functions. The disadvantages of this
modulate the brake pressure applied by the valve design were extremely expensive electri-
driver (Antilock Braking System, (ABS)) and cal control and great mechanical complexity.
systems that can build-up pressure automati- Control with the 2/2 valves of the current
cally (Traction Control System (TCS) and generations offers a less expensive solution.
Electronic Stability Program (ESP)). The following section describes how these
All systems are only available as dual-circuit valves work.
versions for legal reasons. Generation 8, which was introduced onto
the market in 2001, is designed as a fully
modular system. The hydraulic system can
therefore be tailored to meet the require-
ments of the respective vehicle manufacturer
such as with regard to value added functions,
comfort and convenience, vehicle segment
(up to lightweight commercial vehicles), etc.
Generation 8 is immersion-proof which
means that the hydraulic modulator can
withstand brief immersion in water.
1 ESP 8 hydraulic modulator
6 5 4 3 2 1
Fig. 1
1 Electronic control
unit
2 Coil grid
7
3 Coils/solenoid
group
4 Solenoid valves
5 Hydraulic block 8 9
æ UFB0750-1Y
6 DC motor
7 Plunger pump
8 Low-pressure
reservoir
9 Pressure sensor

Hydraulic modulator Design 405
Design greater slip determines the controllable pres-

sure of the axle.
Mechanical system With a 4-channel ABS system (for dual
A hydraulic modulator for ABS/TCS/ESP or X-braking-force distribution), one inlet
consists of an aluminum block into which the and one outlet valve is used for each wheel
hydraulic layout is drilled (Fig. 2). This block resulting in a total of eight valves. This system
also houses the necessary hydraulic function allows each wheel to be controlled individu-
elements (Fig. 1) presented below. ally.
ABS hydraulic modulator Moreover, the system features one pump ele-
With a 3-channel ABS system, this block fea- ment (return pump) and one low-pressure
tures one inlet valve and one outlet valve for reservoir for each brake circuit. Both pump
each front wheel and one inlet valve and one elements are operated by a shared DC motor.
outlet valve for the rear axle – a total of six
valves. This system can only be used in vehi-
cles featuring a dual brake-circuit configura-
tion. The two wheels at the rear axle are not
controlled individually, rather both wheels
are controlled according to the select-low
principle. This means that the wheel with the
2 Hydraulic diagram of 4-channel ABS hydraulic modulator (X-brake-circuit configuration)
HZ
Brake circuit 1 Brake circuit 2
IV IV
PE
M
AC
Fig. 2
HZ Master cylinder
RZ Wheel-brake
OV OV cylinder
IV Inlet valve
OV Outlet valve
PE Return pump
M Pump motor
RZ RZ RZ RZ AC Low-pressure
æ UFB0749-1E
reservoir
V Front
HL VR VL HR
H Rear
R Right
L Left
406 Hydraulic modulator Design
TCS hydraulic modulator driver in order to stabilize the vehicle. During

Unlike an ABS unit, a TCS system with such a maneuver (partially-active maneuver),
II-brake-circuit configuration also has a the suction path of the pump must be opened
switchover valve and an inlet valve at the despite the high primary pressure.
rear axle (driven wheels) resulting in a total An integrated pressure sensor is used exclu-
of 10 valves. sively in ESP systems. The sensor detects the
In TCS systems with X-brake-circuit con- brake pressure in the master cylinder i.e. the
figuration an additional switchover valve driver command. This is also required for a
and inlet valve are required for each circuit partially-active ESP control maneuver since it
(a total of 12 valves). is important to know the primary pressure
applied by the driver pressing the brake pedal.
ESP hydraulic modulator
ESP systems require 12 valves irrespective of Since TCS/ESP systems are required to
the brake-circuit configuration (Fig. 3). In generate pressure automatically, the return
these systems, the two inlet valves, as used in pump is replaced by a self-priming pump
the TCS hydraulic modulator, are replaced in both these systems. An additional non-
by two high-pressure switching valves. The return valve with a specific closing pressure is
difference between the two types of valves is required to prevent the pump from drawing
that the high-pressure switching valve can unwanted media from the wheels.
switch against higher differential pressures
(> 0.1 MPa). With ESP it may be necessary to
increase the brake pressure specified by the
3 Hydraulic diagram of ESP hydraulic modulator (X-brake-circuit configuration)
HZ
Brake circuit 1 Brake circuit 2
HSV SV
P
U
Fig. 3 PE
HZ Master cylinder IV IV M
RZ Wheel-brake
AC
cylinder
EV Inlet valve
AV Outlet valve
USV Switchover valve
OV OV
SV High-pressure
switching valve
PE Return pump
M Pump motor
AC Low-pressure
RZ RZ RZ RZ
æ UFB0752-1E
reservoir
V Front
H Rear HL VR VL HR
R Right
L Left
ABS versions 407
왘 ABS versions
Evolution of the ABS system have enabled the weight and dimensions of
Technological advances in the areas of ABS systems to be more than halved since
쐌 solenoid-valve design and manufacturing, the first-generation ABS2 in 1978. As a result,
쐌 assembly and component integration, modern systems can now be accommodated
쐌 electronic circuitry (discrete components even in vehicles with the tightest space re-
replaced by hybrid and integrated circuits strictions. Those advances have also lowered
with microcontrollers), the cost of ABS systems to the extent that it
쐌 testing methods and equipment (separate has now become standard equipment on all
testing of electronic and hydraulic systems types of vehicle.
before combination in the hydraulic modula-
tor), and
쐌 sensor and relay technology
1 Evolution of ABS configurations
Hydraulic-modulator development
ABS 2
ABS 5.0
ABS 5.3/5.7
Weight
6,200 g ABS 8
3,800 g 2,600 g
< 2,000 g
Development of electronics
ABS 2 ABS 5.0 ABS 5.3 ABS 5.3/5.7 ABS 8
Fig. 1
Historical development
of ABS showing techno-
æ UFB0700E
logical advances:
1989 1993 1995 2001 Decreasing weight
accompanied by
increasing processing
power.
408 Hydraulic modulator Pressure modulation
Pressure modulation system can of course control several wheels if

they lock up simultaneously.
Modulation with ABS control
The pressure of an ABS/TCS/ESP system
is modulated using solenoid valves. The out-
let valves and the inlet and high-pressure
switching valves (for TCS and ESP systems)
are switching valves that are closed at zero
current and can adopt two positions – closed
or open.
In contrast, the inlet valves and switchover
valves are both open at zero current and are 1 Pressure modulation in ABS hydraulic modulator
used as control valves for the first time in

generation 8. This is advantageous not only
for instantaneous braking power and braking a
comfort but also for noise emissions. Pres-

sures of up to 200 bar can be modulated IV IV
using the standard valve set. Special systems PE
for even higher pressures and also for greater M
AC
throughput, usually required in the commer-
cial vehicle sector, can be designed based OV OV
on further developments of the generation
8 modular system.
All valves are controlled using coils which HL VR
are controlled via the add-on control unit.
b
Pressure modulation with ABS hydraulic

modulator IV IV
In the event of ABS brake application, the PE
driver first generates the brake pressure at the M
AC
wheel by pressing the brake pedal. This can be
done without switching the valves since the OV OV
inlet valve is open at zero current and the
outlet valve is closed at zero current (Fig. 1a).
The pressure maintenance state is generated HL VR
Fig. 1 when the inlet valve is closed (Fig. 1b).
a Pressure build-up If a wheel locks up, the pressure from this c
b Pressure wheel is released by opening the relevant
maintenance
outlet valve (Fig. 1c). The brake fluid can also IV IV
c Pressure reduction
escape from the wheel-brake cylinder to the PE
relevant low-pressure reservoir. M
IV Inlet valve AC
OV Outlet valve This chamber performs the role of a buffer.
PE Return pump It can receive the brake fluid quickly and OV OV
M Pump motor efficiently. The return pump in the circuit,
AC Low-pressure
which is driven by a shared motor via an
reservoir
eccentric element, reduces the pressure speci-
æ UFB0781Y
V Front
HL VR
H Rear fied by the driver. The motor is actuated
R Right based on demand i.e. the motor is actuated
L Left according to the rotational speed. The ABS
Hydraulic modulator Pressure modulation 409
Pressure modulation with ESP hydraulic switching valve that is closed at zero current.
modulators These two valves are required for active/par-
Pressure is modulated with ESP control using tially-active brake intervention (Fig. 2).
ESP hydraulics in the same way as described
for ABS. Unlike with ABS, however, the Pressure generation with ESP
wheel-brake cylinder and master cylinder are The pressure generation chain consists of two
also connected via a switchover valve that is self-priming pumps and a motor. Plunger
open at zero current and a high-pressure pumps are used, as for ABS, but these pumps
can generate pressure without requiring the
2 Pressure modulation in ESP hydraulic modulator primary pressure applied when the driver
presses the brake pedal. These pumps are
a driven by a DC motor based on demand. The
motor drives an eccentric bearing located on
HSV SV the shaft of the motor.
P
U
IV PE IV TCS/ESP pumps can build up pressure inde-

M
pendently of the driver or increase the brak-
AC ing pressure already generated by the driver.
OV OV
These systems are therefore able to initiate
brake application. To do this, the switchover
valve is closed and the inlet valve or high-
HL VR pressure switching valve is opened. Fluid
can then be drawn from the brake fluid
reservoir via the master cylinder and pressure
b can be built up in the wheel-brake cylinders
(Fig. 2c). This is required not only for
HSV SV TCS/ESP functions, but also for many addi-
P
U
tional convenience functions (value added
IV PE IV functions such as the brake assist (HBA)).
M
Fig. 3
AC Demand-based control of the pump motor a Pressure build-up
when braking
OV OV reduces noise emission during pressure
b Pressure reduction
generation and regulation. The pumps can with ABS control
be equipped with damping elements to c Pressure build-up
HL VR satisfy vehicle manufacturer’s strict low noise via self-priming
emission requirements. pump due to TCS
or ESP intervention
c
IV Inlet valve
HSV SV OV Outlet valve
P SV Switchover valve
U
HSV High-pressure
IV PE IV switching valve
M
PE Return pump
AC M Pump motor
AC Low-pressure
OV OV reservoir
æ UFB0782Y
V Front
H Rear
HL VR R Right
L Left
410 Hydraulic modulator Pressure modulation
With ESP, there are basically three different Fully-active control

applications: If the ESP controller detects an unstable
쐌 Passive – as previously described for ABS vehicle state, the switchover valves are
control. closed. This prevents the pumps from being
쐌 Partially-active – when the pressure speci- hydraulically short-circuited via the
fied by the driver is insufficient to stabilize switchover/high-pressure switching valve
the vehicle. which would prevent pressure generation.
쐌 Fully-active – when pressure is generated to The high-pressure switching valves are
stabilize the vehicle without the driver simultaneously opened. The self-priming
pressing the brake pedal. pump now pumps brake fluid to the relevant
wheel or wheels in order to build up pressure.
Both pressure generation cases above, in ad- If pressure generation is only required in
dition to ESP control, are used for a whole se- one wheel, (for yaw rate compensation), the
ries of additional functions such as adaptive inlet valves of the other wheels are closed.
cruise control and the brake assist. To reduce the pressure, the outlet valves are
opened and the high-pressure switching
Partially-active control valves and switchover valves return to their
For partially-active control, the high-pressure original position. The brake fluid flows from
switching valve must be able to open the suc- the wheels to the reservoirs. The pumps then
tion path of the pump against high pressures. empty the reservoirs.
This is required since the driver has already
generated a high pressure, but this pressure is
insufficient to stabilize the vehicle.
The high-pressure switching valve is
designed with two stages so that the valve can
open against the high pressure. The first stage
of the valve is opened via the magnetic force
of the energized coil; the second stage via the
hydraulic area difference.
If the ESP controller detects an unstable
vehicle state, the switchover valves (open at
zero current) are closed and the high-pres-
sure switching valve (closed at zero current)
is opened. The two pumps then generate
additional pressure in order to stabilize the
vehicle.
Once the vehicle is stabilized, the outlet
valve is opened and the excess pressure in the
controlled wheel escapes to the reservoir. As
soon as the driver releases the brake pedal, the
fluid is pumped from the reservoir back to
the brake fluid reservoir.
Electronic Control Unit Development of hydraulic modulators 411
왘 Development of hydraulic modulators
Control of hydraulic modulators increased and reduced either continuously or

An electronic control unit processes the infor- incrementally (and therefore gradually).
mation received from the sensors and gener-
ates the control signals for the hydraulic mod- Hydraulic modulators with 2/2 solenoid valves
ulator. The hydraulic modulator incorporates While version ABS2S operates with 3/3 sole-
a series of solenoid valves that can open or noid valves, the successor systems ABS5 and
close the hydraulic circuits between the mas- ABS8 feature 2/2 solenoid valves with two hy-
ter cylinder and the wheel-brake cylinders. draulic connections and two valve positions.
The inlet valve between the master cylinder and
Hydraulic modulators with 3/3 solenoid valves the wheel-brake cylinder controls pressure
In 1978 version ABS2S was the first antilock build-up, while the outlet valve between the
braking system to go into series production. wheel-brake cylinder and the return pump con-
In this ABS system the electronic control unit trols pressure release. There is a solenoid valve
switches the 3/3 solenoid valves of the hy- pair for each wheel-brake cylinder (Fig. 1b).
draulic modulator to three different valve posi- – In the “pressure build-up” position, the inlet
tions. There is a solenoid valve for each wheel- valve connects the master cylinder to the
brake cylinder (Fig. 1a). wheel-brake cylinder so that the brake pres-
– The first (zero-current) position connects sure built up in the master cylinder can be
the master cylinder and the wheel-brake applied to the wheel-brake cylinder when
cylinder; the wheel brake pressure can rise. the brakes are applied.
– The second position (excitation at half the – In the “pressure maintenance” position, the
maximum current) separates the wheel inlet valve blocks the connection between
brake from the master cylinder and return the master cylinder and wheel-brake cylin-
line so that the wheel brake pressure der during rapid wheel deceleration (risk
remains constant. of locking up) and thus prevents the brake
– The third position (excitation at the maxi- pressure from rising any higher. The outlet
mum current) separates the master cylinder valve is also closed.
and simultaneously connects the wheel – If the wheel deceleration continues to in-
brake and return line so that the wheel crease, the inlet valve continues to block in
brake pressure decreases. the “decrease pressure” position. In addition
the return pump pumps out the brake fluid
By applying these settings in the appropriate via the open outlet valve so that the brake
sequences, the brake pressure can thus be pressure in the wheel-brake cylinder drops.
1 Comparison of ABS systems
a b
1 1
2 2
Fig. 1
3 3 a ABS2
4b
b ABS5
4a 4b 1 Damping chamber
2 Constrictor
æ UFB0545-2Y
3 Return pump
5 5 4a 3/3 solenoid valve
4b 2/2 solenoid valves
5 Accumulator
chamber
412 Sensotronic brake control Purpose and function
Sensotronic brake control (SBC)

Sensotronic brake control (SBC) is an Basic functions
electrohydraulic brake system that com- As with a conventional braking system,
bines the functions of a brake servo unit sensotronic brake control must be capable of
and the ABS (antilock braking system) 쐌 reducing the speed of the vehicle,
equipment, including ESP (electronic 쐌 bringing the vehicle to a halt, and
stability program). The mechanical opera- 쐌 keeping the vehicle stationary when it is
tion of the brake pedal is redundantly mea- stopped.
sured by the actuator unit and transmitted
to the control unit. There, control com- As an active braking system, it also performs
mands are calculated according to specific the tasks of
algorithms and passed to the hydraulic 쐌 operating the brakes,
modulator where they are converted into 쐌 amplifying the brake force, and
pressure modulating operations for the 쐌 modulating the brake force.
brakes. If the electronics fail, a hydraulic
fallback system is automatically available. SBC is an electronic control system with
hydraulic actuators. Braking force distribu-
Purpose and function tion takes place electronically to each wheel
in response to driving conditions. A vacuum
By utilizing its “brake-by-wire” capabilities, source for the brake servo function is no
SBC can control the hydraulic pressure in longer required. The self-diagnosis capabil-
the wheel brake cylinders independently of ity enables an early warning function for
driver input. As a result, functions beyond detection of possible system faults.
those performed by ABS (antilock braking SBC uses hydraulic standard wheel
system), TCS (traction control system) and brakes. Because of the fully electronic pres-
ESP (electronic stability program) can be sure control, SBC can be easily networked
implemented. One example is the conve- with vehicle handling systems. It thus meets
nient method of brake application for ACC all the demands made of future braking
(adaptive cruise control). systems.
1 SBC components in the car
1 2 3 4 1 1
Fig. 1
1 Active wheel speed
Motr
C
onic
sensor with direction
A
N
sensing
2 Engine management SB
C
ECU
3 SBC ECU
4 Yaw rate and lateral
acceleration sensor
5 Hydraulic modulator
CA
N
(for SBC, ABS, TCS

æ UFB0704-2Y
and ESP)
6 Actuator unit with
pedal travel sensor
7 Steering angle
5 1 6 7
sensor

Sensotronic brake control Purpose and function 413
By using a high-pressure accumulator, SBC optimum friction value is fully utilized. This
is capable of extremely rapid dynamic pres- results in a significant reduction of total
sure increases and thus offers the potential braking distance for hesitant drivers. The
for achieving short braking distances and highly dynamic braking force build-up of
excellent vehicle handling stability. Brake SBC exceeds that of conventional systems in
pressure modulation and active braking are this regard.
silent and produce no brake pedal feedback.
Consequently, SBC also satisfies demands Soft stop assist
for greater levels of comfort. SBC provides comfortable braking that stops
Braking characteristics can be adapted to the car with no jerking by automatically
the driving conditions (e.g. sharper response reducing the pressure just before the car
at high speeds or with more dynamic dri- comes to a complete stop. If more decelera-
ving styles). “Duller” pedal characteristics tion is desired, this function is not activated
allow the reduction of the braking effect, and SBC minimizes the braking distance.
which is necessitated by physics, to be
signaled to the driver before fading due Traffic jam assist
to overheating occurs. When traffic jam assist is activated, SBC
builds up a higher drag torque, which means
Additional SBC functions that the driver does not have to constantly
The auxiliary functions provided by SBC alternate between the accelerator and brake.
make a significant extra contribution to The vehicle is automatically braked and, if
safety and convenience. necessary, brought to a complete stop and
kept at a complete stop. This function can be
Hill hold control activated at speeds of 50 - 60 km/h.
After hill hold control is activated by a sig-
nificant increase in brake pressure while the Brake wiping
vehicle is stationary, the vehicle remains “Brake wiping” is an operation whereby the
braked without the need to keep the pedal film of water is regularly removed from the
down. The hill hold control is automatically brake disks in wet weather. It results in
released as soon as the driver has built up shorter stopping distances in wet conditions.
sufficient engine torque by depressing the You can take the information for activating
accelerator. This allows the driver to start the this function from the windshield wiper
car on a hill, for example, without activating signal, for example.
the parking brake system. Likewise, in other
situations in which the vehicle would roll
out of position if not braked, the driver does 2 Evolution of brake systems
not need to keep his or her foot on the brake

at all times once hill hold control has been
activated.
Brake-by-Wire- EMB electromechanical brake
Systems SBC electrohydraulic brake
Enhanced brake assist function EBS electronic braking system (trucks)
If the driver abruptly releases the accelerator, Electronic ESP electronic stability program
an automatically regulated brake pressure brake control TCS traction control system
build-up takes place that gently applies the systems
ABS antilock braking system
brake pads. If panic braking follows, this al- Unassisted
æ UFB0736-1E
Power-assisted brake Conventional

lows the brake to “grab” more quickly and brake
Externally powered brake braking systems
thus allows a shorter total braking distance.
If the system detects panic braking, the
brake pressure is briefly increased until the
414 Sensotronic brake control Design, method of operation
Design brake pressure to be applied, which is built

up without pedal feedback.
The sensotronic brake control system
consists of the following components: SBC sensors
쐌 Actuator unit (Fig. 1, Item 6) Four pressure sensors measure the pressure
쐌 Vehicle dynamics sensors (1, 4 and 7) individually for each wheel circuit (Fig. 3).
쐌 Discrete (separate) ECU (3) A pressure sensor measures the storage pres-
쐌 Hydraulic modulator with add-on sure of the high-pressure accumulator. The
ECU (5). driver’s braking request is calculated by the
pedal travel sensor attached to the actuator
Those components are interconnected by unit and a pressure sensor that detects the
electrical control leads and hydraulic pipes. brake pressure applied by the driver.
Fig. 1 shows where they are fitted in a car. The pedal travel sensor consists of two
redundant, independent angle sensors.
Actuator unit Together with the pressure sensor for the
The actuator unit consists of: driver’s brake pressure, they allow a threefold
쐌 Master cylinder with expansion tank, detection of the driver’s request, and the sys-
쐌 Pedal travel simulator and tem can continue to work without errors
쐌 Pedal travel sensor. even if one of these systems fails.
Pedal travel simulation

The pedal travel simulator makes it possible Method of operation
to realize a suitable force-distance curve and
an appropriate damping of the brake pedal. Normal operation
In that way, the driver obtains the same Fig. 3 illustrates the components of SBC in
“brake feel” with the sensotronic brake the form of a block diagram. An electric
control system as with a very well designed motor drives a hydraulic pump. This charges
conventional braking system. a high-pressure accumulator to a pressure
between approx. 90 and 130 bar, which is
Sensors of the SBC systems monitored by the accumulator pressure sen-
The SBC sensors consist of the vehicle sor. The four independent wheel pressure
dynamics sensors familiar from ESP and modulators are supplied by this accumulator
the actual SBC sensors. and adjust the required pressure individually
for each wheel. The pressure modulators
Vehicle dynamics sensors themselves each consist of two valves with
The ESP sensors consist of four wheel speed proportional control characteristics and a
sensors, the yaw rate sensor, a steering angle pressure sensor.
sensor and, where applicable, a lateral accel- In normal operation, the isolating valves
eration sensor. These sensors provide the interrupt the connection to the actuator.
ECU with data relating to the speed and The system is in “brake-by-wire” mode. It
movement status of the wheels and driving electronically detects the driver’s braking re-
states such as cornering. Control functions quest and transmits it “by wire” to the wheel
such as ABS, TCS and ESP are executed in pressure modulators. The interaction of the
the familiar manner. engine, valves and pressure sensors is con-
trolled by the electronics, which are installed
If the vehicle is fitted with ACC (adaptive in the add-on ECU in hybrid technology.
cruise control), a radar system measures the They have two microcontrollers that moni-
distance to the vehicle in front. From these tor each other. The essential feature of these
data, the SBC control unit calculates the electronics is their extensive self-diagnosis,
Sensotronic brake control Method of operation 415
which monitors the plausibility of every sys- Braking in the event of system failure
tem state at all times. In this way, the driver For safety reasons, the SBC system is de-
can be notified of any failures before critical signed so that in the event of any serious
states occur. If components fail, the system errors (such as failure of the power supply),
automatically provides the optimal remain- the system is switched to a state in which the
ing partial function to the driver. An exten- vehicle can be braked even without active
sive fault memory allows prompt diagnostics brake force support. When de-energized, the
and repair in the event of a fault. isolating valves establish a direct connection
An intelligent interface with CAN bus es- to the actuator (Fig. 3) and thus allow a direct
tablishes the connection to the discrete ECU. hydraulic connection from the actuator unit
The following functions are integrated there: to the wheel brake cylinders.
쐌 ESP (electronic stability program), To maintain optimum function even if
쐌 TCS (traction control system), the system fails, the plunger pistons in the
쐌 ABS (antilock braking system), illustration serve as a medium separator be-
쐌 Driver brake request calculation and tween the active circuit of the SBC and the
쐌 SBC auxiliary functions (assist functions). conventional front axle brake circuit. These
prevent any gas that might escape from the
high-pressure accumulator from reaching
the brake circuit of the front wheels, which
would diminish the instantaneous braking
power in the event of a system failure.
3 Interaction of SBC functional modules
Brake functions Vehicle control systems

Brake operation Brake power assistance
p Braking force distribution Signal interfaces
s CAN
u ABS, TCS, ESP (gateway)
s
Control unit (ECU)
Pedal C
travel simulator A
pi
N
Hydraulic Intelligent interface
up u M
modulator p
Brake pressure modulators High-pressure
Isolating accumulator
valves
p p p p
Hydraulic energy
u u u u source
p p p p
æ UFB0635-1E
Brakes
LF RF LR RR
416 Overview of common-rail systems Areas of application
Overview of common-rail systems

The demands placed on diesel-engine Areas of application
fuel-injection systems are continuously
increasing. Higher pressures, faster switching The common-rail fuel-injection system
times, and a variable rate-of-discharge curve for engines with diesel direct injection
modified to the engine operating state have (Direct Injection, DI) is used in the following
made the diesel engine economical, clean, vehicles:
and powerful. As a result, diesel engines have 쐌 Passenger cars ranging from high-economy
even entered the realm of luxury- 3-cylinder engines with displacements of
performance sedans. 800 cc, power outputs of 30 kW (41 HP),
torques of 100 Nm, and a fuel consump-
One of the advanced fuel-injection systems is tion of 3.5 l/100 km through to 8-cylinder
the common-rail (CR) fuel-injection system. engines in luxury-performance sedans
The main advantage of the common-rail sys- with displacements of approx. 4 l, power
tem is its ability to vary injection pressure outputs of 180 kW (245 HP), and torques
and timing over a broad scale. of 560 Nm.
This was achieved by separating pressure gen- 쐌 Light-duty trucks with engines producing
eration (in the high-pressure pump) from the up to 30 kW/cylinder, and
fuel-injection system (injectors). The rail here 쐌 Heavy-duty trucks, railway locomotives, and
acts as a pressure accumulator. ships with engines producing up to approx.
200 kW/cylinder
1 Common-rail fuel-injection system taking the example of a five-cylinder diesel engine
2
5
3
4
Fig. 1
1 Fuel return line
2 High-pressure fuel 7
6
line to injector
3 Injector
4 Fuel rail
5 Rail-pressure sensor
æ UMK1991Y
6 High-pressure fuel 8
line to rail
7 Fuel return line
8 High-pressure pump

Overview of common-rail systems Areas of application, design 417
The common-rail system is a highly flexible Design

system for adapting fuel injection to the
engine. This is achieved by: The common-rail system consists of the
쐌 High injection pressure up to approx. following main component groups
1,600 bar, in future up to 1,800 bar. (Figs. 1 and 2):
쐌 Injection pressure adapted to the operating 쐌 The low-pressure stage, comprising the fuel-
status (200...1,800 bar). supply system components.
쐌 Variable start of injection. 쐌 The high-pressure system, comprising com-
쐌 Possibility of several pre-injection events ponents such as the high-pressure pump,
and secondary injection events (even fuel rail, injectors, and high-pressure fuel
highly retarded secondary injection lines.
events). 쐌 The electronic diesel control (EDC), consist-
ing of system modules, such as sensors, the
In this way, the common-rail system helps electronic control unit, and actuators.
to raise specific power output, lower fuel con-
sumption, reduce noise emission, and The key components of the common-rail
decrease pollutant emission in diesel engines. system are the injectors. They are fitted
Today common rail has become the most with a rapid-action valve (solenoid valve or
commonly used fuel-injection system for piezo-triggered actuator) which opens and
modern, high-rev passenger-car direct- closes the nozzle. This permits control of the
injection engines. injection process for each cylinder.
2 System modules of an engine control unit and a common-rail fuel-injection system
Electronic diesel control (EDC): engine management, sensors, interface
Fuel supply system Air-intake and

(low-pressure stage) exhaust-gas systems
Engine
1
2
3
æ NMK1871E
Signals
Fig. 2
High-pressure section Diesel fuel
2 Fuel rail
3 Injectors
418 Overview of common-rail systems Design, operating concept
All the injectors are fed by a common fuel Operating concept

rail, this being the origin of the term “com-
mon rail”. In the common-rail fuel-injection system, the
One of the main features of the common- functions of pressure generation and fuel in-
rail system is that system pressure is variable jection are separate. The injection pressure is
dependent on the engine operating point. generated independent of the engine speed
Pressure is adjusted by the pressure-control and the injected fuel quantity. The electronic
valve or the metering unit (Fig. 3). diesel control (EDC) controls each of the
components.
The modular design of the common-rail sys-
tem simplifies modification of the system to Pressure generation
different engines. Pressure generation and fuel injection are
separated by means of an accumulator vol-
ume. Fuel under pressure is supplied to the
accumulator volume of the common rail
ready for injection.
A continuously operating high-pressure
pump driven by the engine produces the
Fig. 3 desired injection pressure. Pressure in the fuel
a Pressure control on
rail is maintained irrespective of engine speed
the high-pressure
side by means of
or injected fuel quantity. Owing to the almost
pressure-control uniform injection pattern, the high-pressure
valve for passenger- pump design can be much smaller and its
car applications 3 Examples of high-pressure control for drive-system torque can be lower than con-
b Pressure control on common-rail systems
ventional fuel-injection systems. This results
the suction side with
in a much lower load on the pump drive.
a metering unit
flanged to the high- a
pressure pump (for The high-pressure pump is a radial-piston
4 5 6
passenger cars and pump. On commercial vehicles, an in-line
commercial vehicles) 8 fuel-injection pump is sometimes fitted.
1 7 7 7 7
c Pressure control on
the suction side with 2 3
Pressure control
a metering unit and
The pressure control method applied is
additional control
with a pressure-
largely dependent on the system.
control valve (for b
passenger cars) 9 5 6 Control on the high-pressure side
10 On passenger-car systems, the required rail
11 High-pressure pump 1 7 7 7 7 pressure is controlled on the high-pressure
12 Fuel inlet 2 3 side by a pressure-control valve (Fig. 3a, 4).
13 Fuel return
14 Pressure-control
Fuel not required for injection flows back
valve to the low-pressure circuit via the pressure-
15 Fuel rail c control valve. This type of control loop allows
16 Rail-pressure sensor 11 5 6 rail pressure to react rapidly to changes in
17 Injector connection
10
operating point (e.g. in the event of load
18 Return fuel 8 7 7 7 7 changes).
1
connection
æ SMK1993Y
19 Pressure-relief valve 2 3
10 Metering unit
11 Pressure-control
valve
Overview of common-rail systems Operating concept 419
Control on the high-pressure side was The injector opening times and system
adopted on the first common-rail systems. pressure determine the quantity of fuel
The pressure-control valve is mounted prefer- delivered. At a constant pressure, the fuel
ably on the fuel rail. In some applications, quantity delivered is proportional to the
however, it is mounted directly on the high- switching time of the solenoid valve. This is,
pressure pump. therefore, independent of engine or pump
speed (time-based fuel injection).
Fuel-delivery control on the suction side
Another way of controlling rail pressure is Potential hydraulic power
to control fuel delivery on the suction side Separating the functions of pressure genera-
(Fig. 3b). The metering unit (10) flanged on tion and fuel injection opens up further de-
the high-pressure pump makes sure that the grees of freedom in the combustion process
pump delivers exactly the right quantity of compared with conventional fuel-injection
fuel to the fuel rail in order to maintain the systems; the injection pressure is more or less
injection pressure required by the system. freely selectable within the program map.
In a fault situation, the pressure-relief valve The maximum injection pressure at present
(9) prevents rail pressure from exceeding a is 1,600 bar; in future this will rise to
maximum. 1,800 bar.
Fuel-delivery control on the suction side The common-rail system allows a further
reduces the quantity of fuel under high reduction in exhaust-gas emissions by intro-
pressure and lowers the power input of the ducing pre-injection events or multiple injec-
pump. This has a positive impact on fuel con- tion events and also attenuating combustion
sumption. At the same time, the temperature noise significantly. Multiple injection events
of the fuel flowing back to the fuel tank is re- of up to five per injection cycle can be gener-
duced in contrast to the control method on ated by triggering the highly rapid-action
the high-pressure side. switching valve several times. The nozzle-
needle closing action is hydraulically assisted
Two-actuator system to ensure that the end of injection is rapid.
The two-actuator system (Fig. 3c) combines
pressure control on the suction side via the
metering unit and control on the high-pres-
sure side via the pressure-control valve, thus
marrying the advantages of high-pressure-
side control and suction-side fuel-delivery
control (see the section on “Common-rail
system for passenger cars”).
Fuel injection
The injectors spray fuel directly into the en-
gine’s combustion chambers. They are sup-
plied by short high-pressure fuel lines con-
nected to the fuel rail. The engine control
unit controls the switching valve integrated in
the injector to open and close the injector
nozzle.
420 Overview of common-rail systems Operating concept
Control and regulation Correction functions for calculating

Operating concept fuel injection
The engine control unit detects the accelera- A number of correction functions are avail-
tor-pedal position and the current operating able to compensate for tolerances between the
states of the engine and vehicle by means of fuel-injection system and the engine (see the
sensors (see the section on “Electronic diesel section on “Electronic diesel control”):
control”). The data collected includes: 쐌 Injector delivery compensation
쐌 Crankshaft speed and angle 쐌 Zero delivery calibration
쐌 Fuel-rail pressure 쐌 Fuel-balancing control
쐌 Charge-air pressure 쐌 Average delivery adaption
쐌 Intake air, coolant temperature, and fuel
temperature Additional functions
쐌 Air-mass intake Additional open- and closed-loop control
쐌 Road speed, etc. functions perform the tasks of reducing ex-
haust-gas emissions and fuel consumption,
The electronic control unit evaluates the or providing added safety and convenience.
input signals. In sync with combustion, Some examples are:
it calculates the triggering signals for the pres- 쐌 Control of exhaust-gas recirculation
sure-control valve or the metering unit, the 쐌 Boost-pressure control
injectors, and the other actuators (e.g. the 쐌 Cruise control
EGR valve, exhaust-gas turbocharger 쐌 Electronic immobilizer, etc.
actuators, etc.).
Integrating EDC in an overall vehicle system
The injector switching times, which need to opens up a number of new opportunities,
be short, are achievable using the optimized e.g. data exchange with transmission control
high-pressure switching valves and a special or air-conditioning system.
control system.
The angle/time system compares injection A diagnosis interface permits analysis of
timing, based on data from the crankshaft stored system data when the vehicle is
and camshaft sensors, with the engine state serviced.
(time control). The electronic diesel control
(EDC) permits a precise metering of the in- Control unit configuration
jected fuel quantity. In addition, EDC offers As the engine control unit normally has
the potential for additional functions that can a maximum of only eight output stages
improve engine response and convenience. for the injectors, engines with more than
eight cylinders are fitted with two engine
Basic functions control units. They are coupled within the
The basic functions involve the precise con- “master/slave” network via an internal, high-
trol of diesel-fuel injection timing and fuel speed CAN interface. As a result, there is also
quantity at the reference pressure. In this way, a higher microcontroller processing capacity
they ensure that the diesel engine has low available. Some functions are permanently
consumption and smooth running character- allocated to a specific control unit (e.g. fuel-
istics. balancing control). Others can be dynami-
cally allocated to one or other of the control
units as situations demand (e.g. to detect
sensor signals).
Diesel boom in Europe 421
왘 Diesel boom in Europe
Diesel engine applications It is torque, and not engine performance,

At the start of automobile history, the spark- that is the decisive factor for engine power.
ignition engine (Otto cycle) was the drive unit Compared to a gasoline engine without
for road vehicles. The first time a diesel engine supercharging, a driver can experience more
was mounted on a truck was 1927. Passenger “driving pleasure” with a diesel engine of
cars had to wait until 1936. lower performance. The image of the “stinking
The diesel engine made strong headway in slowcoach” is simply no longer true for diesel-
the truck sector due to its fuel economy and engined cars of the latest generation.
long service life. By contrast, the diesel engine
in the car sector was long relegated to a Environmental compatibility
fringe existence. It was only with the introduc- The clouds of smoke that diesel-engined cars
tion of supercharged direct-injection diesel produced when driven at high loads are a
engines – the principle of direct injection was thing of the past. This was brought about by
already used in the first truck diesel engines – improved fuel-injection systems and electronic
that the diesel engine changed its image. diesel control (EDC). These systems can me-
Meanwhile, the percentage of diesel-engined ter fuel quantity with high precision, adjusting
passenger cars among new registrations is it to the engine operating point and environ-
fast approaching 50% in Europe. mental conditions. This technology also meets
prevailing exhaust-gas emission standards.
Features of the diesel engine Oxidation-type catalytic converters, that
What is the reason for the boom in diesel remove carbon monoxide (CO) and hydro-
engines in Europe? carbons (HC) from exhaust gas, are standard
equipment on diesel engines. Future, more
Fuel economy stringent exhaust-gas emission standards, and
Firstly, fuel consumption compared to gasoline even U.S. legislation, will be met by other ex-
engines is still lower – this is due to the greater haust-gas treatment systems, such as particu-
efficiency of the diesel engine. Secondly, late filters and NOx accumulator-type catalytic
diesel fuel is subject to lower taxes in most converters.
European countries. For people who travel a
lot, therefore, diesel is the more economical 왘 Typical torque and power curves of a
passenger-car diesel engine
alternative despite the higher purchase price.
Nm kW
Driving pleasure
Almost all diesel engines on the market are 360 100
supercharged. This produces a high cylinder 320 90

charge at low revs. The metered fuel quantity M
280 80
Torque M
Power P
can also be high, and this produces high en-

gine torque. The result is a torque curve that 240 70
P
permits driving at high torque and low revs. 200 60
160 50
40
æ SMK2023E
0 1,000 2,000 3,000 4,000 rpm

Engine speed
422 Overview of common-rail systems Common-rail system for passenger cars
Common-rail system for High-pressure control

On first-generation common-rail systems,
passenger cars rail pressure is controlled by the pressure-
Fuel supply control valve. The high-pressure pump
In common-rail systems for passenger cars, (type CP1) generates the maximum delivery
electric fuel pumps or gear pumps are used to quantity, irrespective of fuel demand. The
deliver fuel to the high-pressure pump. pressure-control valve returns excess fuel to
the fuel tank.
Systems with electric fuel pump Second-generation common-rail systems
The electric fuel pump is either part of the in- control rail pressure on the low-pressure side
tank unit (in the fuel tank) or is fitted in the by means of the metering unit (Figs. 1 and 2).
fuel line (in-line). It intakes fuel via a pre-fil- The high-pressure pump (types CP3 and
ter and delivers it to the high-pressure pump CP1H) need only deliver the fuel quantity
at a pressure of 6 bar (Fig. 3). The maximum that the engine actually requires. This lowers
delivery rate is 190 l/h. To ensure fast engine the energy demand of the high-pressure
starting, the pump switches on as soon as the pump and reduces fuel consumption.
driver turns the ignition key. This builds up Third-generation common-rail systems
the necessary pressure in the low-pressure feature piezo-inline injectors (Fig. 3).
circuit when the engine starts.
The fuel filter (fine filter) is fitted in the If pressure is only adjustable on the low-pres-
supply line to the high-pressure pump. sure side, it takes too long to lower the pres-
sure in the fuel rail when rapid negative load
Systems with gear pump changes occur. Adapting pressure to dynamic
The gear pump is flanged to the high-pres- changes in load conditions is then too slow.
sure pump and is driven by its input shaft This is particularly the case with piezo-inline
(Figs. 1 and 2). In this way, the gear pump injectors due to their very low internal leak-
starts delivery only after the engine has age. For this reason, some common-rail sys-
started. Delivery rate is dependent on the tems are equipped with an additional pres-
engine speed and reaches rates up to 400 l/h sure-control valve (Fig. 3) besides the high-
at pressures up to 7 bar. pressure pump and metering unit. This
A fuel pre-filter is fitted in the fuel tank. two-actuator system combines the advantages
The fine filter is located in the supply line to of control on the low-pressure side with the
the gear pump. dynamic response of control on the high-
pressure side.
Combination systems Another advantage compared with control
There are also applications where the two on the low-pressure side only is that the high-
pump types are used. The electric fuel pump pressure side is also controllable when the
improves starting response, in particular for engine is cold. The high-pressure pump then
hot starts, since the delivery rate of the gear delivers more fuel than is injected and pres-
pump is lower when the fuel is hot, and sure is controlled by the pressure-control
therefore, thinner, and at low pump speeds. valve. Compression heats the fuel, thus elimi-
nating the need for an additional fuel heater.
Overview of common-rail systems Common-rail system for passenger cars 423
1 Example of a second-generation common-rail system for a 4-cylinder engine
Fig. 1
6 1 High-pressure pump
1 5 8 (CP3) with fitted
geared presupply
pump and metering
unit
2 Fuel filter with water
separator and heater
(optional)
2 3 Fuel tank
4 Pre-filter
3 5 Fuel rail
7
æ SMK2017Y
7 Solenoid-valve
4 injector
8 Pressure-relief valve
2 Example of a second-generation common-rail system with two-actuator system for a V8 engine Fig. 2
5 1 High-pressure pump
(CP3) with fitted
8 geared presupply
1 6 pump and metering
9 5 unit
(optional)
3 Fuel tank
4 Pre-filter
2 5 Fuel rail
3 7 Solenoid-valve
7
æ SMK2018Y
injector
8 Pressure-control valve
4 9 Function module
(distributor)
3 Example of a third-generation common-rail system with two-actuator system for a 4-cylinder engine
Fig. 3
1 8 5 6 1 High-pressure pump
(CP1H) with
metering unit
(optional)
3 Fuel tank
2 4 Pre-filter
5 Fuel rail
3
7 6 Rail-pressure sensor
æ SMK2019Y
7 Piezo-inline injector
9 8 Pressure-control
4 valve
9 Electric fuel pump
424 Overview of common-rail systems Common-rail system for passenger cars
System diagram for passenger cars Data exchange between the various sections
Fig. 4 shows all the components in a takes place via the CAN bus in the “Inter-
common-rail system for a fully equipped, faces” (B) section:
4-cylinder, passenger-car diesel engine. 쐌 Starter motor
Depending on the type of vehicle and its 쐌 Alternator
application, some of the components may 쐌 Electronic immobilizer
not be fitted. 쐌 Transmission control
쐌 Traction Control System (TSC)
The sensors and setpoint generators (A) 쐌 Electronic Stability Program (ESP)
are not depicted in their real installation
position to simplify presentation. Exceptions The instrument cluster (13) and the air-con-
are the exhaust-gas treatment sensors (F) and ditioning system (14) are also connectable to
the rail-pressure sensor as their installation the CAN bus.
positions are required to understand the sys-
tem. Two possible combined systems are described
(a or b) for exhaust-gas treatment.
Fig. 4
Engine, engine management, and high-pressure C Fuel-supply system (low-pressure stage)
fuel-injection components 19 Fuel filter with overflow valve
17 High-pressure pump 20 Fuel tank with pre-filter and Electric Fuel Pump,
18 Metering unit EFP (presupply pump)
25 Engine ECU 21 Fuel-level sensor
26 Fuel rail
27 Rail-pressure sensor D Additive system
28 Pressure-control valve (DRV 2) 22 Additive metering unit
29 Injector 23 Additive control unit
30 Glow plug 24 Additive tank
31 Diesel engine (DI)
M Torque E Air supply
32 Exhaust-gas recirculation cooler
A Sensors and setpoint generators 33 Boost-pressure actuator
1 Pedal-travel sensor 34 Turbocharger (in this case with Variable Turbine
2 Clutch switch Geometry (VTG))
3 Brake contacts (2) 35 Control flap
4 Operator unit for vehicle-speed controller (cruise control) 36 Exhaust-gas recirculation actuator
5 Glow-plug and starter switch (“ignition switch”) 37 Vacuum pump
6 Road-speed sensor
7 Crankshaft-speed sensor (inductive) F Exhaust-gas treatment
8 Camshaft-speed sensor (inductive or Hall sensor) 38 Broadband lambda oxygen sensor, type LSU
9 Engine-temperature sensor (in coolant circuit) 39 Exhaust-gas temperature sensor
10 Intake-air temperature sensor 40 Oxidation-type catalytic converter
11 Boost-pressure sensor 41 Particulate filter
12 Hot-film air-mass meter (intake air) 42 Differential-pressure sensor
43 NOx accumulator-type catalytic converter
B Interfaces 44 Broadband lambda oxygen sensor, optional NOx sensor
13 Instrument cluster with displays for fuel consumption,
engine speed, etc.
14 Air-conditioner compressor with operator unit
15 Diagnosis interface
16 Glow control unit
CAN Controller Area Network
(on-board serial data bus)
Overview of common-rail systems Common-rail system for passenger cars 425
4 Common-rail diesel fuel-injection system for cars
B 19 C D
CAN
13
17 22
20
18
21
14
15
27
26 28
16
23 24
A
29
1 30
25
2
32
31
3
M
E
4
35
5
36 37
34
6 33
7 F
a
38 39 39
8
40 41
9
42
10 b
39 38 44 39
11 43 41
æ SMK1819-1Y
12 42
426 Overview of diesel fuel-injection systems
왘 Overview of diesel fuel-injection systems
Areas of application Requirements

Diesel engines are characterized by high fuel Ever stricter statutory regulations on noise and
economy. Since the first volume-production exhaust-gas emissions and the desire for more
fuel-injection pump was introduced by Bosch economical fuel consumption continually place
in 1927, fuel-injection systems have experi- greater demands on the fuel-injection system
enced a process of continuous development. of a diesel engine.
Diesel engines are used in a wide variety of Basically, the fuel-injection system is required
design for many different purposes (Fig. 1 and to inject a precisely metered amount of fuel at
Table 1), for example high pressure into the combustion chamber
쐌 To drive mobile power generators in such a way that it mixes effectively with the
(up to approx. 10 kW/cylinder) air in the cylinder as demanded by the type
쐌 As fast-running engines for cars of engine (direct or indirect-injection) and its
and light-duty trucks present operating status. The power output
(up to approx. 50 kW/cylinder) and speed of a diesel engine is controlled by
쐌 As engines for construction-industry means of the injected fuel quantity as it has no
and agricultural machinery air intake throttle.
(up to approx. 50 kW/cylinder)
쐌 As engines for heavy trucks, Mechanical control of diesel fuel-injection
omnibuses and tractor vehicles systems is being increasingly replaced by
(up to approx. 80 kW/cylinder) Electronic Diesel Control (EDC) systems.
쐌 To drive fixed installations such All new diesel-injection systems for cars
as emergency power generators and commercial vehicles are electronically
(up to approx. 160 kW/cylinder) controlled.
쐌 As engines for railway locomotives
and ships (up to 1,000 kW/cylinder)
1 Applications for Bosch diesel fuel-injection systems
M M M A/P P/H ZWM ZWM

MW MW MW MW CW CW
PF PF PF PF PF PF
Fig. 1
M, MW,
VE VE VE VE VE VE
A, P, H,
ZWM,
CW In-line fuel-injection VR VR VR VR VR
pumps of increas-
ing size
PF Discrete fuel- UIS UIS UIS UIS UIS
injection pumps
VE Axial-piston pumps UPS UPS UPS UPS
VR Radial-piston PF(R) PF(R)
æ UMK1563-4Y
pumps
UIS Unit injector system CR CR CR CR CR CR
UPS Unit pump system
CR Common-rail
system
Overview of common-rail systems Common-rail system for commercial vehicles 427
Common-rail system for (Figs. 1 and 2). In many applications, it is

mounted on the engine.
commercial vehicles
Fuel supply Fuel filtering
Presupply As opposed to passenger-car systems, the fuel
Common-rail systems for light-duty trucks filter (fine filter) is fitted to the pressure side.
differ very little from passenger-car systems. For this reason, an exterior fuel inlet is re-
Electric fuel pumps or gear pumps are used quired, in particular when the gear pump is
for fuel presupply. On common-rail systems flanged to the high-pressure pump.
for heavy-duty trucks, only gear pumps are
used to deliver fuel to the high-pressure
pump (see the subsection “Gear-type supply
pumps” in the section “Fuel supply in the
low-pressure stage”). The presupply pump is
normally flanged to the high-pressure pump
1 Common-rail system for commercial vehicles with high-pressure pump (CP3)
7 8 9
5 Fig. 1
11 Fuel tank
12 Pre-filter
4 3
6 13 Fuel filter
14 Gear presupply
pump
(CP3.4)
1 16 Metering unit
æ SMK2025Y
18 Fuel rail
2 19 Pressure-relief valve
10 Injector
2 Common-rail system for commercial vehicles with high-pressure pump (CPN2)
6 7 8 9 Fig. 2
5 3 11 Fuel tank
12 Pre-filter
4 13 Fuel filter
14 Gear presupply
pump
(CPN2.2)
1 16 Metering unit
æ SMK2026Y
18 Fuel rail
2 19 Pressure-relief valve
10 Injector
428 Overview of common-rail systems Common-rail system for commercial vehicles
System diagram for commercial vehicles Program (ESP), oil-grade sensor, trip
Fig. 3 shows all the components in a com- recorder, Active Cruise Control (ACC), brake
mon-rail system for a 6-cylinder commercial- coordinator – up to 30 ECUs). The alternator
vehicle diesel engine. Depending on the type (18) and the air-conditioning system (17) are
of vehicle and its application, some of the also connectable to the CAN bus.
components may not be fitted. Three systems are described for exhaust-
Only the sensors and setpoint generators gas treatment: a purely DPF system (a)
are depicted at their real position to simplify mainly for the U.S. market, a purely SCR
presentation, as their installation positions system (b) mainly for the EU market, and
are required to understand the system. a combined system (c).
Data exchange to the various sections takes
place via the CAN bus in the “Interfaces” (B)
section (e.g. transmission control, Traction
Control System (TCS), Electronic Stability
Fig. 3 20 SCR control unit

Engine, engine management, and high-pressure 21 Air compressor
fuel-injection components CAN Controller Area Network (on-board serial data bus)
22 High-pressure pump (up to three data buses)
29 Engine ECU
30 Fuel rail C Fuel-supply system (low-pressure stage)
31 Rail-pressure sensor 23 Fuel presupply pump
32 Injector 24 Fuel filter with water-level and pressure sensors
33 Relay 25 Control unit cooler
34 Auxiliary equipment (e.g. retarder, exhaust flap 26 Fuel tank with pre-filter
for engine brake, starter motor, fan) 27 Pressure-relief valve
35 Diesel engine (DI) 28 Fuel-level sensor
36 Flame glow plug (alternatively grid heater)
M Torque D Air intake
37 Exhaust-gas recirculation cooler
A Sensors and setpoint generators 38 Control flap
1 Pedal-travel sensor 39 Exhaust-gas recirculation positioner with exhaust-gas
2 Clutch switch recirculation valve and position sensor
3 Brake contacts (2) 40 Intercooler with bypass for cold starting
4 Engine brake contact 41 Exhaust-gas turbocharger (in this case with variable
5 Parking brake contact turbine geometry) with position sensor
6 Operating switch (e.g. vehicle-speed controller, interme- 42 Boost-pressure actuator
diate-speed regulation, rpm- and torque reduction)
7 Starter switch (“ignition lock”) E Exhaust-gas treatment
8 Turbocharger-speed sensor 43 Exhaust-gas temperature sensor
9 Crankshaft-speed sensor (inductive) 44 Oxidation-type catalytic converter
10 Camshaft-speed sensor 45 Differential-pressure sensor
11 Fuel-temperature sensor 46 Catalyst-coated particulate filter (CSF)
12 Engine-temperature sensor (in coolant circuit) 47 Soot sensor
13 Boost-air temperature sensor 48 Level sensor
14 Boost-pressure sensor 49 Reducing-agent tank
15 Fan-speed sensor 50 Reducing-agent pump
16 Air-filter differential-pressure sensor 51 Reducing-agent injector
52 NOx sensor
B Interfaces 53 SCR catalytic converter
17 Air-conditioner compressor with operator unit 54 NH3 sensor
18 Alternator
19 Diagnosis interface
Overview of common-rail systems Common-rail system for commercial vehicles 429
3 Common-rail diesel fuel-injection system for commercial vehicles
B 23 24 C
CAN
17 22 27 25
28 26
18 G
19
20 31
30
29
21
A 32
1 33
34
37
3
36 35
38
4
39 M
D
5
40
6
7 41
42
8
45 E
43
9
a 44 46
10
48 50
11
49
52 or
12 51 43 54
43
b 44 53
13
14 48 50
49 45 52 or
15 54 or
51 43
43 47
æ SMK1820-1Y
16 c 44 46 53
4 High-pressure components of common-rail system Overview
High-pressure components of common-rail system

The high-pressure stage of the common-rail fuel injected. High-pressure fuel lines inter-
system is divided into three sections: pres- connect the three sections.
sure generation, pressure storage, and fuel
metering. The high-pressure pump assumes Overview
the function of pressure generation. Pres-
sure storage takes place in the fuel rail to The main difference in the various genera-
which the rail-pressure sensor and the pres- tions of common-rail systems lie in the de-
sure-control and pressure-relief valves are sign of the high-pressure pump and the in-
fitted. The function of the injectors is cor- jectors, and in the system functions required
rect timing and metering the quantity of (Table 1).
1 Overview of common-rail systems
CR generation Maximum pressure Injector High-pressure pump
1st generation 1,350...1,450 bar Solenoid-valve injector CP1

Pass. cars Pressure control on high-pressure side by pressure-
control valve
1st generation 1,400 bar Solenoid-valve injector CP2
Comm. veh. Suction-side fuel-delivery control by two solenoid valves
2nd generation 1,600 bar Solenoid-valve injector CP3, CP1H
Pass. cars and Suction-side fuel-delivery control by metering unit
comm. veh.
3rd generation 1,600 bar Piezo-inline injector CP3, CP1H
Pass. cars (in future 1,800 bar) Suction-side fuel-delivery control by metering unit
3rd generation 1,800 bar Solenoid-valve injector CP3.3NH
Comm. veh. Metering unit
Table 1
1 Common-rail fuel-injection system taking the example of a four-cylinder diesel engine
1 2 3
Fig. 1
1 Hot-film air-mass
meter
2 Engine ECU
4 High-pressure
accumulator
(fuel rail)
5 Injector
6 Crankshaft-speed
æ UMK1566-1Y
sensor
7 Engine-temperature
sensor
8 Fuel filter 4 5 6 7 8 9
9 Pedal-travel sensor

Electronic control unit Cleanliness requirements
왘 *SLHUSPULZZYLX\PYLTLU[Z
Cleanliness quality
왘 7YPUJPWSLVM[OLWHY[PJSLHUHS`ZPZZ`Z[LT:,4
;OLZOHYWYPZLPU[OLWLYMVYTHUJLVMUL^
Evolution of particulate-analysis process
HZZLTISPLZLN [OLJVTTVUYHPSZ`Z[LTMVY Particulate Microparticles
OPNOWYLZZ\YLKPLZLSM\LSPUQLJ[PVUYLX\PYLZL_
up to < 1μm
[YLTLWYLJPZPVUPUTHJOPUPUNHUKL]LY[PNO[LY
[VSLYHUJLZHUKMP[Z7HY[PJSLYLZPK\LMYVT[OL Light Electron
WYVK\J[PVUWYVJLZZTH`SLHK[VPUJYLHZLK microscope microscope
^LHYVYL]LU[OL[V[HSMHPS\YLVM[OLHZZLTIS` Growth in information content

;OPZYLZ\S[ZPUOPNOYLX\PYLTLU[ZHUK[PNO[[VS
- Number of particles - EDX analysis
LYHUJLZMVYJSLHUSPULZZX\HSP[`^P[OHJVU[PU\
- Size distribution - Number of particles
V\ZYLK\J[PVUPU[OLWLYTP[[LKWHY[PJSLZPaL - Size distribution
;OLJSLHUSPULZZX\HSP[`VMJVTWVULU[Z - High precision
(focus depth)
PZJ\YYLU[S`KL[LYTPULKPU[OLWYVK\J[PVU
WYVJLZZI`SPNO[TPJYVZJVWLPTHNLHUHS`ZPZ
Z`Z[LTZ;OL`Z\WWS`PUMVYTH[PVUVUWHY[PJSL The electron beam and its action
ZPaLKPZ[YPI\[PVU(KKP[PVUHSPUMVYTH[PVUZ\JO
HZ[OLUH[\YLVM[OLWHY[PJSLZHUK[OLPYJOLTP
JHSJVTWVZP[PVUPZYLX\PYLK[VKL]LSVWPUUV]H
[P]LJSLHUPUNWYVJLZZLZ;OPZPUMVYTH[PVUPZ Backscatter Secondary
electrons electrons
VI[HPULKI`LSLJ[YVUTPJYVZJVWLZ and X-rays from depth
from depth of several
of several Heat μm
Particle-analysis system (SEM) μm
)VZJO\ZLZHWHY[PJSLHUHS`ZPZZ`Z[LTIHZLK Primary electron beam 20 kV
Backscatter electrons to BSE detector (up to 20 keV)
VUHSJHUUPUNESLJ[YVUMPJYVZJVWL:,4 Secondary electrons to SE detector (several eV)
;OPZZ`Z[LTWLYMVYTZHUH\[VTH[LKHUHS`ZPZ X-rays to EDX detector (up to 10 keV)
VMWHY[PJSLZHKOLYPUN[VHWYVK\J[;OLYLZ\S[Z
VM[OLHUHS`ZPZZOV^WHY[PJSLZPaLKPZ[YPI\[PVU
Detection of interaction between
[OLJOLTPJHSJVTWVZP[PVUVM[OLWHY[PJSLZHUK electron beam and sample
PTHNLZVM[OLPUKP]PK\HSWHY[PJSLZ<ZPUN[OPZ
SE detector
PUMVYTH[PVU[OLZV\YJLVM[OLWHY[PJSLZHYL Secondary electrons of sample surface
PKLU[PMPHISL(J[PVUJHU[OLUIL[HRLU[VH]VPK are converted into image signals.
- Plastic image of surface
YLK\JLVY^HZOVMMJLY[HPUWHY[PJSL[`WLZ (REM images).
0U[OPZ^H`ZVS\[PVUZHYLUV[IHZLKVU[OL
BSE detector
PUJYLHZLK\ZLVMJSLHUPUN[LJOUPX\LZI\[VU Backscatter electrons are converted
H]VPKPUNHUKYLK\JPUNYLZPK\HSZVPSPUNK\YPUN into image signals.
Phase composition
[OLWYVK\J[PVUWYVJLZZ TOPO mode plastic image
;OLH\[VTH[LKWHY[PJSLHUHS`ZPZZ`Z[LT
æ :(5 ,
EDX detector
:,4WYV]PKLZ[OLJSLHUSPULZZWYVJLZZ^P[O Characteristic X-ray is converted into
HUHUHS`ZPZZ`Z[LT[OH[WYVK\JLZPTWVY[HU[ "energy-dispersive" spectrum.
Identification of chemical elements
PUMVYTH[PVUVU[OL[`WLVMYLZPK\HSZVPSPUN
;OLWYLJPZLPKLU[PMPJH[PVUVMWHY[PJSLZHUK[OLPY
ZV\YJLZPZ]P[HS[VKL]LSVWPUNUL^JSLHUPUN
[LJOUPX\LZ
432 High-pressure components of common-rail system Injector
Injector of sensors to detect the crankshaft position

and the camshaft position (phase detection).
On a common-rail diesel injection system, An optimum mixture formation is re-
the injectors are connected to the fuel rail by quired to reduce exhaust-gas emissions and
short high-pressure fuel lines. The injectors comply with continuous demands to reduce
are sealed to the combustion chamber by a the noise of diesel engines. This calls for
copper gasket. The injectors are fitted into injectors with very small pre-injection quan-
the cylinder head by means of taper locks. tities and multiple injection events.
Depending on the injection-nozzle design,
common-rail injectors are intended for There are presently three different injector
straight or inclined mounting in direct- types in serial production:
injection diesel engines. 쐌 Solenoid-valve injectors with one-part
armature
One of the system’s features is that it gener- 쐌 Solenoid-valve injectors with two-part
ates an injection pressure irrespective of armature
engine speed or injected fuel quantity. The 쐌 Injector with piezo actuator
start of injection and injected fuel quantity
are controlled by the electrically triggered
injector. The injection time is controlled
by the angle/time system of the Electronic
Diesel Control (EDC). This requires the use
1 Solenoid-valve injector (functional schematic)
a b c
1
Fig. 1
a Resting position
b Injector opens
c Injector closes 2 11
11 Fuel-return 3
12 Solenoid coil 12
4
13 Overstroke spring
14 Solenoid armature 13
5
15 Valve ball
6
16 Valve-control
14
chamber
17 Nozzle spring
7 15
18 Pressure shoulder
of nozzle needle
19 Chamber volume
8
10 Injection orifice
11 Solenoid-valve
spring 9
12 Outlet restrictor 16
13 High-pressure
æ UMK1855-1Y
connection 10
14 Inlet restrictor
15 Valve plunger
(control plunger)
16 Nozzle needle
High-pressure components of common-rail system Injector 433
Solenoid-valve injector Injector opens (start of injection)

Design To begin with, the injector is in its resting po-
The injector can be subdivided into a number sition. The solenoid valve is triggered by the
of function modules: “pickup current”. This makes the solenoid
쐌 The hole-type nozzle (see the section on valve open very rapidly (Fig. 1b). The re-
“Injection nozzle”) quired rapid switching times are achieved by
쐌 The hydraulic servo system controlling solenoid-valve triggering in the
쐌 The solenoid valve ECU at high voltages and currents.
The magnetic force of the now triggered
Fuel is conveyed by the high-pressure connec- electromagnet exceeds the force of the valve
tion (Fig. 1a, 13) via a supply passage to the spring. The armature raises the valve ball
injection nozzle and via an inlet restrictor from the valve seat and opens the outlet
(14) to the valve-control chamber (6). The restrictor. After a short time the increased
valve-control chamber is connected to the pickup current is reduced to a lower holding
fuel return (1) via the outlet restrictor (12) current in the electromagnet. When the outlet
which can be opened by a solenoid valve. restrictor opens, fuel flows from the valve-
control chamber to the cavity above and then
Operating concept via the fuel-return line to the fuel tank. The
The function of the injector can be subdi- inlet restrictor (14) prevents a complete pres-
vided into four operating states when the sure compensation. As a result, pressure in
engine and the high-pressure pump are the valve-control chamber drops. Pressure in
operating: the valve-control chamber falls below the
쐌 Injector closed (with high pressure pressure in the nozzle chamber, which is still
applied) the same as the pressure in the fuel rail. The
쐌 Injector opens (start of injection) reduction in pressure in the valve-control
쐌 Injector fully open chamber reduces the force acting on the con-
쐌 Injector closes (end of injection) trol plunger and opens the nozzle needle. Fuel
injection commences.
The operating states are caused by the balance
of forces acting on the injector components. Injector fully open
When the engine is not running and the fuel The rate of movement of the nozzle needle is
rail is not pressurized, the nozzle spring closes determined by the difference in the flow rates
the injector. through the inlet and outlet restrictors. The
control plunger reaches its upper stop and
Injector closed (resting position) dwells there on a cushion of fuel (hydraulic
In its resting position, the injector is not trig- stop). The cushion is created by the flow of
gered (Fig. 1a). The solenoid-valve spring fuel between the inlet and outlet restrictors.
(11) presses the valve ball (5) onto the seat of The injector nozzle is then fully open. Fuel is
the outlet restrictor (12). Inside of the valve- injected into the combustion chamber at a
control chamber, the pressure rises to the pressure approaching that in the fuel rail.
pressure in the fuel rail. The same pressure is The balance of forces in the injector is
also applied to the chamber volume (9) of the similar to that during the opening phase.
nozzle. The forces applied by the rail pressure At a given system pressure, the fuel quantity
to the end faces of the control plunger (15), injected is proportional to the length of time
and the force of the nozzle spring (7) retain that the solenoid valve is open. This is entirely
the nozzle needle closed against the opening independent of the engine or pump speed
force applied to its pressure shoulders (8). (time-based injection system).
Injector closes (end of injection) Another feature of the fuel-quantity map is

When the solenoid valve is no longer trig- the flat curve that occurs with small triggering
gered, the valve spring presses the armature periods. The flat curve is caused by the sole-
down and the valve ball closes the outlet re- noid armature rebounding on opening. In this
strictor (Fig. 1c). When the outlet restrictor section, the injected fuel quantity is indepen-
closes, pressure in the control chamber rises dent on the triggering period. This allows
again to that in the fuel rail via the inlet re- small injected fuel quantities to be represented
strictor. The higher pressure exerts a greater as stable. Only after the armature has stopped
force on the control plunger. The force on rebounding does the injected fuel quantity
the valve-control chamber and the nozzle- curve continue to rise linearly as the triggering
spring force then exceed the force acting period becomes longer.
on the nozzle needle, and the nozzle needle Injection events with small injected fuel
closes. The flow rate of the inlet restrictor de- quantities (short triggering periods) are used
termines the closing speed of the nozzle nee- as pre-injection in order to suppress noise.
dle. The fuel-injection cycle comes to an end Secondary injection events help to enhance
when the nozzle needle is resting against its soot oxidation in selected sections of the
seat, thus closing off the injection orifices. operating curve.
This indirect method is used to trigger the Program maps without fuel-quantity flat curve
nozzle needle by means of a hydraulic servo The increasing stringency of emission-control
system because the forces required to open the legislation has lead to the use of the two sys-
nozzle needle rapidly cannot be generated tem functions: injector delivery compensation
directly by the solenoid valve. The “control (IMA) and zero delivery calibration (NMK), as
volume” required in addition to the injected well as to short intervals in injection between
fuel quantity reaches the fuel-return line via pre-injection, main injection, and secondary
the restrictors in the control chamber. injection events. With injectors that have no
In addition to the control volume, there are flat-curve section, IMA allows a precise ad-
also leakage volumes through the nozzle-nee- justment of the pre-injection fuel quantity
dle and valve-plunger guides. The control and when new. NMK corrects fuel-quantity drifts
leakage volumes are returned to the fuel tank over time in the low-pressure section. The key
via the fuel-return line and a collective line condition for deploying these two system
that comprises an overflow valve, high-pres- functions is a constant, linear rise in quantity,
sure pump, and pressure-control valve. i.e. there is no flat curve in the fuel-quantity
map (Fig. 2c). If the valve plunger/nozzle nee-
Program-map variants dle unit is operated in nominal mode without
Program maps with fuel-quantity flat curve lift-stop at the same time, this represents a
With injectors, a distinction is made in the fully ballistic operating mode of the valve
program map between ballistic and non- plunger and there is no kink in the fuel-quan-
ballistic modes. The valve plunger/nozzle tity map.
needle unit reaches the hydraulic stop if the
triggering period in vehicle operation is of Injector variants
sufficient length (Fig. 2a). The section until A distinction is made between two different
the nozzle needle reaches its maximum stroke solenoid-valve concepts with solenoid-valve
is termed ballistic mode. The ballistic and injectors:
nonballistic sections in the fuel-quantity 쐌 Injectors with one-part armature
map, where the injected fuel quantity is ap- (1-spring system)
plied for the triggering period (Fig. 2b), is 쐌 Injectors with two-part armature
separated by a kink in the program map. (2-spring system)
The short intervals between injection events on closing can end faster by decoupling the
are ensured when the armature can return to armature masses and adapting the setting
its resting position very rapidly on closing. parameters. This helps to achieve shorter
This is best achieved by a two-part armature intervals between two injection events with
with an overstroke stop. During the closing the two-part armature concept.
process, the armature plate moves down by
positive locking. The bottoming-out of the
armature plate is limited by an overstroke
stop. As a result, the armature reaches its
resting position faster. Armature rebound
2 Needle lift and fuel-quantity maps of an injector with lift-stop
a
Valve plunger/nozzle needle unit
at hydraulic stop
Nonballistic section
Needle lift
Ballistic section
Accumulation due to rail pressure
Time t
b
Rail pressure
quantity
Flat curve
Nonballistic
Einspritzmenge
section
Injected-fuel
Ballistic
section
c
Rail pressure
quantity
Einspritzmenge
Injected-fuel
Full ballistic section without lift-stop

æ UMK1983E
Triggering period
Triggering the solenoid-valve injector Holding-current phase

In its resting position, the injector’s high- In order to reduce power loss in the ECU and
pressure solenoid valve is not triggered and is injector, the current is dropped to approx.
therefore closed. The injector injects when 13 A in the holding-current phase.
the solenoid valve opens. The energy which becomes available when
Triggering the solenoid valve is divided into pickup current and holding current are
five phases (Figs. 3 and 4). reduced is routed to the booster-voltage
capacitor.
Opening phase
Initially, in order to ensure tight tolerances Switchoff
and high levels of reproducibility for the in- When the current is switched off in order
jected fuel quantity, the current for opening to close the solenoid valve, the surplus energy
the solenoid valve features a steep, precisely is also routed to the booster-voltage capacitor.
defined flank and increases rapidly up to
approx. 20 A. This is achieved by means of a Recharging the step-up chopper
booster voltage of up to 50 V. It is generated in Recharging takes place by means of a step-up
the control unit and stored in a capacitor chopper integrated in the ECU. The energy
(booster-voltage capacitor). When this volt- tapped during the opening phase is recharged
age is applied across the solenoid valve, the at the start of the pickup phase until the
current increases several times faster than it original voltage required to open the solenoid
does when only battery voltage is used. valve is reached.
Pickup-current phase
During the pickup-current phase, battery
voltage is applied to the solenoid valve and
assists in opening it quickly. Current control
limits pickup current to approx. 20 A.
3 Triggering sequence of a high-pressure solenoid valve for a single injection event
a b c d e
Solenoid-valve
current IM
Solenoid-valve
needle lift hM
Fig. 3
a Opening phase
b Pickup-current
phase Injected fuel
quantity Q
æ SAE0743-1E
c Transition to holding-
current phase
d Holding-current
phase Time t
e Switchoff
4 Common-rail system: Block diagram of the triggering phases for a cylinder group
2 I 4
a Opening phase
3 3 I
6
1
5
I 7 7 6
I
b Pickup-current phase
c Transition to I
holding-current phase
I I
I
d Holding-current phase
I Fig. 4
e Switchoff 11 Battery
12 Current control
I I 13 Solenoid windings
of the high-pressure
solenoid valves
14 Booster switch
15 Booster-voltage
capacitor
f Recharging the DC/DC converter:
16 Free-wheeling
step-up chopper recharging the energy Energy transfer
accumulator (9) (from 9 to 5) diodes for energy
recovery and high-
9 speed quenching
10 17 Cylinder selector
1
æ SMK1757-1E
switch
5 18 DC/DC switch
8
19 DC/DC coil
10 DC/DC diode
I Current flow
Piezo-inline injector acting on the nozzle needle. Such forces

Design and requirements occurred as a result of the push rod used
The design of the piezo-inline injector on previous solenoid-valve injectors. On
is divided into its main modules in the aggregate, this design effectively reduces the
schematic (see Fig. 5): moving masses and friction, thus enhancing
쐌 Actuator module (3) injector stability and drift compared to con-
쐌 Hydraulic coupler or translator (4) ventional systems.
쐌 Control or servo valve (5) In addition, the fuel-injection system
쐌 Nozzle module (6) allows the implementation of very short in-
tervals (“hydraulic zero”) between injection
The design of the injector took account of the events. The number and configuration of
high overall rigidity required within the actu- fuel-metering operations can represent up
ator chain composed of actuator, hydraulic to five injection events per injection cycle
coupler, and control valve. Another design in order to adapt the requirements to the
feature is the avoidance of mechanical forces engine operating points.
A direct response of the needle to actuator
5 Construction of the piezo-inline injector operation is achieved by coupling the servo
valve (5) to the nozzle needle. The delay
1 2
between the electric start of triggering and
hydraulic response of the nozzle needle is
about 150 microseconds. This meets the con-
tradictory requirements of high needle speeds
and extremely small reproducible injected
fuel quantities.
As a result of this principle, the injector also
includes small direct leakage points from the
high-pressure section to the low-pressure cir-
cuit. The result is an increase in the hydraulic
efficiency of the overall system.
3
Operating concept
Function of the 3/2-way servo valve in the
CR injector
The nozzle needle on piezo-inline injector
4 is controlled indirectly by a servo valve.
The required injected fuel quantity is then
controlled by the valve triggering period.
In its non-triggered state, the actuator is
Fig. 5 5 in the starting position and the servo valve
1 Fuel return
2 High-pressure
is closed (Fig. 6a), i.e. the high-pressure
connection section is separated from the low-pressure
3 Piezo actuator 6 section.
module
4 Hydraulic coupler
(translator)
æ UMK1974-1Y
5 Servo valve
(control valve)
6 Nozzle module with 7
nozzle needle
7 Injection orifice
The nozzle is kept closed by the rail pressure bration (NMK). The pre-injection quantity
exerted in the control chamber (3). When the can then be selected at will, and IMA can
piezo actuator is triggered, the servo valve minimize the quantity spread in the program
opens and closes the bypass passage (Fig. 6b). map using full ballistic mode (see Fig. 7).
The flow-rate ratio between the outlet restric-
tor (2) and the inlet restrictor (4) lowers pres-
sure in the control chamber and the nozzle
(5) opens. The control volume flows via the 7 Injection-quantity program map of the
servo valve to the low-pressure circuit of the piezo-inline injector
overall system.
To start the closing process, the actuator mm 3
Lift 5
is discharged and the servo valve releases
the bypass passage. The control chamber is 100
then refilled by reversing the inlet and outlet a
restrictors, and pressure in the control cham- b
80 c
ber is raised. As soon as the required pressure 0 d
Injected-fuel quantity
is attained, the nozzle needle starts to move 0.1 0.3

and the injection process ends.
60
The valve design described above and
the greater dynamic design of the actuator
system result in much shorter injection times 40
compared to injectors of conventional design, e Fig. 7
i.e. push rod and 2/2-way valve. Ultimately, Injected fuel quantities
this has a positive impact on exhaust-gas 20 at different injection
emissions and engine performance. Due to pressures
a 1,600 bar
æ UMK1984E
requirements regarding the engine in EU 4, 0 b 1,200 bar
the injector program maps were optimized to 0 0.4 0.8 1.2 ms c 1,000 bar
apply corrective functions (injector delivery Triggering period d 800 bar
compensation (IMA) and zero delivery cali- e 250 bar
6 Function of the servo valve
a b c
Fig. 6
1 a Start position
b Nozzle needle opens
(bypass closed,
normal function
2 with outlet and
inlet restrictors)
3 c Nozzle needle
closes (bypass
4 6 open, function with
two inlet restrictors)
1 Servo valve
(control valve)
2 Outlet restrictor
æ UMK1985E
5
3 Control chamber
4 Inlet restrictor
Rail pressure Leakage-oil pressure Control-chamber pressure 5 Nozzle needle
6 Bypass
Function of the hydraulic coupler To generate an injection event, a voltage

Another key component in the piezo-inline (110...150 V) is applied to the actuator
injector is the hydraulic coupler (Fig. 8, 3) until the equilibrium of forces between the
that implements the following functions: switching valve and the actuator is exceeded.
쐌 Translates and amplifies the actuator This increases the pressure in the coupler, and
stroke. a small leakage volume flows out of the cou-
쐌 Compensates for any play between the pler via the piston guide clearances into the
actuator and the servo valve (e.g. caused by low-pressure circuit of the injector. The pres-
thermal expansion). sure drop caused in the coupler has no im-
쐌 Performs a failsafe function (automatic pact on injector function for a triggering
safety cutoff of fuel injection if electrical period lasting several milliseconds.
decontacting fails). At the end of the injection process, the
quantity missing in the hydraulic coupler
The actuator module and the hydraulic coupler needs refilling. This takes place in the reverse
are immersed in the diesel fuel flow at a pres- direction via the guide clearances of the
sure of about 10 bar. When the actuator is not plungers as a result of the pressure difference
triggered, pressure in the hydraulic coupler is in between the hydraulic coupler and the low-
equilibrium with its surroundings. Changes in pressure circuit of the injector. The guide
length caused by temperature are compensated clearances and the low-pressure level are
by small leakage-fuel quantities via the guide matched to fill up the hydraulic coupler fully
clearances of the two plungers. This maintains before the next injection cycle starts.
the coupling of forces between actuator and
switching valve at all times.
8 Function of the hydraulic coupler

Voltage
1
2
Coupler pressure
pSystem
Recharging
pK < pSystem
Volume change
in coupler
Fig. 8 Rail pressure Leakage

pK > pSystem
1 Low-pressure fuel
æ UMK1986E
Coupler pressure
rail with valve 10 bar
2 Actuator
1 bar Time t
3 Hydraulic coupler
(translator)
Triggering the common-rail piezo Benefits of the piezo-inline injector

in-line injector 쐌 Multiple injection with flexible start of
The injector is triggered by an engine control injection and intervals between individual
unit whose output stage was specially designed injection events.
for these injectors. A reference triggering volt- 쐌 Production of very small injected fuel
age is predetermined as a function of the rail quantities for pre-injection.
pressure of the set operating point. The voltage 쐌 Small size and low weight of injector
signal is pulsed (Fig. 9) until there is a mini- (270 g compared to 490 g).
mum deviation between the reference and the 쐌 Low noise (–3 dB [A]).
control voltage. The voltage rise is converted 쐌 Lower fuel consumption (–3%).
proportionally into a piezoelectric actuator 쐌 Lower exhaust-gas emission (–20%).
stroke. The actuator stroke produces a pressure 쐌 Increased engine performance (+7%).
rise in the coupler by means of hydraulic
translation until the equilibrium of forces is
exceeded at the switching valve, and the valve
opens. As soon as the
switching valve reaches 9 Triggering sequences of the piezo-inline injector for an injection event
its end position, pres-
sure in the control
chamber starts to
a
drop via the needle, Voltage
and injection ends.
Current
Valve lift
Coupler pressure
c
Injection rate
Needle lift
Fig. 9
a Current and voltage
curves for triggering
0.5 the injector
æ UMK1987E
0.0 0.5 1.0 1.5 ms

Time t b Valve-lift curve and
coupler pressure
c Valve-lift curve and
injection rate
442 The piezoelectric effect
왘 The piezoelectric effect
In 1880 Pierre Curie and his brother Jacques The change in length ∆x results from the
discovered a phenomenon that is still very lit- following when a voltage U is applied:
tle known today, but is present in the everyday U / δ = ∆x (using quartz as an example:
lives of millions of people: the piezoelectric deformation of about 10–9 cm at U = 10 V)
effect. For example, it keeps the pointers of
a crystal clock operating in time. The piezoelectric effect is not only used in
quartz clocks and piezo-inline injectors, it has
Certain crystals (e.g. quartz and turmaline) are many other industrial applications, either as a
piezoelectric: Electric charges are induced on direct or inverse effect:
the crystal surface by exerting a compression Piezoelectric sensors are used for knock
or elongation force along certain crystal axes. control in gasoline engines. For example, they
This electrical polarization arises by shifting detect high-frequency engine vibrations as a
positive and negative ions in the crystal rela- feature of combustion knock. Converting me-
tive to each other by exerting force (see Fig., chanical vibration to electric voltage is also
b). The shifted centers of charge gravity within used in the crystal audio pickup of a record
the crystal compensate automatically, but an player or crystal microphones. The piezoelec-
electric field forms between the end faces of tric igniter (e.g. in a firelighter) causes me-
the crystal. Compressing and elongating the chanical pressure to produce the voltage to
crystal create inverse field directions. generate a spark.
On the other hand, if an electrical voltage On the other hand, if an alternating voltage
is applied to the end faces of the crystal, the is applied to a piezoelectrical crystal, it vi-
effect reverses (inverse piezoelectric effect): brates mechanically at the same frequency
The positive ions in the electric field migrate as the alternating voltage. Oscillating crystals
toward the negative electrode, and negative are used as stabilizers in electrical oscillating
ions toward the positive electrode. The crystal circuits or as piezoelectric acoustic sources
Principle of the then contracts or expands depending on the to generate ultrasound.
piezoelectric effect direction of the electric field strength (see When used in clocks, the oscillating quartz
(represented as a Fig., c). is excited by an alternating voltage whose
unit cell) frequency is the same as the quartz’s natural
The following applies to piezoelectric field frequency. This is how an extremely time-con-
a Quartz crystal SiO2
strength Ep: stant resonant frequency is generated. In a
Ep = δ ∆x/x calibrated quartz, it deviates by only approx.
b Piezoelectric effect:
∆x/x: relative compression or elongation 1/1,000 second per year.
When the crystal
is compressed, δ: piezoelectric coefficient, numeric value
negative O2– ions 109 V/cm through 1011 V/cm
shift upward,
positive Si4+ ions
shift downward:
a b c +
Electric charges
Si4+
are induced at the
+
crystal surface. ¯ ¯ + ¯ ¯ +
O 2¯
¯ ¯ ¯ ¯ ¯ ¯
c Inverse piezoelectric
effect:
By applying an + + + + + +
æ SAN0170Y
electrical voltage, + + ¯ + + ¯
O2– ions shift ¯
upward, Si4+ ions ¯
shift downward:
The crystal contracts.
Where does the word “electronics” come from? 443
왘 Where does the word “electronics” come from?
This term actually goes back to the ancient Even the first “Electronic Engineer” already
Greeks. For them, the word “electron” meant existed in the 19th century. Fleming was listed
amber. Its force of attraction on woollen in the 1888 edition of “Who’s Who”, pub-
threads or similar was known to Thales von lished during the reign of Queen Victoria. The
Milet over 2,500 years ago. official title was “Kelly’s Handbook of Titled,
Landed and Official Classes”. The Electronic
Electrons, and therefore electronics as such, Engineer can be found under the title “Royal
are extremely fast due to their very small mass Warrant Holders”, that is the list of persons
and electric charge. The term “electronics” who had been awarded a Royal Warrant.
comes directly from the word “electron”. What was this Electronic Engineer’s job?
The mass of an electron has as little effect He was responsible for the correct functioning
on a gram of any given substance as a 5 gram and cleanliness of the gas lamps at court.
weight has on the total mass of our earth. And why did he have such a splendid title?
Because he knew that “electrons” in ancient
The word “electronics” was born in the Greece stood for glitter, shine, and sparkle.
20th century. There is no evidence available
as to when the word was used for the first Source:
time. It could be Sir John Ambrose Fleming, “Basic Electronic Terms” (“Grundbegriffe der Elek-
one of the inventors of the electron tube in tronik”) – Bosch publication (reprint from the “Bosch
about 1902. Zünder” (Bosch Company Newspaper)), 1988.
æ LAE0047Y
This page left been blank intentionally.

444 High-pressure components of common-rail system High-pressure pumps
High-pressure pumps The pump plunger inside of the high-pres-

sure pump compresses the fuel. At three
Design and requirements delivery strokes per revolution, the radial-
The high-pressure pump is the interface be- piston pump produces overlapping delivery
tween the low-pressure and high-pressure strokes (no interruption in delivery), low
stages. Its function is to make sure there is al- drive peak torques, and an even load on the
ways sufficient fuel under pressure available pump drive.
in all engine operating conditions. At the On passenger-car systems, torque reaches
same time it must operate for the entire 16 Nm, i.e. only 1/9th of the drive torque
service life of the vehicle. This includes required for a comparable distributor injec-
providing a fuel reserve that is required for tion pump. As a result, the common-rail
quick engine starting and rapid pressure rise system places fewer demands on the pump-
in the fuel rail. drive system than conventional fuel-injection
The high-pressure pump generates a con- systems. The power required to drive the
stant system pressure for the high-pressure pump increases in proportion to the pressure
accumulator (fuel rail) independent of fuel in the fuel rail and the rotational speed of the
injection. For this reason, fuel – compared to pump (delivery quantity).
conventional fuel-injection systems – is not On a 2-liter engine, the high-pressure pump
compressed during the injection process. draws a power of 3.8 kW at nominal speed
and a pressure of 1,350 bar in the fuel rail
A 3-plunger radial-piston pump is used as the (at a mechanical efficiency of approx. 90%).
high-pressure pump to generate pressure in The higher power requirements of common-
passenger-car systems. 2-plunger in-line fuel- rail systems compared to conventional fuel-
injection pumps are also used on commercial injection systems is caused by leakage and
vehicles. Preferably, the high-pressure pump control volumes in the injector, and – on the
is fitted to the diesel engine at the same point high-pressure pump CP1 – the pressure drop
as a conventional distributor injection pump. to the required system pressure across the
The pump is driven by the engine via cou- pressure-control valve.
pling, gearwheel, chain, or toothed belt.
Pump speed is therefore coupled to engine The high-pressure radial-piston pumps used
speed via a fixed gear ratio. in passenger cars are lubricated by fuel. Com-
mercial-vehicle systems may have fuel- or
oil-lubricated radial-piston pumps, as well as
oil-lubricated 2-plunger in-line fuel-injection
1 Bosch high-pressure pumps for common-rail systems
pumps. Oil-lubricated pumps are more ro-
Pressure
Pump
in bar
Lubrication bust against poor fuel quality.
CP1 1,350 Fuel
CP1+ 1,350 Fuel High-pressure pumps are used in a number
CP1H 1,600 Fuel of different designs in passenger cars and
CP1H-OHW 1,100 Fuel commercial vehicles. There are versions of
CP3.2 1,600 Fuel pump generations that have different delivery
CP3.2+ 1,600 Fuel rates and delivery pressures (Table 1).
CP3.3 1,600 Fuel
CP3.4 1,600 Oil
CP3.4+ 1,600 Fuel
Table 1 CP2 1,400 Oil
H Increased pressure CPN2.2 1,600 Oil
section CPN2.2+ 1,600 Oil
+ Higher delivery rate CPN2.4 1,600 Oil
OHW Off-Highway
High-pressure components of common-rail system High-pressure pumps 445
Radial-piston pump (CP1) behind the inlet. If the delivery pressure of

Design the presupply pump exceeds the opening
The drive shaft in the housing of CP1 is pressure (0.5 to 1.5 bar) of the safety valve,
mounted in a central bearing (Fig. 1, 1). the fuel is pressed through the restriction
The pump elements (3) are arranged radially bore of the safety valve into the lubrication
with respect to the central bearing and offset and cooling circuit of the high-pressure
by 120°. The eccenter (2) fitted to the drive pump. The drive shaft with its eccenter moves
shaft forces the pump plunger to move up the pump plunger up and down to mimic the
and down. eccentric lift. Fuel passed through the high-
pressure pump’s inlet valve (4) into the ele-
Force is transmitted between the eccentric ment chamber and the pump plunger moves
shaft and the delivery plunger by means of downward (inlet stroke).
a drive roller, a sliding ring mounted on the When the bottom-dead center of the pump
shaft eccenter, and a plunger base plate at- plunger is exceeded, the inlet valve closes,
tached to the plunger base plate. and the fuel in the element chamber can no
longer escape. It can then be pressurized be-
Operating concept yond the delivery pressure of the presupply
Fuel delivery and compression pump. The rising pressure opens the outlet
The presupply pump – an electric fuel pump valve (5) as soon as pressure reaches the level
or a mechanically driven gear pump – deliv- in the fuel rail. The pressurized fuel then
ers fuel via a filter and water separator to the passes to the high-pressure circuit.
inlet of the high-pressure pump (6). The inlet
is located inside of the pump on passenger-
car systems with a gear pump flanged to the
high-pressure pump. A safety valve is fitted
1 High-pressure pump (schematic, cross-section)
4 5
Fig. 1
1 Drive shaft
2 Eccenter
æ UMK1573-1Y
3 Pump element with

pump plunger
4 Inlet valve
5 Outlet valve
6 Fuel inlet
The pump plunger continues to deliver fuel Delivery rate

until it reaches its top-dead center position As the high-pressure pump is designed for
(delivery stroke). The pressure then drops so high delivery quantities, there is a surplus
that the outlet valve closes. The remaining of pressurized fuel when the engine is idling
fuel is depressurized and the pump plunger or running in part-load range. On first-
moves downward. generation systems with a CP1, excess fuel
When the pressure in the element chamber delivered is returned to the fuel tank by the
exceeds the pre-delivery pressure, the inlet pressure-control valve on the fuel rail. As
valve reopens, and the process starts over. the compressed fuel expands, the energy
imparted by compression is lost; overall
Transmission ratio efficiency drops. Compressing and then
The delivery quantity of a high-pressure expanding the fuel also heats the fuel.
pump is proportional to its rotational speed.
In turn, the pump speed is dependent on
the engine speed. The transmission ratio
between the engine and the pump is deter-
mined in the process of adapting the fuel-
injection system to the engine so as to limit
the volume of excess fuel delivered. At the
same time it makes sure that the engine’s fuel
demand at WOT is covered to the full extent.
Possible gear ratios are 1:2 or 2:3 relative to
the crankshaft.
2 High-pressure pump (CP1), variant with mounted pressure-control valve (3D view)
1 2 3
Fig. 2
11 Flange
12 Pump housing
13 Engine cylinder head
14 Inlet connection
15 High-pressure inlet
16 Return connection
æ UMK2022Y
17 Pressure-control
valve
18 Barrel bolt
19 Shaft seal 10 9 8
10 Eccentric shaft
Radial-piston pump (CP1H) The solenoid valve is triggered by a PWM

Modifications signal.
An improvement in energetic efficiency is
possible by controlling fuel delivery by the
high-pressure pump on the fuel-delivery side
(suction side). Fuel flowing into the pump 4 Metering unit design
element is metered by an infinitely variable
solenoid valve (metering unit, ZME). This
valve adapts the fuel quantity delivered to the
rail to system demand. This fuel-delivery con-
trol not only drops the performance demand 1
of the high-pressure pump, it also reduces the
maximum fuel temperature. This system de-
signed for the CP1H was taken over by the 2 Fig. 4
CP3. 3 11 Plug with electrical
Compared to the high-pressure pump CP1, interface
the CP1H is designed for higher pressures up 4
12 Magnet housing
to 1,600 bar. This was achieved by reinforcing 5
13 Bearing
14 Armature with
the drive mechanism, modifying the valve 6
tappet
units, and introducing measures to increase
15 Winding with coil
the strength of the pump housing. body
7
The metering unit is mounted on the high- 8 16 Cup
pressure pump (Fig. 3, 13). 9 17 Residual air-gap
washer
10
Design of the metering unit (ZME) 18 Magnetic core
9 19 O-ring
Fig. 4 shows the design of the metering unit.
æ UMK2016Y
11
10 Plunger with control
The plunger operated by solenoid force frees 12 slots
up a metering orifice depending on its posi- 11 Spring
9
tion. 12 Safety element
3 High-pressure pump (CP1H) with metering unit (exploded view)
8 Fig. 3
9
11 Flange
10
7 12 Eccentric shaft
13 Bushing
11
14 Drive roller
12
6 15 Pump housing
16 Plate
13 17 Spring
18 Engine cylinder
head
19 Return-flow
14 connection
15 10 Overflow valve
11 Inlet connection
æ UMK2015Y
12 Filter
13 Metering unit
1 2 3 4 5 14 Cage
15 Pump plunger
Radial-piston pump (CP3) Variants

Modifications Pumps of the CP3 family are used in both
The CP3 is a high-pressure pump with suc- passenger cars and commercial vehicles.
tion-side fuel-delivery control by means of a A number of different variants are used
metering unit (ZME). This control was first depending on the delivery rate required. The
used on the CP3 and was assumed later on size, and thus the delivery rate, increases from
the CP1H. the CP3.2 to the CP3.4. The oil-lubricated
The principle design of the CP3 (Fig. 5) is CP3.4 is only used on heavy-duty trucks. On
similar to the CP1 and the CP1H. The main light-duty trucks and vans, pumps primarily
difference in features are: designed for passenger cars may also be used.
쐌 Monobloc housing: This construction A special feature of systems for medium-
reduces the number of leak points in the duty and heavy-duty trucks is the fuel filter
high-pressure section, and permits a higher located on the pressure side. It is situated
delivery rate. between the gear pump and the high-pressure
쐌 Bucket tappets: Transverse forces arising pump, and permits a greater filter storage
from the transverse movement of the capacity before requiring a change. The high-
eccenter drive roller are not removed di- pressure pump requires an external connec-
rectly by the pump plungers but by buckets tion for the fuel inlet in any case, even if the
on the housing wall. The pump then has gear pump is flanged onto the high-pressure
greater stability under load and is capable pump.
of withstanding higher pressures. Poten-
tially, it can withstand pressures up to
1,800 bar.
5 High-pressure pump CP3 with metering unit and mounted gear presupply pump
æ UMK2014Y
In-line piston pump (CP2) Lube oil is supplied either directly via the
Design mounting flange of the CP2 or a side-
The oil-lubricated, quantity-controlled high- mounted inlet.
pressure pump (CP2) is only used on com-
mercial vehicles. This is a 2-plunger pump The drive gear ratio is 1:2. The CP2 is there-
with an in-line design, i.e. the two pump fore mountable together with conventional
plungers are arranged adjacently (Fig. 6). in-line fuel-injection pumps.
A gear pump with a high gear ratio is lo-
cated on the camshaft extension. Its function Operating concept
is to draw fuel from the fuel tank and route it Fuel enters the pump element and the com-
to the fine filter. From there, the fuel passes pressed fuel is conveyed to the fuel rail via a
through another line to the metering unit lo- combined inlet/outlet valve on the CP2.
cated on the upper section of the high-pres-
sure pump. The metering unit controls the
fuel quantity delivered for compression de-
pendent on actual demand in the same way as
other common-rail high-pressure pumps of
the recent generation.
6 High-pressure pump CP2
1
7
2
8 Fig. 6
11 Zero delivery
9 restrictor
12 Metering unit
3 13 Internal gear
14 Pinion
10 15 Gear presupply
4 pump
11 16 High-pressure
connection
17 Two-part inlet/outlet
valve
5 18 C-coated plunger
19 Plunger return
æ UMK2013Y
12 spring
10 Oil inlet
11 C-coated roller bolt
12 Concave cam
4 High-pressure components of common-rail system Fuel rail (high-pressure accumulator)
Fuel rail (high-pressure Operating concept

The pressurized fuel delivered by the high-
accumulator) pressure pump passes via a high-pressure fuel
Function line to the fuel-rail inlet (4). From there, it is
The function of the high-pressure accumula- distributed to the individual injectors (hence
tor (fuel rail) is to maintain the fuel at high the term “common rail”).
pressure. In so doing, the accumulator vol-
ume has to dampen pressure fluctuations The fuel pressure is measured by the rail-
caused by fuel pulses delivered by the pump pressure sensor and controlled to the required
and the fuel-injection cycles. This ensures value by the pressure-control valve. The pres-
that, when the injector opens, the injection sure-relief valve is used as an alternative to
pressure remains constant. On the one hand, the pressure-control valve – depending on
the accumulator volume must be large system requirements – and its function is to
enough to meet this requirement. On the limit fuel pressure in the fuel rail to the maxi-
other hand, it must be small enough to en- mum permissible pressure. The highly com-
sure a fast enough pressure rise on engine pressed fuel is routed from the fuel rail to the
start. Simulation calculations are conducted injectors via high-pressure delivery lines.
during the design phase to optimize the per-
formance features. The cavity inside the fuel rail is permanently
Besides acting as a fuel accumulator, the filled with pressurized fuel. The compressibil-
fuel rail also distributes fuel to the injectors. ity of the fuel under high pressure is utilized
to achieve an accumulator effect. When fuel
Design is released from the fuel rail for injection, the
The tube-shaped fuel rail (Fig. 1, 1) can have pressure in the high-pressure accumulator
as many designs as there are engine mounting remains virtually constant, even when large
variants. It has mountings for the rail-pres- quantities of fuel are released.
sure sensor (5) and a pressure-relief valve or
pressure-control valve (2).
1 Common rail with attached components
4
5
1
Fig. 1
1 Fuel rail
2 Pressure-control
2
valve
3 Return line from
fuel rail to fuel tank
æ SMK1996Y
3 6
4 Inlet from high-
pressure pump
6 Fuel line to injector
High-pressure components of common-rail system High-pressure sensors 451
High-pressure sensors pressures and thinner ones for lower pres-

sures). When the pressure is applied via the
Application pressure connection (4) to one of the dia-
In automotive applications, high-pressure phragm faces, the resistances of the bridge re-
sensors are used for measuring the pressures sistors change due to diaphragm deformation
of fuels and brake fluids. (approx. 20 μm at 1,500 bar).
The 0...80 mV output voltage generated by
Diesel rail-pressure sensor the bridge is conducted to an evaluation cir-
In the diesel engine, the rail-pressure sensor cuit which amplifies it to 0...5 V. This is used
measures the pressure in the fuel rail of the as the input to the ECU which refers to a
common-rail accumulator-type injection sys- stored characteristic curve in calculating the
tem. Maximum operating (nominal) pressure pressure (Fig. 2).
pmax is 160 MPa (1,600 bar). Fuel pressure is
controlled by a closed control loop, and re-
mains practically constant independent of
load and engine speed. Any deviations from 1 High-pressure sensor
the setpoint pressure are compensated for by
a pressure-control valve. 2 cm
1
Gasoline rail-pressure sensor
As its name implies, this sensor measures the
pressure in the fuel rail of the DI Motronic
2
with gasoline direct injection. Pressure is a
function of load and engine speed and is 3
5...12 MPa (50...120 bar), and is used as an
actual (measured) value in the closed-loop Fig. 1
rail-pressure control. The rpm and load- 4 1 Electrical
dependent setpoint value is stored in a map connection (socket)
5 2 Evaluation circuit
and is adjusted at the rail by a pressure
3 Steel diaphragm
æ UMK1576Y
control valve.
with deformation
resistors
Brake-fluid pressure sensor p 4 Pressure connection
Installed in the hydraulic modulator of 5 Mounting thread
such driving-safety systems as ESP, this
high-pressure sensor is used to measure 2 High-pressure sensor (curve, example)
the brake-fluid pressure which is usually
25 MPa (250 bar). Maximum pressure pmax V
can climb to as much as 35 MPa (350 bar).
Pressure measurement and monitoring is 4.5
triggered by the ECU which also evaluates
Output voltage
the return signals.
Design and operating concept

The heart of the sensor is a steel diaphragm
onto which deformation resistors have been
æ UAE0719-2E
vapor-deposited in the form of a bridge 0.5

circuit (Fig. 1, 3). The sensor’s pressure-
measuring range depends on diaphragm 0 pmax
Pressure
thickness (thicker diaphragms for higher
452 High-pressure components of common-rail system Pressure-control valve
Pressure-control valve Pressure-control valve activated

When the pressure in the high-pressure cir-
Function cuit needs to be increased, the force of the
The function of the pressure-control valve electromagnet is added to that of the spring.
is to adjust and maintain the pressure in the The pressure-control valve is activated and
fuel rail as a factor of engine load, i.e.: closes until a state of equilibrium is reached
쐌 It opens when the rail pressure is too high. between the high pressure and the combined
Part of the fuel then returns from the fuel force of the electromagnet and the spring.
rail via a common line to the fuel tank. At this point, it remains in partly open posi-
쐌 It closes when the rail pressure is too low, tion and maintains a constant pressure.
thus sealing the high-pressure side from the Variations in the delivery quantity of the
low-pressure side. high-pressure pump and the withdrawal
of fuel from the fuel rail by the injectors are
Design compensated by varying the valve aperture.
The pressure-control valve (Fig. 1) has a The magnetic force of the electromagnet
mounting flange which attaches it to the is proportional to the control current. The
high-pressure pump or the fuel rail. The control current is varied by pulse-width
armature (5) forces a valve ball (6) against the modulation. A pulse frequency of 1 kHz is
valve seat in order to seal the high-pressure sufficiently high to prevent adverse armature
stage from the low-pressure stage; this is movement or pressure fluctuations in the
achieved by the combined action of a valve fuel rail.
spring (2) and an electromagnet (4) which
force the armature downwards. Designs
Fuel flows around the whole of the arma- The pressure-control valve DRV1 is used in
ture for lubrication and cooling purposes. first-generation common-rail systems. Sec-
ond- and third-generation CR systems oper-
Operating concept ate using the two-actuator concept. Here, the
The pressure-control valve has two closed rail pressure is adjusted by both a metering
control loops: unit as well as a pressure-control valve. In this
쐌 A slower, closed electrical control loop for
setting a variable average pressure level in 1 Pressure-control valve DRV1 (section)
the fuel rail.
쐌 A faster hydromechanical control loop for
balancing out high-frequency pressure
pulses.
1
Fig. 1
11 Electrical Pressure-control valve not activated
connections 2
The high pressure present in the fuel rail or
12 Valve spring 3
at the high-pressure pump outlet is applied
13 Armature
14 Valve housing
to the pressure-control valve via the high-
4
15 Solenoid coil pressure fuel supply. As the deengerized
5
16 Valve ball electromagnet exerts no force, the high-
17 Support ring pressure force is greater than the spring force.
18 O-ring The pressure-control valve opens to a greater
19 Filter 12
or lesser extent depending on the delivery 6
10 High-pressure 11
æ UMK2007Y
quantity. The spring is dimensioned to 7

fuel supply
11 Valve body
maintain a pressure of approx. 100 bar. 10 8
12 Drain to low- 9
pressure circuit
High-pressure components of common-rail system Pressure-control valve, pressure-relief valve 453
case, either the pressure-control valve DRV2 Pressure-relief valve

is used or the DRV3 variant for higher pres-
sures. This control strategy achieves lower Function
fuel heating and eliminates the need for a fuel The pressure-relief valve has the same func-
cooler. tion as a pressure limiter. The latest version
of the internal pressure-relief valve now has
The DRV2/3 (Fig. 2) differs from the DRV1 an integrated limp-home function. The pres-
in the following features: sure-relief valve limits pressure in the fuel
쐌 Hard seal to the high-pressure interface rail by releasing a drain hole when pressure
(bite edge). exceeds a certain limit. The limp-home func-
쐌 Optimized magnetic circuit (lower power tion ensures that a certain pressure is main-
consumption). tained in the fuel rail to permit the vehicle to
쐌 Flexible mounting concept (free plug continue running without any restriction.
orientation).
Design and operating concept
The pressure-relief valve (Fig. 3) is a mechan-
ical component. It consists of the following
2 Pressure-control valve DRV2 parts:
쐌 A housing with an external thread for
screwing to the fuel rail.
쐌 A connection to the fuel-return line
to the fuel tank (3).
1 2 3 4 5 6 7 8
쐌 A movable plunger (2). Fig. 2
쐌 A plunger return spring (5). 1 Filter
2 Bite edge
9 At the end which is screwed to the fuel rail, 3 Valve ball
there is a hole in the valve housing which 4 O-ring
is sealed by the tapered end of the plunger 5 Union bolt with
circlip
resting against the valve seat inside the valve
6 Armature
æ UMK2005Y
housing. At normal operating pressure, a 7 Solenoid coil

spring presses the plunger against the valve 8 Electrical
seat so that the fuel rail remains sealed. connection
Only if the pressure rises above the maximum 9 Valve spring
system pressure is the plunger forced back
3 Pressure-relief valve DBV4 against the action of the spring by the pres-
sure in the fuel rail so that the high-pressure
fuel can escape. The fuel is routed through
passages into a central bore of the plunger
and returned to the fuel tank via a common
6
line. As the valve opens, fuel can escape from
the fuel rail to produce a reduction in fuel-
rail pressure.
Fig. 3
5
1 Valve insert
2 Valve plunger
3 Low-pressure
æ UMK2003Y
section
4 Valve holder
1 2 3 4 5 Spring
6 Diaphragm disc
454 Electronic diesel control System overview
Electronic Diesel Control (EDC)

Electronic control of a diesel engine allows In addition, diesel engine development has
fuel-injection parameters to be varied pre- been influenced by the high levels of com-
cisely for different conditions. This is the fort and convenience demanded in modern
only means by which a modern diesel engine cars. Noise levels, too, are subject to more
is able to satisfy the many demands placed and more stringent requirements.
upon it. The EDC (Electronic Diesel Con- As a result, the performance demanded
trol) system is subdivided into three areas, of fuel-injection and engine-management
“Sensors and desired-value generators”, systems has also increased, specifically with
“Control unit”, and “Actuators”. regard to:
쐌 High fuel-injection pressures
System overview 쐌 Rate-of-discharge curve control
쐌 Pre-injection and, where applicable,
Requirements secondary injection
Present-day development in the field of 쐌 Variation of injected fuel quantity, charge-
diesel technology is focused on lowering air pressure, and start of injection to suit
fuel consumption and exhaust-gas emissions operating conditions
(NOx, CO, HC, particulate), while increasing 쐌 Temperature-dependent excess-fuel quan-
engine performance and torque. In recent tity for starting
years this has led to an increase in the popu- 쐌 Control of idle speed independent of
larity of the direct-injection (DI) diesel en- engine load
gine, which uses much higher fuel-injection 쐌 Controlled exhaust-gas recirculation
pressures than indirect-injection (IDI) en- (cars)
gines with whirl or prechamber systems. 쐌 Cruise control
Due to the more efficient mixture formation 쐌 Tight tolerances for injection duration
and the absence of flow-related losses be- and injected fuel quantity, and mainte-
tween the whirl chamber/prechamber and nance of high precision over the service
the main combustion chamber, the fuel life of the system (long-term perfor-
consumption of direct-injection engines mance)
is 10 ... 20% lower than that achieved by
indirect-injection designs.
1 Main components of EDC
Sensors and setpoint generators ECU Actuators
Pedal-travel sensor
Injectors
Air-mass sensor ADC
Rail-pressure sensor
Boost-pressure sensor Function
Temperature sensors processor
(air and coolant) Intake-duct switchoff
Lambda oxygen Boost-pressure actuator
sensor Exhaust-gas recirculation
Wheel-speed sensors RAM actuator
(crankshaft, Throttle-valve actuator
camshaft) Flash
EPROM A/C compressor
Brake switch
Clutch switch EEPROM Auxiliary heating
Ignition switch Radiator fan
Mon- Rail-pressure control valve
æ UMK1988E
Glow-plug control
unit itoring Electronic shutoff valve
CAN module (EAB)
Fault diagnosis Diagnosis lamp

Electronic diesel control System overview 455
Conventional mechanical governing of engine home mode in the idle-speed range). EDC,
speed uses a number of adjusting mechanisms therefore incorporates a number of control
to adapt to different engine operating condi- loops.
tions and ensures a high mixture formation
quality. Nevertheless, it is restricted to a sim- Electronic diesel control allows data exchange
ple engine-based control loop and there are a with other electronic systems, such as the
number of important influencing variables Traction Control System (TCS), Electronic
that it cannot take account of or cannot re- Transmission Control (ETC), or Electronic
spond quickly enough to. Stability Program (ESP). As a result, the en-
As demands have increased, what was orig- gine management system can be integrated in
inally a straightforward system using electric the vehicle’s overall control system network,
actuator shafts has developed into the pre- thereby enabling functions such as reduction
sent-day EDC, a complex electronic control of engine torque when the automatic trans-
system capable of processing large amounts mission changes gear, regulation of engine
of data in real time. It can form part of an torque to compensate for wheel spin, dis-
overall electronic vehicle control system abling of fuel injection by the engine immo-
(“drive-by-wire”). And a result of the increas- bilizer, etc.
ing integration of electronic components, the The EDC system is fully integrated in the
control-system circuitry can be accommo- vehicle’s diagnostic system. It meets all OBD
dated in a very small space. (On-Board Diagnosis) and EOBD (European
OBD) requirements.
Operating concept
Electronic Diesel Control (EDC) is capable of System modules
meeting the requirements listed above as a re- Electronic Diesel Control (EDC) is divided
sult of microcontroller performance that has into three system modules (Fig. 1):
risen considerably in the last few years.
In contrast with diesel-engine vehicles with 1. Sensors and setpoint generators detect
conventional mechanically controlled fuel- operating conditions (e.g. engine speed)
injection pumps, the driver of a vehicle and setpoint values (e.g. switch position).
equipped with EDC has no direct control They convert physical variables into electrical
over the injected fuel quantity via the acceler- signals.
ator pedal and cable. The injected fuel quan-
tity is actually determined by a number of 2. The electronic control unit processes data
different influencing variables. They include: from the sensors and setpoint generators
쐌 The vehicle response desired by the driver based on specific open- and closed-loop con-
(accelerator-pedal position) trol algorithms. It controls the actuators by
쐌 The engine operating status means of electrical output signals. In addi-
쐌 The engine temperature tion, the control unit acts as an interface to
쐌 Interventions by other systems (e.g. TCS) other systems and to the vehicle diagnostic
쐌 The effect on exhaust-gas emission levels, system.
etc.
3. Actuators convert electrical output signals
The control unit calculates the injected fuel from the control unit into mechanical para-
quantity on the basis of all these influencing meters (e.g. the solenoid valve for the fuel-
variables. Start of delivery can also be varied. injection system).
This demands a comprehensive monitoring
concept that detects inconsistencies and initi-
ates appropriate actions in accordance with
the effects (e.g. torque limitation or limp-
456 Electronic diesel control Common-rail system for passenger cars
Common-rail system for passenger cars

1 Overview of EDC components for common-rail systems in passenger cars
ECU Injectors
Pedal-travel sensor EDC 16 C / EDC 7 (max. 8 per ECU)
with low-idle switch
and kickdown switch Signal inputs
Sensor evaluation * *
Engine speed Atmospheric-pressure
(crankshaft) sensor
Functions:
Engine speed and - Idle-speed control
cylinder detection - Intermediate-speed
(camshaft)
regulation
- Smooth-running
Rail pressure control
Rail-pressure control valve
- Active-surge damper High-pressure pump
- External torque
Charge-air pressure intervention Exhaust-gas recirculation
- Vehicle immobilizer actuator
- Fuel-delivery control
and limitation Boost-pressure actuator*
Charge-air temperature
- Vehicle-speed
controller (FGR)* Throttle-valve actuator
Engine temperature - Vehicle-speed
(coolant) limitation*
- Cylinder shutoff Auxiliary heating*
Exhaust-gas - Fuel-quantity control

temperature* - Rail-pressure control Intake-duct switchoff*
- Start-of-injection control
- Pre-injection control
Air mass or Radiator-fan control*
exhaust-gas - Secondary-injection
signal check-back control*
- Additional special Starter motor*
adaptations*
Road speed
Additional output stages*
Diagnostic functions:
- System diagnostics
Terminal 15 - Substitute functions Actuators
- Engine diagnosis
Terminal 50* Solenoid-valve output

stages K ISO interface
Power output stages (e.g. diagnostics)
Clutch switch (altern- Signal outputs L
ative P/N contact on
automatic transmission) CAN communcation
Diagnosis communication Glow-plug control unit
End-of-line programming
Multistage switch Engine-speed signal
for maximum-speed Power supply (TD)
limitation (HGB) and Diagnosis lamp
vehicle-speed controller
+ 12V
(FGR) Main relay CAN CAN interface
æ SAE0750-1E
Sensors and setpoint generators Communication

*optional
Electronic diesel control Common-rail system for commercial vehicles 457
Common-rail system for commercial vehicles

2 Overview of EDC components for common-rail systems in commercial vehicles
ECU Injectors
Pedal-travel sensor EDC 16 / EDC 7 (max. 8 per ECU)
with low-idle switch
and kickdown switch Signal inputs * *
Engine speed Sensor evaluation
(crankshaft)
Engine speed and Functions:

cylinder detection - Idle-speed control
(camshaft) - Intermediate-speed
regulation
Rail pressure
- Smooth-running control Metering unit
- Active-surge damper High-pressure pump
Charge-air pressure Exhaust-gas recirculation
- External torque
intervention actuator
Oil pressure - Vehicle immobilizer Boost-pressure actuator*/
- Fuel-delivery control VTG*
Charge-air temperature
and limitation
Engine temperature Auxiliary heating*
- Vehicle-speed controller
(coolant) (FGR)*
- Vehicle-speed limitation Cold-start system
Fuel temperature
- Cylinder shutoff
Differential pressure Fuel heater*
(particulate filter) - Rail-pressure control
- Start-of-injection control
Exhaust-gas Intake-duct switchoff*
temperature* - Multiple-injection control
- Additional special
Air mass or adaptations* Engine brake valve*/
exhaust-gas retarder
signal check-back**
Radiator-fan control
Terminal 15 Diagnostic functions:
On-board diagnosis Radiator-fan coupling
Terminal 50* (OBD)
Clutch switch (altern- Substitute functions
ative P/N contact on Starter motor*
Engine diagnosis
automatic transmission)
Engine brake switch Additional output stages*
Solenoid-valve output
Road speed stages Actuators
Power output stages
Multistage switch
for maximum-speed Signal outputs K
limitation (HGB) and CAN communcation
vehicle-speed controller L ISO interface
(FGR) Diagnosis communication
(e.g. diagnostics)
End-of-line programming
Radiator-fan speed
Power supply
Glow-plug control unit**
VTG rpm*
Engine-speed signal (TD)
Brake switch 12V / 24V Diagnosis lamp
+
æ UAE0986E
Main relay CAN CAN interface

Sensors and setpoint generators
Communication
*optional **light-duty only
458 Electronic diesel control Data processing
Data processing The required degree of accuracy together

with the diesel engine’s outstanding dynamic
The main function of the electronic diesel response requires high-level computing
control (EDC) is to control the injected fuel power. The output signals trigger output
quantity and the injection timing. The com- stages that supply sufficient power for the
mon-rail fuel-injection system also controls actuators (e.g. high-pressure solenoid valves
injection pressure. Furthermore, on all sys- for the fuel-injection system, exhaust-gas re-
tems, the engine ECU controls a number circulation positioners, and boost-pressure
of actuators. For all components to operate actuators). Apart from this, a number of
efficiently, the EDC functions must be pre- other auxiliary-function components
cisely matched to every vehicle and every (e.g. glow relay and air-conditioning system)
engine. This is the only way to optimize com- are triggered.
ponent interaction (Fig. 2).
Faulty signal characteristics are detected by
The control unit evaluates the signals sent by output-stage diagnostic functions for the
the sensors and limits them to the permitted solenoid valves. Furthermore, signals are
voltage level. Some input signals are also exchanged with other systems in the vehicle
checked for plausibility. Using these input via the interfaces. The engine ECU monitors
data together with stored program maps, the the complete fuel-injection system as part of
microprocessor calculates injection timing a safety strategy.
and its duration. This information is then
converted to a signal characteristic which
is aligned to the engine’s piston strokes.
This calculation program is termed the “ECU
software”.
1 Schematic using the example of a current regulator
I–actual
PT1
I–setpoint
PT1 ¯
l
DT1
Min Max PWM–setpoint

x ÷ + x
x
æ SAE0987Y
Electronic diesel control Data processing 459
2 Electronic Diesel Control (EDC): Basic sequence
Fuel control circuit 1 (fuel-injection components) Air control circuit

Fuel control circuit 2 (engine) Data and signal flow
"Diversion" via driver
EDC ECU Data exchange with

other systems
¯ Traction control system
Triggering the fuel- ¯ Transmission control
injection components ¯ A/C control, etc.
CAN
Control and
Fuel-injection control triggering of the
remaining actuators
Sensors and
Driver commands setpoint generators
¯ Driver command ¯ Pedal-travel sensor
¯ Cruise control ¯ Wheel-speed sensors
¯ Engine brake, etc. ¯ Switch, etc.
Engine Cylinder-charge
control system
¯ Supercharging
¯ Exhaust-gas recirculation
Air
Actuators
¯ Electropneum. converter
¯ Continuous-operation
Fuel braking system
¯ Radiator fan
¯ Glow-plug control, etc.
Fuel-injection components
¯ In-line fuel-injection pumps
¯ Distributor injection pumps
¯ Unit Injector / Unit Pump
æ SMK1793-1E
¯ Common-rail high-pressure
pump and injectors
¯ Nozzle holders and nozzles
460 Electronic diesel control Fuel-injection control
Fuel-injection control In order that the engine can run with optimal
combustion under all operating conditions,
An overview of the various control functions the ECU calculates exactly the right injected
which are possible with the EDC control fuel quantity for all conditions. Here, a num-
units is given in Table 1. Fig. 1 opposite shows ber of parameters must be taken into ac-
the sequence of fuel-injection calculations count. On a number of solenoid-valve-con-
with all functions, a number of which are trolled distributor pumps, the sol-enoid
special options. These can be activated in the valves for injected fuel quantity and start of
ECU by the workshop when retrofit equip- injection are triggered by a separate pump
ment is installed. ECU (PSG).
1 EDC variants for road vehicles: Overview of functions
Fuel-injection system In-line injection Helix-controlled Solenoid-valve- Unit Injector Common Rail
pumps distributor controlled System and System
injection pumps distributor Unit Pump
injection pumps System
PE VE-EDC VE-M, VR-M UIS, UPS CR
Function
Injected-fuel-quantity limitation 쐌 쐌 쐌 쐌 쐌
External torque intervention 쐌3 쐌 쐌 쐌 쐌
Vehicle-speed limitation 쐌3 쐌 쐌 쐌 쐌
Vehicle-speed control
(Cruise Control) 쐌 쐌 쐌 쐌 쐌
Altitude compensation 쐌 쐌 쐌 쐌 쐌
Boost-pressure control 쐌 쐌 쐌 쐌 쐌
Idle-speed control 쐌 쐌 쐌 쐌 쐌
Intermediate-speed control 쐌3 쐌 쐌 쐌 쐌
Active surge damping 쐌2 쐌 쐌 쐌 쐌
BIP control – – 쐌 쐌 –
Intake-tract switch-off – – 쐌 쐌2 쐌
Electronic immobilizer 쐌2 쐌 쐌 쐌 쐌
Controlled pilot injection – – 쐌 쐌2 쐌
Glow control 쐌2 쐌 쐌 쐌2 쐌
A/C switch-off 쐌2 쐌 쐌 쐌 쐌
Auxiliary coolant heating 쐌2 쐌 쐌 – 쐌
Cylinder-balance control 쐌2 쐌 쐌 쐌 쐌
Control of injected fuel
quantity compensation 쐌2 – 쐌 쐌 쐌
Table 1 Fan (blower) triggering – 쐌 쐌 쐌 쐌

1 Only control-sleeve EGR control 쐌2 쐌 쐌 쐌2 쐌
in-line injection
Start-of-injection control
pumps
with sensor 쐌 1, 3 쐌 쐌 – –
2 Passenger cars only
3 Commercial vehicles Cylinder shutoff – – 쐌3 쐌3 쐌3
only
Electronic diesel control Fuel-injection control 461
1 Calculation of fuel-injection process in the ECU
Inputs
Vehicle-speed controller
Accelerator-pedal sensor
(Cruise Control),
(driver input)
vehicle-speed limiter
Inputs from
other systems
(e.g. ABS, TCS, ESP)
CAN
Calculations
Selection of the required
External torque intervention
injected fuel quantity
+/-
Idle-speed controller,
or controller for injected-fuel- Injected-fuel-quantity limit
quantity compensation
Smooth-running controller Active surge damper
Start Drive mode

Start quantity
Switch
Control of start of injection, Fuel-quantity metering

and/or start of delivery (pump map)
Triggering
Timing-device Solenoid-valve Pump ECU

triggering triggering triggering
æ NMK1755E
Start quantity When adjusting the stipulated idle speed, the

For starting, the injected fuel quantity is cal- idle-speed control must cope with heavily
culated as a function of coolant temperature fluctuating requirements. The input power
and cranking speed. Start-quantity signals needed by the engine-driven auxiliary equip-
are generated from the moment the starting ment varies considerably.
switch is turned (Fig. 1, switch in “Start” posi- At low electrical-system voltages, for in-
tion) until a given minimum engine speed is stance, the alternator consumes far more
reached. power than it does when the voltages are
The driver cannot influence the start quan- higher. In addition, the power demands from
tity. the A/C compressor, the steering pump, and
the high-pressure generation for the diesel in-
Drive mode jection system must all be taken into account.
When the vehicle is being driven normally, Added to these external load moments is
the injected fuel quantity is a function of the the engine’s internal friction torque which is
accelerator-pedal setting (accelerator-pedal highly dependent on engine temperature,
sensor) and of the engine speed (Fig. 1, and must also be compensated for by the
switch in “Drive” position). Calculation idle-speed control.
depends upon maps which also take other
influences into account (e. g. fuel and intake- In order to regulate the desired idle speed,
air temperature). This permits best-possible the controller continues to adapt the injected
alignment of the engine’s output to the fuel quantity until the actual engine speed
driver’s wishes. corresponds to the desired idle speed.
Idle-speed control Maximum-rpm control

The function of idle speed control (LLR) is to The maximum-rpm control ensures that
regulate a specific setpoint speed at idle when the engine does not run at excessive speeds.
the accelerator pedal is not operated. This can To avoid damage to the engine, the engine
vary depending on the engine’s particular op- manufacturer stipulates a permissible maxi-
erating mode. For instance, with the engine mum speed which may only be exceeded for
cold, the idle speed is usually set higher than a very brief period.
when it is hot. There are further instances
when the idle speed is held somewhat Above the rated-power operating point,
higher. For instance, when the vehicle’s electri- the maximum-speed governor reduces the
cal-system voltage is too low, when the air- injected fuel quantity continuously, until
conditioning system is switched on, or when just above the maximum-speed point when
the vehicle is freewheeling. When the vehicle fuel-injection stops completely. In order
is driven in stop-and-go traffic, together with to prevent engine surge, a ramp function
stops at traffic lights, the engine runs a lot of is used to ensure that the drop-off in fuel
the time at idle. Considerations concerning injection is not too abrupt. This is all the
emissions and fuel consumption dictate, more difficult to implement, the closer the
therefore, that idle speed should be kept as nominal performance point and maximum
low as possible. This, of course, is a disadvan- engine speed are to each other.
tage with respect to smooth-running and
pulling away.
Intermediate-speed control If cruise control has been switched off, the

Intermediate-speed control (ZDR) is used on driver only needs to shift the lever to the
commercial vehicles and light-duty trucks restore position to reselect the last speed
with power take-offs, e.g. crane), or for special setting.
vehicles (e.g. ambulances with a power gener-
ator). With the control in operation, the en- The operator controls can also be used for a
gine is regulated to a load-independent inter- step-by-step change of the selected speed.
mediate speed.
With the vehicle stationary, the intermedi- Vehicle-speed limiter
ate-speed control is activated via the cruise- Variable limitation
control operator unit. A fixed rotational Vehicle-speed limitation (FGB, also called
speed can be called up from the data store at the limiter) limits the maximum speed to
the push of a button. In addition, this opera- a set value, even if the driver continues to de-
tor unit can be used for preselecting specific press the accelerator pedal. On very quiet ve-
engine speeds. The intermediate-speed con- hicles, where the engine can hardly be heard,
trol is also applied on passenger cars with this is a particular help for the driver who can
automated transmissions (e.g. Tiptronic) to no longer exceed speed limits inadvertently.
control the engine speed during gearshifts.
The vehicle-speed limiter keeps the injected
Vehicle-speed controller fuel quantity down to a limit corresponding
(cruise control) to the selected maximum speed. It can be de-
Cruise control allows the vehicle to be driven activated by pressing the lever or depressing
at a constant speed. It controls the vehicle the kickdown switch. In order to reselect the
speed to the speed selected by the driver with- last speed setting, the driver only needs to
out him/her needing to press the accelerator press the lever to the restore position.
pedal. The driver can set the required speed The operator controls can also be used for
either by operating a lever or by pressing but- a step-by-step change of the selected speed.
tons on the steering wheel. The injected fuel
quantity is either increased or decreased until Fixed limitation
the desired (set) speed is reached. In a number of countries, fixed maximum
speeds are mandatory for certain classes of
On some cruise-control applications, the vehicles (for instance, for heavy trucks).
vehicle can be accelerated beyond the current Vehicle manufacturers also limit the maxi-
set speed by pressing the accelerator pedal. mum speeds of their heavy vehicles by in-
As soon as the accelerator pedal is released, stalling a fixed speed limit which cannot be
cruise control regulates the speed back down deactivated.
to the previously set speed.
If the driver depresses the clutch or brake In the case of special vehicles, the driver
pedal while cruise control is activated, control can also select from a range of fixed, pro-
is terminated. On some applications, the con- grammed speed limits (for instance, when
trol can be switched off by the accelerator workers are standing on the platform of a
pedal. garbage truck).
Active-surge damping Smooth-running control (SRC)/

Sudden engine-torque changes excite the ve- Control of injected-fuel-quantity
hicle’s drivetrain, which, as a result, goes into compensation (MAR)
bucking oscillation. These oscillations are Presuming a constant injection duration,
perceived by the vehicle’s occupants as un- not all of the engine’s cylinders generate the
pleasant periodic changes in acceleration same torque. This can be due to differences
(Fig. 2, a). The function of the active-surge in cylinder-head sealing, as well as differences
damper (ARD) is to reduce these changes in in cylinder friction, and hydraulic-injection
acceleration (b). components. These differences in torque out-
Two different methods are used: put lead to rough engine running and an in-
쐌 In case of sudden changes in the torque re- crease in exhaust-gas emissions.
quired by the driver (through the accelera-
tor pedal), a precisely matched filter func- Smooth-running control (LRR) or fuel-
tion reduces drivetrain excitation (1). balancing control (MAR) have the function
쐌 The speed signals are used to detect drive- of detecting these differences based on the
train oscillations which are then damped resulting fluctuations in engine speed, and
by an active control. In order to counteract to compensate by adjusting the injected fuel
the drivetrain oscillations (2), the active quantity in the cylinder affected. Here, the
control reduces the injected fuel quantity rotational speed at a given cylinder after in-
when rotational speed increases, and injection is compared to a mean speed. If the
creases it when speed drops. particular cylinder’s speed is too low, the in-
jected fuel quantity is increased; if it is too
high, the fuel quantity is reduced (Fig. 3).
2 Example of active-surge damper (ARD) 3 Example of smooth-running control (LRR)
Desired (setpoint) speed: 800 rpm

rpm
1,000 a
Engine speed n
Actual speed:
Cyl. 1 Cyl. 2 Cyl. 3 Cyl. 4
800 rpm 800 790 820 790
b
mm3
a
Injected fuel quantity
25 Injected
1+2 fuel = + +
Fig. 2 b quantity
20
a Without active-surge 1
æ NMK1557-1E
damper 2
æ NMK1792E
b With active-surge 12
damper
0 1 2s
1 Filter function Time t
2 Active correction
Smooth-running control is a convenience fea- Altitude compensation

ture. Its primary object is to ensure that the Atmospheric pressure drops as altitude in-
engine runs smoothly at near-idle. The increases so that the cylinder is charged with less
jected-fuel-quantity compensation function combustion air. This means that the injected
is aimed at not only improving comfort at fuel quantity must be reduced accordingly,
idle, but also at reducing exhaust-gas emis- otherwise excessive soot will be emitted.
sions in the medium-speed ranges by ensur- In order that the injected fuel quantity can
ing identical injected fuel quantities for all be reduced at high altitudes, the atmospheric
cylinders. pressure is measured by the ambient-pressure
On commercial vehicles, smooth-running sensor in the ECU. This reduces the injected
control is also known as the AZG (adaptive fuel quantity at higher elevations. Atmospheric
cylinder equalization). pressure also has an effect on boost-pressure
control and torque limitation.
Injected-fuel-quantity limit
There are a number of reasons why the fuel Cylinder shutoff
quantity actually required by the driver, or If less torque is required at high engine
that which is physically possible, should not speeds, very little fuel needs to be injected. As
always be injected. The injection of such fuel an alternative, cylinder shutoff can be applied
quantities could have the following effects: to reduce torque. Here, half of the injectors
쐌 Excessive exhaust-gas emissions are switched off (commercial-vehicle UIS,
쐌 Excessive soot UPS, and CRS). The remaining injectors then
쐌 Mechanical overloading due to high torque inject correspondingly more fuel which can
or excessive engine speed be metered with even higher precision.
쐌 Thermal overloading due to excessive tem- When the injectors are switched on and off,
peratures of the exhaust gas, coolant, oil, or special software algorithms ensure smooth
turbocharger transitions without noticeable torque
쐌 Thermal overloading of the solenoid valves changes.
if they are triggered too long
To avoid these negative effects, a number

of input variables (for instance, intake-air
quantity, engine speed, and coolant tempera-
ture) are used to generate this limitation fig-
ure. The result is that the maximum injected
fuel quantity is limited and with it the maxi-
mum torque.
Engine-brake function
When a truck’s engine brake is applied, the
injected fuel quantity is either reduced to
zero, or the idle fuel quantity is injected. For
this purpose, the ECU detects the position of
the engine-brake switch.
Injector delivery compensation Average delivery adaption

New functions are added to common-rail The deviation of the actually injected fuel
(CR) and UIS/UPS systems to enhance the quantity from the setpoint value is required
high precision of the fuel-injection system to adapt exhaust-gas recirculation and
further, and ensure them for the service life of charge-air pressure correctly. The average de-
the vehicle. livery adaption (MMA), therefore determines
With injector delivery compensation the average value of the injected fuel quantity
(IMA), a mass of measuring data is detected for all cylinders from the signals received
for each injector during the injector manu- from the lambda oxygen sensor and the air-
facturing process. The data is then affixed mass sensor. Correction values are then calcu-
to the injector in the form of a data-matrix lated from the setpoint and actual values (see
code. With piezo-inline injectors, data on lift “Lambda closed-loop control for passenger-
response is included. This data is transferred car diesel engines”).
to the ECU during vehicle production. While The MMA teach-in function ensures a con-
the engine is running, these values are used to stant level of favorable exhaust-gas emission
compensate for deviations in metering and values in the lower part-load range for the
switching response. service life of the vehicle.
Zero delivery calibration Pressure-wave correction

The reliable mastery of small pre-injection Injection events trigger pressure waves in the
events for the service life of the vehicle is line between the nozzle and the fuel rail in all
vitally important to achieve the required level CR systems. These pressure pulses affect sys-
of comfort (through reduced noise) and ex- tematically the injected fuel quantity of later
haust-gas emission targets. There must be injection events (pre-injection/main injec-
some form of compensation for fuel-quantity tion/secondary injection) within a combus-
drifts in the injectors. For this reason, a small tion cycle. The deviations of later injection
quantity of fuel is injected in one cylinder in events are dependent on the fuel quantity
overrun conditions in second- and third- previously injected, the time interval between
generation CR systems. injection events, rail pressure, and fuel tem-
The wheel-speed sensor detects the resulting perature. The control unit can calculate a
torque increase as a minor dynamic change correction factor by including these parame-
in engine speed. This increase in torque, ters in suitable compensation algorithms.
which remains imperceptible to the driver, However, extremely high application re-
is clearly linked to the injected fuel quantity. sources are required for this correction func-
The process is then repeated for all cylinders tion. The benefit is the possibility of flexibly
and at various operating points. A teach-in adjusting the interval between pre-injection
algorithm detects minor changes in pre- and main injection, for example, in order to
injection quantity and corrects the injector optimize combustion.
triggering period accordingly for all pre-
injection events.
Electronic control unit Injector delivery compensation 467
왘 Injector delivery compensation
Functional description
왘 EMI characteristic curve without IMA
Injector delivery compensation (IMA) is a soft-
ware function to make fuel quantity metering 1,600 bar 1,200 bar
more precise and increase injector efficiency Fig. 1

600 bar
on the engine. The feature has the function of Curves of various
Injected-fuel
correcting injected fuel quantity to the setpoint injectors as a function
quantity
value over the entire program map individually of rail pressure.
æ SMK1999E
300 bar IMA reduces curve
for every injector in a CR system. This re-
duces system tolerances and exhaust-gas spread.
EMI Injected-fuel-
emission spread. The compensation values Injection time
quantity indicator
required for IMA represent the difference from
the setpoint value of each factory test point,
and are inscribed on each injector in encoded 왘 Considering the matrix in the injection
calculation
form.
The entire engine environment is corrected Setpoint quantity Q Δ Q Triggering-period
+
calculation
by means of a correction program map that
uses compensation values to calculate a cor- Rail pressure p
Injected-fuel
quantity
rection quantity. At the end of the line of the Flash EPROM

compensation Injection time
car assembly plant, the EDC compensation program map

Triggering-period
values belonging to the injectors fitted and x
program map
their cylinder assignment are programmed in
the electronic control unit using EOL program- For all injectors
of identical type
ming. The compensation values are also re-
Data matrix code
programmed when an injector is replaced at Plaintext code
the customer service workshop.
EEPROM Cylinder 1
Fig. 2
Necessity for this function Compensation values
EEPROM
injector-specific Calculation of injector
æ SMK2000E
The technical resources required for a further Compensation val

Cylinder 2
EEPROM
injector-specific triggering period based
Compensation val
restriction of the manufacturing tolerances for injector-specific
Cylinder 3
on setpoint quantity, rail
Compensation val
injectors rise exponentially and appear to be Cylinder 4
injector-specific pressure, and correction
financially unprofitable. IMA is a viable solution values
to increase efficiency, enhance the metering
precision of fuel quantity injected in the en- 왘 Process chain
gine, and reduce exhaust-gas emissions. BOSCH BOSCH
Measured values in testing

The end-of-line test measures every injector at
Injector test Injector label
several points that are representative for the
spread of the particular injector type. Devia- Engine manufacturer BOSCH
tions from setpoint values at these points

(compensation values) are calculated and Fig. 3
then inscribed on the injector head. Fitting the injector 100% retest Schematic of process
chain from injector
Engine manufacturer Vehicle manufacturer
delivery compensation
æ SMK2002E
Inj. 1 EDC programming

Inj. 2
Inj. 3
with IMA data at Bosch through to
Inj. 4
end-of-line programming
EDC
Code
readout at the vehicle manufac-
by camera
turer’s plant
468 Electronic diesel control Lambda closed-loop control
Lambda closed-loop control A broadband lambda oxygen sensor in the

exhaust pipe (Fig. 1, 7) measures the residual
for passenger-car diesel oxygen content in the exhaust gas.
engines This is an indicator of the A/F ratio (excess-
Application air-factor lambda λ). The lambda oxygen-
The lawmakers are continually increasing the sensor signal is adapted while the engine is
severity of legislation governing exhaust-gas running. This ensures a high level of signal
emission limits for cars powered by diesel accuracy throughout the sensor’s service life.
engines. Apart from the measures taken to The lambda oxygen-sensor signal is used as
optimize the engine’s internal combustion, the basis for a number of lambda functions,
the open and closed-loop control of functions which will be described in more detail in the
related to exhaust-gas emissions are continu- following.
ing to gain in importance. Introduction of
lambda closed-loop control offers major Lambda closed-loop control circuits are used
potential for reducing emission-value spread to regenerate NOX accumulator-type catalytic
in diesel engines. converters.
1 System overview of lambda closed-loop control for passenger-car diesel engines (example)
2
λ control
4
6
3
Fig. 1 8
1 Diesel engine
2 Diesel injection
7
component
(here, common-
rail injector)
3 Control flap
4 Hot-film air-mass
1
meter
5 Exhaust-gas
turbocharger
(here, VTG version)
6 Engine ECU for
æ UMA0031Y
EDC
7 Broadband lambda
oxygen sensor
8 EGR valve
Electronic diesel control Lambda closed-loop control 469
Lambda closed-loop control is designed for Adaption

all passenger-car fuel-injection systems with In overrun mode (trailing throttle), lambda
engine control units dating dating from the oxygen-sensor adaption takes into account
EDC16 generation. the deviation of the measured oxygen con-
centration from the fresh-air oxygen concen-
Basic functions tration (approx. 21%). As a result, the system
Pressure compensation “learns” a correction value which is used at
The unprocessed lambda oxygen-sensor every engine operating point to correct the
signal is dependent on the oxygen concentra- measured oxygen concentration. This leads
tion in the exhaust gas and the exhaust-gas to a precise, drift-compensated lambda out-
pressure at the sensor installation point. The put signal for the service life of the lambda
influence of pressure on the sensor signal oxygen sensor.
must, therefore, be compensated.
Lambda-based EGR control
The pressure-compensation function incorpo- Compared with air-mass-based exhaust-gas
rates two program maps, one for exhaust-gas recirculation, detecting oxygen concentration
pressure, and one for pressure dependence of in the exhaust gas allows tighter emission
the lambda oxygen-sensor output signal. tolerance bands for an automotive manufac-
These two maps are used to correct the sensor turer’s entire vehicle fleet. For future limits,
output signal with reference to the particular an emission advantage of approx. 10 to 20%
operating point. can be gained in this way for the exhaust-gas
test.
2 Operating concept of average delivery adaption in “indirect control” mode
Engine ECU
Lambda oxygen Calculating the

sensor injected fuel + - Desired injected
quantity from fuel quantity
lambda
Hot-film +
air-mass meter Start-of-injection
control +
Injection system Program map -

for EGR +
desired-value
Engine-speed sensor Desired air mass
Adaption
Air-mass program map
controller
EGR valve
Boost-pressure
control
Turbocharger
æ UAE0899-1E
470 Electronic diesel control Lambda closed-loop control
Average delivery adaption fuel quantity correction, it can be imple-

Average delivery adaption supplies a precise mented without delay even during dynamic
injection quantity signal to form the setpoint changes of state.
for the exhaust-gas-related closed control
loop. Correction of exhaust-gas recirculation These correction quantities are stored in
plays a major role in emissions here. Average the EEPROM of the ECU and are available
delivery adaption operates in the lower part- immediately the engine is started.
load range and determines the average devia-
tion in the injected fuel quantity of all cylin- Basically speaking, there are two average-
ders. delivery adaption operating modes. They dif-
fer in the way they apply detected deviations
Fig. 2 (previous page) shows the basic struc- in injected fuel quantity:
ture of average delivery adaption and its in-
fluence on the exhaust-gas-related closed Operating mode: Indirect Control
control loops. In Indirect Control mode (Fig. 2), a precise in-
jection quantity setpoint is used as the input
The lambda oxygen-sensor signal and the air- variable in various exhaust-gas-related refer-
mass signal are used to calculate the actually ence program maps. The injected fuel quan-
injected fuel mass, which is then compared to tity is not corrected during the fuel-metering
the desired injected fuel mass. Differences are process.
stored in an adaption map in defined “learn-
ing points”. This procedure ensures that,
when the operating point requires an injected
3 Full-load smoke limitation using the lambda closed-loop control: Principle of operation
Engine ECU
Calculation of the
Hot-film preliminary quantity
air-mass meter
Desired-value map λdesired -

for smoke limitation +
Engine-speed sensor
+
+
λactual Lambda controller
Lambda oxygen Smoke-limitation quantity

sensor Lowest
value
Calculation of the
injected fuel quantity
Injection system
æ UAE0900E
Electronic diesel control Lambda closed-loop control 471
Operating mode: Direct Control Detection of undesirable combustion

In Direct Control mode, the quantity devia- The lambda oxygen sensor signal helps to
tion is used in the metering process to correct detect the occurrence of undesirable combus-
the injected fuel quantity so that the actual tion in overrun mode. It is detected if the
fuel quantity injected coincides more pre- lambda oxygen-sensor signal drops below a
cisely with the reference injected fuel quan- calculated threshold. In this case, the engine
tity. In this case, this is (more or less) can be switched off by closing a control flap
a closed quantity control loop. and the EGR valve. The detection of undesir-
able combustion represents an additional
Full-load smoke limitation engine safeguard function.
Fig. 3 shows the block diagram of the control
structure for full-load smoke limitation using Summary
a lambda oxygen sensor. The objective here A lambda-based exhaust-gas recirculation
is to determine the maximum fuel quantity system can substantially reduce emission-
which may be injected without exceeding a value spread over a manufacturer’s vehicle
given smoke-emission value. fleet due to production tolerances or aging
drift. This is achieved by using average
The signals from the air-mass meter and delivery adaption.
the engine-speed sensor are applied together
with a smoke-limitation map to determine the Average delivery adaption supplies a precise
desired air/fuel ratio value λdesired. This, in injection quantity signal to form the setpoint
turn, is applied together with the air mass to for the exhaust-gas-related closed control
calculate the precontrol value for the maxi- loop. The precision of these control loops is
mum permissible injected fuel quantity. increased as a result. Correction of exhaust-
gas recirculation plays the major role on emis-
This form of control is already in serial pro- sions here.
duction, and has a lambda closed-loop control
imposed on it. The lambda controller calcu- In addition, the application of lambda closed-
lates a correction fuel quantity from the dif- loop control permits the precise metering of
ference between the desired air/fuel ratio λde- the full-load smoke quantity and detection
sired and the actual air/fuel ratio value of undesirable combustion in overrun (trail-
λactual. The maximum full-load injected fuel ing throttle) mode.
quantity is the total of the pilot-control quan-
tity and the correction quantity. Furthermore, the lambda oxygen sensor’s
high-precision signal can be used in a
This control architecture permits a high level lambda closed control loop to regenerate
of dynamic response due to pilot control, and NOX catalytic converters.
improved precision due to the superimposed
lambda control loop.
472 Closed-loop and open-loop control
왘 Closed-loop and open-loop control
Application values (z1, z2) occurring within the control

The closed-loop and open-loop control appli- loop. Examples of closed-loop systems in a
cations are of vital importance for various vehicle:
on-board systems. 쐌 Lambda closed-loop control
The term (open-loop) control is used in 쐌 Idle-speed control
many cases, not only for the process of con- 쐌 ABS/TSC/ESP control
trolling, but also for the entire system in which 쐌 Air conditioning (interior temperature)
control takes place (for this reason, the gen-
eral term “control unit” is used, although it Open-loop control
may perform a closed-loop control function). Open-loop control is the process within a sys-
Accordingly, arithmetic processes run in con- tem in which one or several parameters act as
trol units to calculate both closed-loop and input variables affecting other parameters due
open-loop functions. to intrinsic laws governing the system. A fea-
ture of open-loop control is the open action
Closed-loop control sequence across an individual transfer ele-
Closed-loop control is a process in which a ment or the open control loop.
parameter (controlled variable x) is detected An open control loop (Fig. 1b) is an
continuously, compared to another parameter arrangement of elements that interact on each
(reference variable w1), and adapted to the other in a loop structure. It may interact in any
reference variable in an adjustment process possible way with other systems as an entity
depending on the result of the comparison. within a higher-level system. The open control
The resulting action takes place in a closed loop can only counter the impact of a distur-
circuit (closed control loop). bance value measured by the control unit
Closed-loop control has the function of (e.g. z1); other disturbance values (e.g. z2)
adjusting the value of controlled variables to a may act unimpaired. Examples of open-loop
value specified by a reference variable, despite systems in a vehicle:
any disturbance influences that may occur. 쐌 Electronic Transmission Control (ETC)
The closed control loop (Fig. 1a) is a 쐌 Injector delivery compensation and pres-
closed-loop control circuit with a discrete ac- sure-wave correction for calculating injected
Fig. 1
a Closed control tion. Controlled variable x acts within a loop fuel quantity
loop configuration in a form of negative feedback.
b Open control loop Contrary to open-loop control, closed-loop
c Block diagram of control considers the impact of all disturbance
a digital closed-
control loop
1 Closed-loop and open-loop control applications
w Reference variable a z1 z2 b z1 z2
x Controlled variable w y x w
y xA
(closed loop) Con- Actuator Closed Control Open
unit Actuator loop
troller loop
xA Controlled variable
(open loop)
y Manipulated c Closed- z
variable w w* loop y* y x
A controller D Holding Closed
Scanner
z1, z2 Disturbance values D (e.g. μC) A unit loop
æ UAN0168E
T x*
T Sampling time D Scanner Measur-
A ing unit
* Digital signal
values T
A Analog
D Digital
Electronic diesel control Torque-controlled EDC systems 473
Torque-controlled given torque from the engine. At the same

time, but independent of the driver’s require-
EDC systems ments, via the interfaces other vehicle systems
The engine-management system is continu- submit torque demands resulting from the
ally being integrated more closely into the power requirements of the particular compo-
overall vehicle system. Through the CAN bus, nent (e.g. air conditioner, alternator). Using
vehicle dynamics systems such as TCS, and these torque-requirement inputs, the engine
comfort and convenience systems such as management calculates the output torque to
cruise control, have a direct influence on the be generated by the engine and controls the
Electronic Diesel Control (EDC). Apart from fuel-injection and air-system actuators ac-
this, much of the information registered cordingly. This method has the following
and/or calculated in or by the engine man- advantages:
agement system must be passed on to other 쐌 No single system (for instance, boost pres-
ECUs through the CAN bus. sure, fuel injection, pre-glow) has a direct
effect on engine management. This enables
In order to be able to incorporate the EDC the engine management to also take into
even more efficiently in a functional alliance account higher-level optimization criteria
with other ECUs, and implement other (such as exhaust-gas emissions and fuel
changes rapidly and effectively, it was neces- consumption) when processing external
sary to make far-reaching changes to the requirements, and thus control the engine
newest-generation controls. These changes in the most efficient manner,
resulted in the torque-controlled EDC which 쐌 Many of the functions which do not di-
was introduced with the EDC16. The main rectly concern the engine management
feature is the changeover of the module can be designed to function identically
interfaces to the parameters, as commonly for diesel and gasoline engines.
encountered in practice in the vehicle. 쐌 Extensions to the system can be imple-
mented quickly.
Engine parameters
Essentially, an IC engine’s output can be de-
fined using the three parameters: power P, 1 Example of the torque and power-output curves as
a function of engine speed for two passenger-car
engine speed n, and torque M. diesel engines with approx. 2.2 l displacement
For 2 diesel engines. Fig. 1 compares typical
curves of torque and power as a function of kW
engine speed. Basically speaking, the follow-
75
Power output
ing equation applies:

50 b
P = 2·π·n·M
25
a
In other words, it suffices to use the torque as 0
the reference (command) variable. Engine N·m
power then results from the above equation.
Since power output cannot be measured di- 300 b
Torque
rectly, torque has turned out to be a suitable

200
reference (command) variable for engine
management. 100 a
æ NMM0556E
Fig. 1
0 a Year of manufacture
Torque control 1968
0 1,000 2,000 3,000 4,000 rpm
When accelerating, the driver uses the accel- Engine speed b Year of manufacture
erator pedal (sensor) to directly demand a 1998
474 Electronic diesel control Torque-controlled EDC systems
Engine-management sequence Internal torque demands

Fig. 2 shows (schematically) the processing of At this stage, the idle-speed control and the
the setpoint inputs in the engine ECU. In or- active surge damper intervene.
der to be able to fulfill their assignments effi- For instance, if demanded by the situation,
ciently, the engine management’s control in order to prevent mechanical damage, or
functions all require a wide range of sensor excessive smoke due to the injection of too
signals and information from other ECUs in much fuel, the torque limitation reduces the
the vehicle. internal torque requirement.
In contrast to the previous engine-manage-
Propulsion torque ment systems, limitations are no longer only
The driver’s input (that is, the signal from applied to the injected fuel-quantity, but in-
the accelerator-pedal sensor) is interpreted stead, depending upon the required effects,
by the engine management as the request also to the particular physical quantity in-
for a propulsive torque. The inputs from the volved.
cruise control and the vehicle speed limiter The engine’s losses are also taken into
are processed in exactly the same manner. account (e.g. friction, drive for the high-pres-
Following this selection of the desired sure pump). The torque represents the en-
propulsive torque, should the situation arise, gine’s measurable effects to the outside. The
the vehicle-dynamics system (TCS, ESP) in- engine management, though, can only gener-
creases the desired torque value when there is ate these effects in conjunction with the cor-
the danger of wheel lockup and decreases it rect fuel injection together with the correct
when the wheels show a tendency to spin. injection point, and the necessary marginal
conditions as apply to the air-intake system
Further external torque demands (e.g. boost pressure and EGR rate). The re-
The drivetrain’s torque adaptation must be quired injected fuel quantity is determined
taken into account (drivetrain transmission using the current combustion efficiency. The
ratio). This is defined for the most part by the calculated fuel quantity is limited by a protec-
ratio of the particular gear, or by the torque- tive function (for instance, protection against
converter efficiency in the case of automatic overheating),
transmissions. On vehicles with an automatic- and if necessary can be varied by smooth-
gearbox, the transmission control stipulates running control (SRC). During engine start,
the torque requirement during the actual gear the injected fuel quantity is not determined
shift. Apart from reducing the load on the by external inputs such as those from the dri-
transmission, reduced torque at this point rever, but rather by the separate “start-quantity
sults in a comfortable, smooth gear shift. In control” function.
addition, the torque required by other engine-
powered units (for instance, air-conditioner Actuator triggering
compressor, alternator, servo pump) is deter- Finally, the desired values for the injected fuel
mined. This torque requirement is calculated quantity are used to generate the triggering
either by the units themselves or by the engine data for the injection pump and/or the injec-
management. tors, and for defining the optimum operating
Calculation is based on unit power and ro- point for the intake-air system.
tational speed, and the engine management
adds up the various torque requirements. The
vehicle’s drivability remains unchanged de-
spite varying requirements from the auxiliary
units and changes in the engine’s operating
state.
Electronic diesel control Torque-controlled EDC systems 475
2 Engine-management sequence for torque-controlled diesel injection
Propulsion torque:
Driver input: Selection of the
- Accelerator-pedal desired propulsion
sensor
torque
- Vehicle-speed control
(cruise control)
- Vehicle-speed limitation
Sensor
Input from the vehicle- Coordination of the
signals
dynamics systems: propulsion torque
- TCS
- ESP Data
exchange
Further external torque demands
Drivetrain transm. Coordination of the

drivetrain torque
Input from the
transmission ECU
Engine loading due

to auxiliary units
Internal torque requirements

Idle-speed control Control of the
engine torque
Active-surge damping (internal functions)
Torque limitation
Start quantity
Engine efficiency
Start Drive
Fuel- mode
Fuel-quantity limit
quantity
input
Smooth-running control
Actuator triggering
Inputs: Input:
- Boost pressure - Start of delivery
- EGR rate - Timing device
- ... - Rail pressure
External - ... (depending on system)
inputs
Internal
sequences Intake-air system Injection system
- Turbocharger - Fuel-injection
Data trans- pump
æ NMK1763E
mission - EGR ...

- Injectors ...
possible
through CAN
476 Electronic diesel control Data exchange with other systems
Data exchange with other Glow control unit

The glow control unit (GZS, 5) receives infor-
systems mation from the engine ECU to control glow
Fuel-consumption signal start and duration. It then triggers the glow
The engine ECU (Fig. 1, 3) determines fuel plugs accordingly and monitors the glow
consumption and sends this signal via CAN process, and reports back to the engine ECU
to the instrument cluster or a separate on- on any faults (diagnostic function). The pre-
board computer (6), where the driver is in- glow indicator lamp is usually triggered from
formed of current fuel consumption and/or the engine ECU.
the range that can be covered with the re-
maining fuel in the tank. Older systems Electronic immobilizer
used pulse-width modulation (PWM) for To prevent unauthorized starting and drive-
the fuel-consumption signal. off, the engine cannot be started before a spe-
cial immobilizer (7) ECU removes the block
Starter control from the engine ECU.
The starter motor (8) can be triggered from The driver can signal the immobilizer ECU
the engine ECU. This ensures that the driver that he/she is authorized to use the vehicle,
cannot operate the starter motor with the either by remote control or by means of the
engine already running. The starter motor glow-plug and starter switch (“Ignition” key).
only turns long enough to allow the engine The immobilizer ECU then removes the
to reach a self-sustaining speed reliably. block on the engine ECU to allow engine
This function leads to a lighter, and thus start and normal operation.
lower-priced, starter motor.
1 Possible components involved in the exchange of data with the Electronic Diesel Control (EDC)
3 4 5
1 2
6
Fig. 1
1 ESP ECU
(with ABS and TCS) 7
2 ECU for transmis- 8
sion-shift control 9
3 Engine ECU (EDC)
4 A/C ECU
5 Glow control unit
6 Instrument cluster 10
with onboard
computer
æ UAE0777Y
7 Immobilizer ECU
8 Starter motor
9 Alternator
10 A/C compressor
Electronic diesel control Data exchange with other systems 477
External torque intervention Serial data transmission

In the case of external torque intervention,
the injected fuel quantity is influenced by an-
(CAN)
other (external) ECU (for instance, for trans- Modern-day vehicles are equipped with a
mission-shift control, or TCS). This informs constantly increasing number of electronic
the engine ECU whether the engine torque is systems. Along with their need for extensive
to be changed, and if so, by how much (this exchange of data and information in order
defines the injected fuel quantity). to operate efficiently, the data volumes and
speeds are also increasing at a rapid rate.
Alternator control
By means of a standard serial interface, the Although CAN (Controller Area Network)
EDC can control and monitor the alternator is a linear bus system (Fig. 1) specifically
(9) remotely. The regulator voltage can be designed for automotive applications, it has
controlled, just the same as the complete already been introduced in other sectors
alternator assembly can be switched off. In (for instance, in building automation).
case of low battery power, for instance, the al- Data is relayed in serial form, that is, one
ternator’s charging curve can be improved by after another on a common bus line. All
increasing the idle speed. It is also possible to CAN stations have access to this bus, and
perform simple alternator diagnosis through via a CAN interface in the ECUs, they can
this interface. receive and transmit data over the CAN bus
line. Since a considerable amount of data can
Air conditioner be exchanged and repeatedly accessed on a
In order to maintain comfortable tempera- single bus line, this network results in far
tures inside the vehicle when the ambient fewer lines required. On conventional sys-
temperature is high, the air conditioner (A/C) tems, data exchange takes place point to point
cools down cabin air with the help of an over individually assigned data lines.
A/C compressor (10). Depending on the
engine and operating conditions, the A/C
compressor may draw as much as 30% of the
engine’s output power.
Immediately the driver hits the accelerator

pedal (in other words he/she wishes maxi- 1 Linear bus topology
mum torque), the compressor can be Engine
Transmission-shift
switched off briefly by the engine ECU control management
to concentrate all of the engine’s power to Station 1 Station 2
the wheels. Since the compressor is only

switched off very briefly, this has no notice-
able effect on interior temperature.
CAN
æ UAE0283-2E
ABS/TCS/ESP Instrument cluster

Station 3 Station 4
478 Active steering Purpose, design
Active steering
The development of vehicle steering sys- Purpose
tems is characterized by the consistent
introduction of hydraulic servo assist and The newly developed active steering can
the replacement of ball-and-nut-type steer- affect the steering forces and the steering
ing in the car by the easier and more inex- angle set by the driver. It fulfills the wish for
pensive rack-and-pinion steering. Recently, a direct steering ratio to improve handling at
electromechanical power steering has been low speeds. It also meets requirements for
displacing hydraulic power steering in small ensuring comfort, drivability and straight-
and lightweight cars. By law, however, pure running stability at high speeds. Active steer-
“steer-by-wire” technology is not yet per- ing is an initial step towards a “steer-by-
mitted in motor vehicles. European Union wire” function. Although it does not let the
safety regulations still require a mechanical car drive itself, it provides correction func-
connection between the steering wheel and tions and added comfort and convenience.
the wheels of the vehicle.
Design
All of these developments have the goal of
making vehicle handling as easy as possible The primary difference between active steer-
and to limit steering forces to a logical ing and a “steer-by-wire” system is the fact
amount. The best possible feedback about that the steering train, and thus the driver’s
the contact of the tires to the road is to be mechanical control of the steered front
ensured. This has a decisive impact on the wheels, is maintained during active steering.
driver’s ability to manage his or her task in
the control loop between driver, vehicle and
environment.
1 Variable steering ratio
Steering wheel ΩS= 320° ΩS= 406° ΩS= 512°

angle
Motor angle ΩM= 7000° ΩM = 0° ΩM = -9000°
ΩF= 32° ΩF= 32° ΩF= 32°

Front wheel
angle
iv
iv max indirect
Fig. 1
Changing the ratio
iv mechanical
between the steering
wheel angle and the
iv min direct
median angle of the front
æ SFL0033E
wheels ➞ Reducing the 0

steering effort at low 0 60 100 km/h
speeds and providing Vehicle speed υx
stability at high speeds

Active steering Purpose, design 479
Mechanical system smaller than that set by the driver (Fig. 1).
The steering train consists, as usual, of the When the electric motor is inactive, the
steering wheel, steering column, steering steering wheel controls the vehicle’s wheels
gear and tie rods. The special feature of the directly, as with conventional steering
new active steering is a differential gearbox systems.
(Fig. 2). For this purpose, a planetary gear
(6) is integrated with two input shafts and Hydraulic system
one output shaft in the steering gear. One The principle of differential steering usually
input shaft is connected to the steering requires a hydraulic servo assist to limit the
wheel, and the other drives an electric motor forces applied by hand to a logical amount.
(4) via a worm gear pair (3) as a reduction This is accomplished using an “open center”
stage. The connected ECU processes the steering valve specially adapted to the high
necessary sensor signals, controls the electric performance requirements. The vector su-
motor and monitors the entire steering perimposition of the positioning rates of the
system. drivers and engine can, in certain cases,
cause significantly higher rack-and-pinion
The electric motor and differential gearbox speeds than those of conventional steering
allow steering intervention at the front axle systems. The geometric flow of the vane-
to take place independently of the driver. type power steering pump with flow rate
At low speeds, the effective steering angle controller is designed for the theoretical
at the wheels is greater than the angle set maximum positioning rate. Regulation at
on the steering wheel, as the system adds a the output side provides a highly dynamic
part that is proportional to the steering and quiet energy supply for the active
angle. At high speeds, it subtracts a corre- steering system.
sponding amount so that the wheel angle is
2 Actuators 3 Components of active steering
1 2
Other vehicle sensors
(e.g. ESP): Fig. 2
• Steering wheel angle 1 Servotronic II valve
• Wheel speeds 2 Electromagnetic
• Reference speed block
3 4 3 Worm
4 Electric motor
5 Steering rack
1 6 Planetary gear
7 7 Worm gear
6 2
7 Fig. 3
1 Electronic
4 3 control unit
6 2 Pinion angle sensor
5 3 Substructure
æ SFL0034Y
æ SFL0035E
Direction of travel 4 Block

5 Servotronic II valve
5
6 Actuator module
7 Motor angle sensor
480 Active steering Method of operation
Method of operation tracts a corresponding amount. Therefore,

from the driver’s point of view, the impres-
Activation concept sion given is of a steering ratio that is vari-
The actuator adjustment at the front axle, able over the driving speed. The steering
which takes place independently of the dri- effort remains largely constant over a wide
ver, requires a complex activation concept speed range. Except at extremely low speeds,
that is implemented in an ECU with two a significantly greater steering angle than
processors that communicate with each 180° is required.
other. One processor is responsible for acti-
vating the servomotor, the other for calcu- Driving stabilization
lating the correct control angle. Both proces- To calculate the stabilizing steering interven-
sors monitor each other to ensure that they tion, the vehicle movement variables of yaw
are functioning properly. The movement angular acceleration and lateral acceleration
status of the steering gear is detected using are returned and compared in the stabiliza-
one angle sensor each on the steering pinion tion controller with the setpoint specified by
and servomotor. The steering wheel angle the driver, which is made up of the steering
signal is also used as the setpoint specifica- wheel angle and driving speed (Fig. 4).
tion from the driver. Sensors for yaw veloc-
ity, lateral acceleration and wheel speed, Setpoint value
which are already in place for the driving The setpoint for the control angle to be set,
stability systems (ESP), provide additional which is calculated in the ECU of the active
input signals for the active steering. steering, can be separated into a part with
The ECU is networked to the system using open-loop control and a part with closed-
powertrain CAN and the new chassis CAN loop control. The open-loop control part,
at the required high data rate (see Fig. 3). known simply as the variable steering ratio,
Approximately 100 times a second, the nec- by definition is calculated from the reference
essary data are collected using sensors and variable only, the driver’s steering angle.
evaluated by the ECU. The ECU then de- Additional information from the closed-
cides whether, and by what amount, the loop control system known as the “vehicle”
steering angle needs to be changed. At low provides the basis for the closed-loop part.
speeds, it adds an amount proportional to The partial setpoints meet at a convergence
the steering angle; at high speeds, it sub- point. They modify the response of the vehi-
cle closed-loop con-
4 Yaw moment compensation trol system to the dri-
ver’s steering input.
These steering inter-
ventions are usually
• Different braking forces continuous and thus
generate a yaw moment
do not disturb the
• Required steering driver.
intervention is carried
out automatically
Braking forces
Lateral forces
æ SFL0036E
Fig. 4
Example: braking on
non-uniform road surface
Active steering Method of operation, safety concept, benefits of active steering for the driver 481
Cooperation with driving stability systems comparable to familiar, manageable distur-

Compared to conventional driving stability bances such as those caused by crosswinds
systems based on wheel slip control, driving or rail grooves in the road.
stabilization via steering intervention at the
front axle has different characteristics: Active steering requires no additional con-
쐌 The steering intervention is considerably trol elements because all of its subfunctions
less noticeable to the driver than the are activated automatically when the engine
clearly audible braking intervention. is started. If the internal combustion engine
쐌 The steering intervention is faster than a is not working (while the vehicle is being
radial braking intervention, which re- towed), the active steering, like conventional
quires a certain pressure build-up time. power steering, is disabled. Such situations
쐌 The braking intervention provides are indicated by a lamp in the instrument
superior stabilization performance cluster.
than the steering.
Benefits of active steering
By combining active steering (steering inter- for the driver
vention) and wheel slip control (braking in-
tervention), optimum driving stabilization is 쐌 Driving errors are corrected or compen-
attained. sated for in such a way that the driver is
not surprised by the vehicle’s reaction.
Safety concept 쐌 The steering ratio, which depends on the
driving situation, makes it easier to ma-
If the servomotor has to be switched off due neuver the car at very low speeds, as fewer
to a fault, this path is mechanically blocked. complete turns of the steering wheel are
The planetary gear then rolls off internally required with the same amount of effort.
while the worm gear is blocked, and the ve- 쐌 Greater convenience and comfort at high
hicle remains steerable with no restrictions speeds, as the driver no longer needs to
and at a constant gear ratio. Thus active fear losing control of the car by acciden-
steering is possible even if the mechanical tally applying too much steering force.
connection fails. This is a great advantage 쐌 The steering lead is another comfort and
over pure “steer-by-wire” systems. In active convenience feature. It allows more agile
steering, all relevant input signals are safe- activation in response to the steering
guarded by redundant sensors or measure- command.
ments. Two different processors calculate the
setpoint signal in the ECU. Although the
conversion of the setpoint signal in the
electromechanical converter takes place in
one channel only, no unwanted actuation in
the region relevant to vehicle dynamics is
possible due to the selection of a BLDC
motor.
The safety concept is expanded by adding an

adapted shutoff concept. Function shutoffs
range from temporary or long-term dis-
abling of the driving stabilization to limited
vehicle operation with substitute values all
the way to complete shutoff of the entire
system. The state transitions here are
482 Drive and adjustment systems Power windows
Drive and adjustment systems
Power windows The power window drive units include inte-

gral Hall-effect sensors to monitor motor
Application speed during operation. If a reduction in
Power windows have mechanisms that are speed is detected, the motor’s direction of ro-
driven by electric motors. There are two tation is immediately reversed. The window
types of systems in use (Fig. 1). The available closing force must not exceed 100 N at a
installation space assumes a prominent place spring rate of 10 N/mm. The unit automati-
among the criteria applied in determining cally overrides the force limitation function
which system to install.
쐌 Rod linkage regulator mechanism: The 1 Power windows
drive motor pinion engages with a quad- a

rant gear, which is connected to a rod
Fig. 1
linkage. The use of this type of window
a System with rod regulator mechanism is decreasing.
linkage 쐌 Flexible cable mechanism: The drive motor
b System with flexible turns a cable reel, which operates a flexi-
cable ble cable mechanism.
1 Electric gearmotor
Power window motors 1 4 2
2 Guide rail
3 Coupling
Space limitations inside the door make
narrow drive units imperative (flat motors, b
4 Rod linkage
5 Drive cable Fig. 2). The reduction gearing is a self-in-
hibiting worm gear design. This prevents the
window from opening inadvertently of its
own accord or being opened by force.
æ UKT0065-1Y
Dampers integrated in the gearing mecha-
nism provide good damping characteristics
at the end positions. 1 5 2 3
Power window control

The window is controlled manually by 2 Power window motor with integrated electronics
(force limiter)
means of a rocker switch. For greater conve-
nience, power windows can be linked to a
central or decentralized closing system
which, when the vehicle is locked, closes the
windows automatically or sets them to a
predefined partially open position for venti- 1
lation purposes.
When the windows are closed, a force
limitation device comes into effect. The
device serves to prevent human limbs
from being caught by the closing window.
In Germany, paragraph 30 of the StVZO
(Road Traffic Licensing Regulations) stipu-
lates that the force limitation device must
æ UKT0064-2Y
Fig. 2
be effective when the window is moving
upwards within a range of 200 to 4 mm 2
1 Reduction gear
(worm gear pair) from the upper edge of the window
2 Flat motor aperture.

Drive and adjustment systems Power windows, power sunroofs 483
immediately before the window enters the the use of special control systems that
door seal, allowing the motor to run to its end may be electronic or electromechanical in
position and permitting complete closure of design.
the window. The window position is moni-
tored over its entire range of movement. Open-loop control
With electromechanical control (Fig. 4),
Electronic control (Fig. 3) may be concen- mechanical interlocks on limit switches a)
trated in a central control unit, or the con- and b) ensure that the roof can either slide
trol elements may be dispersed among the or tilt open from the closed position, de-
individual power window motors in order to pending on the polarity applied to Terminals
reduce the complexity of the wiring. Future 1 and 2. Once the sunroof has been tilted or
decentralized electronics will be networked opened, a polarity changeover will initiate
via bus interfaces (LIN bus/CAN bus). The the lowering or closing process. An elec-
advantages offered are fault diagnosis for the tronic control unit featuring integral force
electronics and further reduction of the limitation allows the sliding/tilting sunroof
amount of wiring. to be incorporated in a central locking sys-
tem. The electronic control system includes
Power sunroofs a microcomputer responsible for evaluating
incoming signals and monitoring sliding
Application sunroof position. The closed and end posi-
Power sunroofs can combine tilting and tions for the sliding action of the sunroof
sliding sunroof functions. This requires are monitored with the aid of microswitches
or Hall-effect sensors. Possible auxiliary
3 Power window control unit featuring electronic force functions include:
limitation device
쐌 Preset position control
쐌 Closing via rain sensor
3 쐌 Motor speed control
1 2
M 5 쐌 Electronic motor protection
Fig. 3
æ UKT0060-2Y
1 Microcomputer
Drive systems
CAN
4 2 Relay output stage

Drive for the roof is provided by Bowden 3 Setpoint commands
cables or other torsion and pressure-resis- 4 Multiplex bus
tant control cables. The drive motor is usu- 5 Hall-effect sensors
ally installed in the roof or at the rear of
the vehicle (e.g. in the trunk). Permanently
4 Power sunroof drive unit with electromechanical
control of opening, closing and tilt functions energized worm drive motors with outputs
of approximately 30 W are used as the drive
motors. They are protected against thermal
overload by a thermostatic switch (phasing
out) or (most frequently) by a software
b a thermal protection switch.
1
2
21 Provision must also be made for ensuring
M that the roof can be closed using simple on-
board tools (such as a crank) if the electrical
æ UKT0061-1Y
system fails.
484 Drive and adjustment systems Seat and steering column adjustment
Seat and steering column The optional programmable electric seat ad-
justment (“memory seat”) can recall several
adjustment preset seat positions. Potentiometers or Hall-
Application effect sensors provide seat position feedback.
Electric power seats (Fig. 1) are still primar- On two-door vehicles, a feature designed to
ily installed in passenger cars in the mid- slide the front seat to its extreme forward
price and high-price ranges. The primary position can be incorporated to help rear
function of these systems is to enhance com- passengers enter the vehicle.
fort, but multiple adjustment options com- Electrically adjustable steering columns
bined with space restrictions and difficult are also on the increase as another means of
access to the controls limit the practicality of enhancing driver comfort. The adjustment
manually operated mechanisms. Up to seven mechanism, consisting of a single electric
motors control the following functions: motor and self-inhibiting gear set for each
쐌 Seat cushion height adjustment, front/rear adjustment plane, forms an integral part of
쐌 Fore/aft seat adjustment the steering column. The gear set for fore/aft
쐌 Seat cushion depth adjustment adjustment must be capable of absorbing all
쐌 Backrest tilt adjustment impact forces (crash forces) that might be
쐌 Lumbar support adjustment, height/ applied to the steering column. The adjust-
curvature ment can be triggered in either of two ways,
쐌 Shoulder support tilt angle (top third of by the manual position switch or by the pro-
backrest) grammable seat adjustment. The column
쐌 Head restraint height adjustment can also be tilted upward when the ignition
is switched off to help the driver enter or
Adjustment systems exit the vehicle.
A common seat bottom frame includes four
motors, which drive the height adjustment
gearing or the combined fore/aft and height
adjustment gearing. The unit for adjusting the
seat height is not included on simpler seat de- 1 Electromechanical seat adjustment (example)
signs. Another concept features three identical

motor-and-gear assemblies with four height
and two fore/aft adjustment gear sets. The
motor-and-gear assemblies drive the gear sets 4
via flexible shafts. This is a highly universal
system and can be installed on any seat design.
Modern seats (especially for sportscars) 2
do not merely fasten the lap belt to the seat
Fig. 1 1
frame, they also attach the shoulder strap,
Positioning units for
including its height adjustment, inertia reel
1 Curvature of the
backrest
and seat belt pretensioner mechanism, to the
2 Angle adjustment backrest. This type of seat design ensures
of the backrest optimized seat belt positioning for a wide
3 Depth adjustment range of different occupant sizes at all avail- 5
of the seat cushions able seat positions, and makes an important
4 Height adjustment
contribution to occupant safety. The seat
æ UKP0004-1Y
of the head restraint

5 Height adjustment
frame must be reinforced for this type of
6
of the seat design, while both the gear set components 3
6 Fore/aft seat and their connections to the frame must be
adjustment strengthened.
Heating, ventilation and air conditioning Electronic heater control, electrically controlled air conditioning system 485
Heating, ventilation and air conditioning

The vehicle’s heating, ventilation and air in order to maintain an agreeable tempera-
conditioning systems have the following ture. Electronic heater control dispenses
tasks: with this requirement by automatically
쐌 Providing a pleasant climate for all maintaining the interior temperature of the
occupants of the car at different outside vehicle at the desired level.
temperatures (Fig. 1). With heater units featuring water-side
쐌 Ensuring good visibility through all control, sensors monitor the temperatures of
windows and the windshield. the vehicle’s interior and of the emerging air.
쐌 Creating an environment calculated to The control unit processes this information
minimize driver stress and fatigue. and compares it with the preselected temper-
쐌 More recent units use filters to remove ature. Meanwhile, a solenoid valve installed
particulate matter (pollen, dust) and even in the cooling circuit opens and closes at a
odors from the air. given frequency in response to the signals
which it receives from the control unit. Ad-
Electronic heater control justments in open/close ratio in the cycle pe-
riods regulate the flow rate from the closed
Function position up to the maximum.
In many countries, the performance of the An adjustment flap, actuated by an elec-
heater unit is governed by legal require- tric gearmotor, is the usual means of provid-
ments, with emphasis on the defroster’s ing infinitely variable temperature control in
ability to maintain clear windows and wind- air-side systems (pneumatic linear adjusters
shields (such as EEC Directive 78/317 within are also occasionally used). Sophisticated
the European Union, and MVSS 103 in the systems allowing separate adjustment for the
USA). left and right sides of the vehicle are also
available.
Method of operation
Variations in outside air temperature and Electronically controlled
vehicle speed cause fluctuations in the tem- air conditioning system
perature of the passenger compartment.
Without a control system, the heater con- Function
trols must be constantly readjusted by hand The heater unit alone is not capable of
providing a comfortable environment at
1 Comfortable interior temperature and airflow rate all times. When outside temperatures
as a function of outside temperature
exceed 20°C, the required inside tempera-
tures can be attained only by cooling the air
°C kg /
min using compressor refrigeration systems and
the refrigerant R 134 (Fig. 2, next page). The
28 8
moisture contained in the cooled air is also
Interior temperature t i
extracted as condensate, thus reducing the

Air flow rate m L
26 6
air’s humidity to the desired level.
mL
Automatic climate control is particularly
24 4 useful for vehicles in which both air condi-
ti tioner and heater are installed, because the
22 2 constant monitoring and adjustment re-
quired to maintain a temperate climate pre-
æ UKH0369-1E
20 0 sents the occupants with a complicated task.

–20 –10 0 10 20 30 °C
This rule applies to bus drivers in particular,
Outside temperature ta
as they can only feel the temperature at the
front of the vehicle.

486 Heating, ventilation and air conditioning Electronically controlled air conditioning system
The task of automatic climate control sys- water-side or air-side adjustments (Fig. 3).
tems incorporating a preselection feature is The fresh air (a) taken in by the fan (Item 1)
to automatically maintain the correct tem- is, depending on the temperature, either
perature, air flow, and air distribution in the cooled by the evaporator (2) or heated by
passenger compartment. These parameters the heater core (4) and then, depending on
are mutually interdependent, and changes to the flap position, guided into the desired
one will affect the others. areas of the interior (b, c, f).
By means of various temperature sensors
Method of operation (3, 5, 7) the ECU (8) collects all important
The compressor, powered by the motor influencing and disturbance values and the
(Fig. 2, Item 1) compresses and heats the temperature selected by the occupants on
gaseous coolant. It then cools in the con- the setpoint controller (6) and uses all of
denser (3) and becomes a liquid. The gener- this information to continuously calculate
ated heat dissipates into the environment. the setpoint value. This setpoint value is
An expansion valve (14) injects the cooled compared to the actual temperature, and the
refrigerant into the evaporator (11), where it calculated difference generates reference
evaporates, and the fresh air inflow removes variables in the ECU for regulating the heat-
the required evaporation heat. The moisture ing (4, 11), cooling (2, 10) and air flow (1).
is taken out of the cooled air as condensed Another function controls the position of
water, dehumidifying the air. the air distribution flaps (b, c, d, e, f) de-
At the center of the system is a tempera- pending on the program the occupants have
ture control circuit for interior temperature. selected. Meanwhile, all control circuits con-
The setpoint temperature (as described in tinue to respond to manual inputs.
“Electronic heater control”) determined by Infinitely variable or stepped blower con-
the control unit is achieved by means of trol is used to adjust the air flow to the speci-
2 Refrigerant circuit of an air conditioning system
14 13
Fig. 2
1 12
11 Compressor
12 Solenoid clutch
(for switching the
compressor on/off) 2 9
13 Condenser
14 Auxiliary fan 11
15 High-pressure 3
switch
16 Fluid reservoir with 10
desiccant insert
17 Low-pressure switch
18 Thermostatic switch
or on/off control (for
8
compressor on/off) 7
5
4
10 Condensate drip
æ UKH0371-1E
pan High pressure, liquid 6

11 Evaporator High pressure, gaseous
12 Evaporator fan Suction pressure, liquid
13 Fan switch Suction pressure, gaseous
14 Expansion valve
Heating, ventilation and air conditioning Electronically controlled air conditioning system 487
fied level. Generally, the control system does The defroster mode (“DEF” setting) is a spe-
not process the actual value in this operation. cial case. In order to clear the windows as
rapidly as possible, the temperature control
This type of arrangement is inadequate for must revert to maximum heat and maxi-
dealing with increases in flow rate caused by mum blower speed, while the air flow is di-
aerodynamic pressure at high driving rected through the upper defroster outlets.
speeds. Here, a special control function can On systems with program switches and fully
compensate by responding to increasing ve- automatic units, this operating mode is se-
hicle speeds, initially by reducing the blower lected with a single button; at temperatures
speed to zero, and then, if the flow continues above 0°C, the refrigeration unit is also acti-
to rise, by using a restriction flap to throttle vated to extract humidity from the air. To
the flow of incoming air. Air distribution to prevent unheated air from causing drafts,
the three levels – defroster (b), upper com- the blower is switched off electronically after
partment (c), and footwell (f) – is controlled cold starts in winter, except when “DEF” and
manually, with preselection, or with a fully the cooling function are in operation.
automatic program. Most popular are units
featuring program control buttons, in which The variations described above are used in
each button provides a specific air distribu- both passenger cars and trucks. Buses re-
tion pattern for the three levels. quire more complicated layouts. The passen-
ger compartment can
3 Air conditioner with electronic water-side control (schematic) be divided into several
control zones, in which
the temperature is con-
trolled by electronically
regulating the speed of
the zone’s individual
a b c
water pump.
d
Fig. 3
11 Fan
2 e 12 Evaporator
1
13 Evaporator
6 temperature sensor
5 7 14 Heater core
15 Air exit temperature
3 4 sensor
16 Setpoint controller
17 Interior sensor
(ventilated)
9 18 ECU
8
19 Water drainage
10 10 Compressor
11 Solenoid valve
a Fresh air
æ UKH0370-2Y
11 b Defroster
f c Ventilation
d Recirculated air
e Bypass
f Footwell
488 Vehicle security systems Acoustic signaling devices
Vehicle security systems
Acoustic signaling devices Horns

The mass of the armature together with the
Application flexible diaphragm forms an oscillating sys-
The internationally applicable ECE Regula- tem within the standard horn. When voltage
tion No. 28 specifies that acoustic signals is applied to the solenoid coil controlled by a
produced by motor vehicles must maintain a contact breaker, the armature impacts against
uniform sound quality with no perceptible the magnetic core at the horn’s basic fre-
frequency fluctuations during operation. quency. A fixed tone disk, directly attached to
Operation is permitted only as a hazard the armature, responds to these intense peri-
warning function. In countries where this odic impacts by radiating harmonic waves.
regulation does not apply, the acoustic sig- Legal regulations stipulate that the maximum
naling device is a wearing part subject to sound energy must lie within a band of
heavy use. The use of sirens, signal bells and 1.8-3.55 kHz. This frequency explains the
the like is not permitted, nor is playing horn’s relatively piercing sound. This sound,
melodies by triggering sound generators in a which is essentially emitted along the horn’s
given sequence. Signaling devices must be axis and outwards to the front, can be heard
installed facing forward in the vehicle and above the background noise in traffic, even
must produce the signal at a sufficient loud- over longer distances. The horn’s size is one
ness level at a distance of 2 m. They are sub- of the factors which determine the basic fre-
jected to temperatures ranging from –40°C quency and loudness level.
to +90°C and must be designed to be resis-
tant to moisture, salt spray as well as me- In addition to larger diameters, supertone
chanical shock and vibration. Elastic cou- horns have a higher electrical power input.
plings must be used to decouple electric Their warning signals can therefore still be
horns and fanfare horns from the vehicle heard under extreme conditions (e.g. truck
body, as the horn would otherwise induce driver’s cab).
sympathetic oscillations in the adjacent
bodywork. The resulting feedback would di- Fanfare horns
minish both loudness level and tone quality. The electropneumatic fanfare horn employs
Both electric and electropneumatic horns the same basic actuating system as its stan-
and fanfare horns are sensitive to series re- dard counterpart. The main difference is
sistors in their control circuits. When horns that the armature is allowed to vibrate freely
are installed in pairs, they should be trig- in front of the coil with no impact occur-
gered by relays. The warning produced by ring. The oscillating diaphragm induces vi-
the impact horn is more conspicuous; there- bration of an air column within a tube. The
fore, it is the better choice for vehicles that resonant frequency of the diaphragm and
are frequently used for long-distance over- the air column are tuned to each other and
land truck traffic instead of fanfare horns. determine the signal pitch. The tube is in the
Alternatively, the fanfare horn is superior for shape of a funnel, with a wide opening to
city driving, as pedestrians often find the enhance the efficiency of sound propaga-
standard horn excessively loud and unpleas- tion. The funnel tube is generally coiled to
ant. Both sets of requirements can be ful- minimize the unit’s size.
filled by installing the two horn types to- The presence of upper harmonics in the
gether in a single system with a selector lower range of the frequency spectrum pro-
switch for city or overland traffic. Standard vides the fanfare horn with a rich, melodi-
and fanfare horns both operate at standard- ous sound. The tone is less penetrating than
ized frequencies. High and low tones can be that of a standard horn due to the more
combined to produce a harmonious dual- consistent distribution of sound energy over
tone sound. a wide spectrum.

Vehicle security systems Central locking system 489
Central locking system multiple functions, i.e. they are also used
for the power window drive units and door
Application mirror adjustment. Electromechanical and
The central locking and unlocking of the pneumatic central locking systems are actu-
vehicle doors, trunk and fuel filler flap are ated via electrical contact switches in the ac-
an integral part of the convenience system in tuators or locking cylinders in the driver’s
motor vehicles. Electric or pneumatic actua- door, passenger’s door, trunk lid, and oper-
tors operate the mechanisms in the individ- ating switches in the vehicle interior.
ual locking devices. Infrared and radio remote control systems
enhance operating convenience.
Locking systems In newer comfort and convenience sys-
In pneumatic systems, a central electric tems, the central locking system is locked on
dual-pressure pump (vacuum/gauge pres- starting off (vehicle speed signal) to prevent
sure) and dimensionally stable lines leading the vehicle from being opened from the out-
to the diaphragms of the individual actua- side. The system is unlocked in the event of
tors provide the opening and closing move- an accident (crash sensor).
ment of the locking system. A distinction is made between opening,
The electromechanical central locking sys- closing and locking. In connection with the
tem (Fig. 1) features an electric servomotor opening and closing functions, manual op-
to drive the locking function at each point in eration from the inside (e.g. occupants) is
the system. This electric motor with a reduc- possible at any time. When the system is
tion gear drive unit actuates a mechanical
connection (rod, lever) responsible for the 1 Central locking system with electric motor
locking device (Fig. 2). + +
2 1 2
Open-loop control
All central locking systems feature a control
unit for signal processing. The control unit
is integrated in the housing in pneumatic
dual-pressure pump systems. In electro- 3
Fig. 1
mechanical systems, it is located centrally, or
æ UKT0062-2Y
1 Central switch
to minimize wiring, decentrally in the door 4 4 4 4 2 Contacts in door
modules in connection with multiplex sys- M
-
M
-
M
-
M
-
lock mechanisms
tems. The lines associated with decentral or 3 Control unit
distributed door multiplex systems have 4 Servomotors
2 Central locking mechanism servo unit
1 xx
xxxxxxxxxxxx
xxxxxx xx
xxx
xxxxxxxxxxxxxxxxxxxxxx
xxxxxx
xxxxx
xxxxxxxxxxxxxxxxxxxx xxx
x
x
xxx x
xxxxxxxxxxxxxxx xxxxxx
xxx
xx
x
xx xx
xxxxxx
xxxxxxxxxxxxxxx
xxx
xx
xxxx
4
xxxxxxxxxxxx
3 h
xx Fig. 2
xxxxxxxx
1 Wiring connection
2 Flexible end position
æ UKT0063-1Y
coupling
2 5
3 Gearing
4 Electric motor
5 Adjusting lever
h Stroke
490 Vehicle security systems Central locking systems, locking systems
locked, the locking mechanism is blocked as Depending on their location in the vehicle,
theft deterrence and can only be operated a further distinction is made between the
with the vehicle key from the outside or with following assemblies (with differing scope
the remote control. of functions):
쐌 Side door assembly
Locking systems 쐌 Trunk assembly
쐌 Hood assembly
Application
The function of locking systems is to ensure The most important component of the
and maintain access authorization at all locking system is the door lock mechanism
times. Although lock mechanisms in motor (Fig. 2). Its primary functions include:
vehicles are subject to substantial wear and 쐌 Transmitting the structural forces be-
tear as a result of prolonged cold, wet and tween the door and the vehicle body
dirty conditions over several years, their 쐌 Reliably opening and closing the locking
main design feature must be to reliably mechanism
protect the vehicle and its occupants while 쐌 Evaluating the mechanical or electrical
ensuring smooth and efficient operation. commands
쐌 Storing the logic statuses (mechanical
Structure computer).
The structure of a locking system is made up
of the following components (Fig. 1): Method of operation
쐌 Locking bars on the body pillars Accordingly, the lock mechanism is made
쐌 Side doors, door lock mechanisms, and up of assemblies that perform the following
the associated mechanical and electrical functions:
parts The locking mechanism is responsible for
쐌 Electrical components that make up the transmitting forces to initiate the opening
access authorization facility and radio and closing functions. It consists of a catch,
remote control (frequently assigned to pawl and latch.
locking systems).
1 Structure (example: side door) 2 Mechanical door lock
Side door
Lock Locking
barrel element
(key) (knob) 4 1
Exterior Interior
operating operating
Fig. 1 element element
쐍 Mechanical (handle) 2
connection
Door lock Electronics
–– Electrical signals
(option)
Actuators
Fig. 2
1 Backplate
Mechanical
logic unit
with locking
æ UKT0069-1Y
æ UKT0068-1E
mechanism
2 Catch
Lock Data Power
3 Electrical interface stirrup bus supply 3
4 Bowden cable (B-pillar)
to inner handle
Vehicle security systems Locking systems 491
The corresponding door lock striker is of the outer door handle, the inner opener
mounted on the body pillar. During the and the locking element. For instance, with
closing procedure, the door lock striker en- the door locked, the outer door handle can
gages in the latch. This centers the door, be freely pulled with no response, and it is
while the catch holds it in the closed posi- consequently not possible to open the door
tion (main catch). In turn, the pawl then from the outside.
positively locks the catch in position. In addition, the inner opener (opening
This lock status is initially canceled to ini- from inside) and the locking element (un-
tiate the opening procedure. For this pur- locking from inside) are inoperative in con-
pose, the actuating forces applied at the in- nection with a theft deterrence system.
ner or outer door handles are transmitted to The rear door locks cannot be opened
the pawl. The pawl now releases the catch to from the inside when the childproof lock is
allow the door to be opened. The door lock engaged. However, the occupants are able to
striker is fixed in position while the catch unlock the door from the inside to allow a
swivels into the open position. person providing assistance to open the
door from the outside. (Note the difference
Mechanical locking system compared to the theft deterrence system!)
Logic statuses of the lock mechanism The following status matrix (Table 1) illus-
The logic unit of the lock mechanism per- trates the basic functions of one rear door
mits the mechanical decoupling or isolation variant.
Table 1
1 Basic function of a rear door variant (“single-stroke ejection” type)
Childproof lock
æ UKT0070-1Y
OFF ON
ODH BUS CLC BUTTON IDH

Outer door LIN Central Locking Inner door
handle or CAN locking switch button handle
Command
Open – – – Open
– Unlock Unlock – Unlock
(single-stroke ejector)
– Lock Lock – –
– Activate theft – – –
deterrent
DS
Theft deterrent
activated
Status
VR
locked
ER
unlocked
OPENED
492 Vehicle security systems Locking systems
Design features then introduced in the form of flange-

The lock mechanism is located in the wet mounted locking actuators (electrically and
area inside the door. The adverse effects of pneumatically operated control elements).
water and dust, the considerable impact load These actuators are now integrated in the
that occurs when the door is slammed, and housing of more recent lock mechanisms.
the theft deterrence requirements necessitate Miniature motors carry out the electrical
a sturdy, tough design. commands from the access authorization
facility in the form of mechanical actuating
Currently, the typical locking parts and lever movements. The locking status is stored
mechanisms (Fig. 3) are made of steel mate- mechanically in the lever mechanisms.
rials (precision punched parts, springs, riv- The output from the high-speed drive
eted connections). Plastic components are units is converted in worm gear, multistage
being used to an ever greater extent. How- spur gear or planetary gear mechanisms
ever, they must fulfill their function even af- (plastic materials). Spindle drive systems
ter a side impact or in the event of a vehicle are also used.
fire. Corrosion, friction and wear dictate Sensors in the lock mechanism permit
that the metal components undergo a com- the electrical evaluation of the lock status.
prehensive surface treatment process (vibra- Hall effect sensors and microswitches have
tory grinding, galvanizing, coating). proven particularly suitable for this pur-
pose.
The parts of the locking mechanism are
sheathed in plastic materials to provide Open-by-wire
effective sound deadening in order to satisfy The widespread use of electrical systems in
requirements to minimize closing noise. motor vehicle applications initially gave rise
Higher-grade lock mechanisms normally to the central locking system, which was
feature extensive sound deadening measures soon followed by the theft deterrence system
incorporated in the lever mechanisms and and radio remote control. The childproof
actuators. lock is now switched electrically. Some door
closing and locking systems employ motors
In the past, operation of automotive locks to actuate the catch or the door lock striker
was purely mechanical. Central locking was (servo closing).
3 Locking mechanism (schematic) The obvious next step, therefore, is to use an

electric motor for the opening function, i. e.
1 to swivel the pawl (“open-by-wire”, Fig. 4).
Consequently, development of the “x-by-
wire” function for lock mechanisms has
been consistently pursued.
The main advantage of the “open-by-wire”

function is its interaction with access autho-
rization, particularly in connection with
the use of “passive entry” systems, as they
require either extremely fast unlocking
actuators or “overtake” solutions.
æ UKT0071-1Y
Fig. 3
1 Pawl
Based on the “overtake principle”, the ac-
2 Catch in lock 2 3 tuator begins the opening procedure in the
3 Door lock striker on locking mechanism before the lock logic has
B-pillar or C-pillar been completely released. This function is
Vehicle security systems Locking systems 493
intended to avoid waiting times for the user. 4 Open-by-wire (schematic)

Corresponding design measures reliably rule
out the possibility of being locked in.
Electrical locking system

M 1
Demands for quality and reliability as well
as the pressures of rising costs render it nec- 2
essary in the future to minimize the number
of mechanical parts in the locking system
and replace them by electrical components.
The end point of the development is the
electrical locking system (Fig. 5).
The lock then consists of only the locking
mechanism, opening actuator and the asso-
ciated electronics. The door handles and
other operating elements are equipped with
sensors. Electrical wires replace the mechan-
ical connections to the lock.
æ UKT0072-1Y
Fig. 4
1 Electric motor
The advantages of an electrical locking 2 Pawl
3 4
system are considerable: 3 Catch
쐌 Reduced size and weight of the lock 4 Door lock striker
쐌 Symmetrical design
쐌 Only one lock variant per vehicle (variant 5 Electromechanical door lock (schematic)
encoding at the end of the assembly line)

쐌 Door handles no longer move (or can be
dispensed with completely)
1
Auxiliary functions such as interior lighting,
status indicators and a whole range of other
functions will then be easy to implement in
electrical locking systems, as the lock is
equipped with electronics that can perform 2
these functions.
Communication with the locks, access
authorization and power supply all take
place via the data bus interfaces. Prognoses 3
using fault tree analysis indicate that the
electrical locking system is just as reliable as
conventional systems. Economically feasible
4
systems can be expected with the introduc-
tion of the 42-V vehicle electrical system at
the latest (active redundant power supply).
5 Fig. 5
æ UKT0073-1Y
1 Locking mechanism
2 Electronics
3 Electric motor
4 Gearing
5 Electrical interface
494 Vehicle security systems Biometric systems
Biometric systems Enrollment

In order to be able to recognize a user, a bio-
Function metric system must first of all learn the fin-
“Biometric systems” have the task of deter- gerprint pattern of that person. This process
mining or confirming the identity of per- is known as “enrollment”. It involves the user
sons on the basis of “biometric” features. placing his/her finger on a fingerprint sensor
Currently, approximately ten biometric that generates a grayscale image (typical size:
methods are known, including fingerprint 64,000 to 96,000 pixels at a resolution of
identification, face recognition, iris recogni- 8 bits/pixel, Fig. 1).
tion and voice recognition. Using signal-processing algorithms, a
processor calculates the characteristics
Biometric systems in motor vehicles features within that image, e.g. branches
Fingerprint identification is becoming estab- or nodal points in the line pattern. The bio-
lished in automotive applications, as crimi- metric system then stores these features –
nology has developed a fundamental under- though not as an actual fingerprint pattern –
standing of the techniques of fingerprint in a permanent database, typically in an
distinction. EEPROM memory (capacity requirement:
As fingerprint identification will in the 250 to 600 bytes per fingerprint).
future replace passwords on computers, in
time management and access control sys- When the biometric system checks a per-
tems, and on mobile phones as well, the son’s fingerprint pattern the next time, it
development of sensors and algorithms has again identifies the characteristic features of
already become widespread. Due to the con- the fingerprint and then searches the data-
fined space available on board a motor vehi- base for a matching set of features. If it finds
cle, there is also limited space for installing a match, the person is recognized as an
the sensor. authorized user.
Biometric features Examples of application

Biometric features are features that are Many vehicles already have programmable
uniquely characteristic to a particular person. seat adjustment functions. The control panel
They may either be physiological features typically has three numbered memory but-
(such as the pattern of a fingerprint) or be- tons, which can be assigned to three differ-
havioral characteristics (such as the way a ent drivers, and a teach-in button. The
person walks or writes his/her signature). The drivers of the vehicle will have to select a
advantage of biometric systems is, firstly, for memory button for each driver and then
the greater convenience of the user. Since bio- remember which button contains their own
metric features are part of a person’s individ- settings. The number of drivers that can use
uality, they are carried by the person at all the system is limited by the number of
times and cannot be forgotten or lost. There is memory buttons.
no need to carry a key or a transponder card.
Secondly, they offer the advantage of an at- The use of a biometric system significantly
tainable enhancement in security. Since bio- enhances the level of personalized conve-
metric features are intrinsic and unique to a nience. In this system, a fingerprint sensor
particular individual, they can neither be replaces the memory buttons for the pro-
deliberately given to others nor stolen in the grammable seat adjustment system (Fig. 2).
traditional sense of the word. Therefore, a Instead of pressing a memory button to re-
high level of reliance can be placed on call his/her personal settings, the driver
whether the vehicle really is being driven places a finger on the sensor. Since biometric
by a known driver. features are intrinsic to a particular individ-
Vehicle security systems Biometric systems 495
ual, the biometric system can always identify use as a vehicle immobilizer, it has not yet
a particular driver reliably and apply the been definitively demonstrated that the same
correct personal settings. degree of security is attainable as with the
If a previously unknown driver places present transponder-based immobilizers.
his/her finger on the sensor to store his/her For vehicle access applications, the sensors
personal settings, the biometric system auto- for identifying the fingerprint pattern would
matically performs an enrollment sequence have to be integrated in the vehicle’s outer
and then memorizes the current settings. skin. The technical challenge is then to de-
Once a driver is known to the biometric sign a sufficiently resilient sensor system that
system, only the current settings are stored. is adequately protected against external con-
The advantage for the user lies in the sim- ditions so that effective usability is guaran-
plified human-machine interface. Users no teed at all times – including in winter.
longer have to agree on the allocation of
memory buttons, nor do they have to re-
member which button is theirs. In addition,
the maximum number of users is not lim-
ited to the number of memory buttons 1 Fingerprint sensor based on the direct optical
scanning (DiOSTM) method (example)
available. Ultimately, the number of users is
limited by the biometric system’s memory
capacity for user identities and personal a
settings.
Personalization can be extended beyond 1 2
seat adjustment, mirror adjustment and
steering column adjustment functions.
Theoretically, all configurable vehicle sys-
tems can be connected to the biometric per- 3
sonalization system. It could be conceivable b
2
to personalize air conditioning and auto-
matic transmission settings (sports/econ-
omy). For the radio, a set of personal fa-
vorite stations and preferred volume and 4
æ UAE0921-1Y
tone settings could be selected, while there

could be list of favorite destinations for the 0 1 1 1 1 10 0 0 0 00 0 1 1 1 1 10 0 0
5
navigation system. Similarly, if the vehicle
has configurable displays, a driver’s personal
preferences for the display elements could be
recalled. 2 Seat adjustment via fingerprint sensor
In the case of self-teaching assistance sys-

tems which adapt to the driver’s driving pat-
terns, the “last known settings” regarding
driver characteristics could be retrieved by
the personalization system before starting
off.
Biometric systems can also be used even
before starting the engine as a means of im-
æ UKP0005Y
plementing a vehicle immobilizer and for

vehicle access. Vehicles equipped with this
technology could then be driven entirely
without the use of keys. With regard to the
496 Electromagnetic compatibility | EMC ranges
Electromagnetic compatibility (EMC)

and interference suppression
Electromagnetic compatibility (EMC) EMC ranges
consists of two elements. One is under-
stood as the ability of a device to con- Transmitter and receiver
tinue providing reliable service when ex- All vehicle systems must remain impervi-
posed to electromagnetism from external ous to electromagnetic radiation emitted
sources. The second aspect focuses on from such external sources as extremely
electromagnetic fields generated by the powerful radio transmitters. In other
same device; these should remain mini- words, there must be no threat to the
mal in order to avoid creating interfer- motor vehicle’s operational integrity,
ence that would impinge upon the qual- and malfunctions representing a potential
ity of radio reception, etc. in the vicinity. source of driver irritation are unaccept-
able. Another consideration is that station-
Today’s vehicle contains a multitude of ary receivers should remain unaffected by
systems whose functions are implemented passing traffic. Both considerations are
by electrical or electronic components. governed by national and international
These same functions, if present at all, codes (EU Directives, German StVZO).
used to be carried out fully or predomi-
nantly mechanically. With the exception Electrical and electronic components
of two-way transceivers used in special- Motor vehicles contain an extensive array
purpose applications, a car’s radio was of electrical and electronic components
once the only device in which signal including servo and fan motors, solenoid
reception was of any importance. valves, electronic sensors and control
In contrast, today’s vehicles feature a units with microprocessors. These devices
host of devices that rely on the reception must rely on a single vehicle electrical
of electromagnetic radio waves. Mobile system for their power supply. It is thus
phones, navigation systems, theft-deter- vital to avoid mutual interference and
rent systems with radio remote control, feedback phenomena generated by one
and integrated fax and PC units are now or several systems so that these do not
being installed and used in vehicles. cause malfunction.
This trend has led to a commensurate
increase in the importance of suppressing On-board electronic systems
interference and ensuring reception in Mobile communications equipment –
the vehicle of the operationally vital radio such as the radio – also exists within an in-
waves. terlinked environment including all of the
vehicle’s electronic systems. Every device
is powered via the same vehicle electrical
system, and its reception antennas are lo-
cated in the immediate vicinity of potential
radio-interference sources. These consid-
erations make it imperative that strict lim-
its are imposed on the levels of interfer-
ence emitted by on-board electronic sys-
tems. Compliance with official regulations
is essential, and it is important to maintain
interference-free reception in the vehicle
even when conditions are well below opti-
mal.

Electromagnetic compatibility | EMC between different systems in the vehicle 497
EMC between different to the I/O ports of adjacent systems either

systems in the vehicle directly, through shared conductors such
as the power supply (galvanic coupling),
Shared vehicle electrical system or indirectly, through capacitive and in-
All of the motor vehicle’s electrical systems ductive coupling stemming from electro-
rely on a single shared vehicle electrical magnetic emissions.
system. Because the electric lines leading
to the individual systems are frequently Galvanic coupling
combined within a single wiring harness, Currents for two different electric circuits
feedback pulses can easily travel from one (such as a solenoid valve’s trigger loop and
system to the I/O ports of its neighbor the electric circuit for assessing sensor
(Fig. 1). data) flowing through a single conductive
This transfer of interference can be in path (common ground through the vehicle
the form of signal pulses (abrupt, steep chassis, etc.) will both generate a voltage
jumps in current and voltage) generated owing to the consistent resistance in the
during the switching on and off of various shared conductor (Fig. 2a, next page).
electrical components such as electric Continuing with this sample scenario, the
motors, solenoid valves and actuators. voltage produced by radio-interference
Yet another source is the ignition system’s source u1 has the effect of a supplementary
high-tension circuit. Similar to other inter- signal voltage in signal circuit 2, and could
ference signals (for instance the ripple lead to erroneous interpretation of the
on the power supply), these signal pulses sensor signal. One remedy is to use sepa-
can propagate through the wiring harness. rate return lines for each electric circuit
These interference pulses then proceed (Fig. 2b, next page).
1 Mutual interference between two systems as transmitted through the shared vehicle electrical system (A) and
wiring harnesses (B and C)
A B C
2
I 1
3
5 Fig. 1
II 4
System I:
6
1 Control unit
2 Actuator
3 Sensor
UAE0240-1Y
System II:
4 Control unit
5 Actuator
6 Sensor
498 Electromagnetic compatibility | EMC between different systems in the vehicle
Capacitive coupling Inductive coupling

Capacitive coupling allows variable-peri- Currents recurring with variable periodicity
odicity signals such as pulse voltages and in one conductor can induce voltage pulses
sinusoidal alternating voltages to produce in adjoining circuits. These voltage pulses
interference and crosstalk in adjacent then generate current in the secondary cir-
electric circuits, even without the exis- cuit (Fig. 4). This is the inductive principle
tence of a direct physical link (Fig. 3). exploited in transformer design. One prime
The potential level of capacitive (interfer- factor defining susceptibility to overcou-
ence) voltage is proportional to such fac- pling is the signal’s rise and fall time (or al-
tors as the closeness of the neighboring ternating voltage frequency), reflecting the
conductor paths and the rise rate of the situation encountered with capacitive cou-
pulse-shaped voltage shifts (or the fre- pling. Also significant is the effective mutual
quency of the alternating voltage). The inductance, as determined by such factors
Fig. 2 first step is to separate the conductive as the size of the wires and their relative
a Electric circuit paths while at the same time extending routing. Strategies for avoiding inductive
with shared return the ramp periods during which signals coupling include minimizing the dimensions
conductor path rise and fall (or to limit the frequency of of circuit wires, keeping critical circuits as
b Electric circuit with
the alternating voltages to the absolute far apart as possible from each other, and
separate return
minimum required for the function). the avoidance of parallel conductive paths.
lines
The tendency toward inductive interference
u1, u2 Voltage source is especially pronounced in circuits carrying
Zi Internal resistor low-frequency signals (e.g. coupling into
Za Terminal resistor loudspeaker wiring).
2 Galvanic coupling of interference signals 3 Capacitive coupling of interference signals

Fig. 3
1 Electric circuit 1 a Z i,1
2 Electric circuit 2 1
C1,2
u1 Voltage source
Zi Internal resistor
Z i,1 Z i,2 Z a,2 Z a,1
RE Input resistance
2
CE Input capacitance u1
C1,2 Capacitance
RE CE
UAE0683-1Y
between two u1 u2
conductor paths us,1
us Radio interference
voltage Zk
b 4 Inductive coupling of interference signals

Fig. 4
1 Electric circuit 1 L1
Za,1
2 Electric circuit 2
1 L2
M1,2
Z i,1 Z i,2 Z a,2 Z a,1 Za,2
u1 Voltage source
2 us
u2 Voltage source
Zi Internal resistor Z i,1
Za Terminal resistor Z i,2

L1, L2 Inductance of
u1 u2
UAE0682-1Y
UAE0684-1Y
conductors u1
M1,2 Inductive coupling
us Radio interference u2
voltage
Pulsation in the vehicle electrical system within a vehicle while designing all the
One strategy for dealing with interference susceptible devices (e.g. control units)
pulses in the vehicle electrical system to comply with and exceed – by a certain
entails limiting the amplitude of the inter- safety margin – Class III. If suppressing
ference emanating from the radio-inter- interference at the source proves to be
ference sources. On the other hand, the cheaper or to involve less technical com-
susceptible devices (affected electronic plexity than reducing sensitivity at the re-
components) are designed for insensitivity ceptors, the logical response might be to
to pulses of specific shapes and ampli- shift the definitions to Classes I/II. If the
tudes. The initial step was to list and clas- scenario is inverted, with the shielding
sify the pulses encountered within vehicu- of potential susceptible devices as the
lar electrical systems (Table 1). Special- cheaper and simpler solution, then a
purpose pulse generators can be used move to Classes III/IV is warranted.
to generate the test patterns defined in Because numerous wires are combined
Table 1 as a basis for assessing resistance within a single wiring harness, each indi-
to these interference waves in susceptible vidual conductor is potentially susceptible
devices. Both the test pulses and the test to inductive and capacitive interference.
technology are codified in standards Although reduced in intensity, the resulting
(DIN 40 839, Section 1; ISO 7637, Section 1) voltage pulses in adjacent supply lines can
including definitions of the measurement then appear as spurious signals at the in-
technology for evaluating emissions of put ports and control outputs in neighbor-
pulse-pattern interference. Classifications ing systems. The test technology for simu-
based on pulse amplitude levels facilitate lating crosstalk interference within wiring
effective definition of radio-interference harnesses (as defined in DIN 40 839,
sources and the susceptible devices (inter- Section 3 and ISO 7637, Section 3) uses
ference receptors) within each vehicle. a standardized substitute wiring layout
It would thus be possible to specify Class II (capacitive clip) with a defined wiring
for all of the radio-interference sources capacitance. Test pulses are fed into this
1 Mutual interference within power supply
Test pulses as defined in DIN 40 839, Section 1 Max. pulse amplitude classes
Pulse pattern This is Internal Pulse I II III IY
caused by resistance duration
1 Shutdown 10 Ω 2 ms –25 V –50 V –75 V –100 V
of inductive
consumers
2 Shutdown 10 Ω 50 µs +25 V +50 V +75 V +100 V
of motorized
consumers
3a Steep over- 50 Ω 0.1 µs –40 V –75 V –110 V –150 V
voltages
3 +25 V +50 V +75 V +100 V
b
Table 1
4 Voltage 10 mΩ to 20 s 12 V 12 V 12 V 12 V 1) “Load dump”, i.e.
curve dur- –3 V –5 V –6 V –7 V
when the alternator
ing starting
is feeding high
operation
current to the
5 Alternator 1Ω to 400 ms +35 V +50 V +80 V +120 V battery and the
load dump 1) connection to the
battery is suddenly
interrupted.
layout and through the specimen’s wiring Effects on mobile radio reception
harness to produce overcoupling in the of high-frequency signal feedback
signal and control lines. The effects of in the vehicle electrical system
low-frequency oscillations within vehicu- Along with the pulses and other interfer-
lar electrical systems can be simulated by ence signals described above, high-fre-
producing the desired signals with a signal quency signals constitute yet another
generator and projecting these into the source of undesired interference within
wiring harness through a current clamp. the vehicle electrical system. These signals
This process reflects the procedure de- can stem from periodic switching opera-
scribed above by serving as the basis for tions, as found in high-intensity ignition
the correct balance between the amplitude systems, at the commutators in DC motors
of radiated interference pulses and the re- and from the CPU clock signals that are
sistance to interference of the potentially generated in microprocessor-equipped
susceptible devices (receptors). control units. The interference these sig-
5 Voltage amplitude
dB
a b
Fig. 5 Tr
a As a function
T
of time
b As a function Ti fg1
of frequency A0 H
dB
20
Dek
T Period duration
fg2
Tr Rise time
Ti Pulse duration dB
Amplitude
Amplitude
f0 = T–1 Fundamental
40
Dek
wave
UAE0685-1E
fg Fundamental
frequencies f0 fmin1 fmin2
fmin Periodic minima Time Frequency
H Envelope curve
6 Interference-signal spectrum
dB dB
a b
Radio-interference voltage level
Radio-interference voltage level
Fig. 6
UAE0241-1E
a Wide-band
interference MHz MHz
b Narrow-band Frequency Frequency
interference
nals can induce in mobile communications In the case of radio interference this
receiver equipment may impair reception equipment is in the form of a selective
or even render it impossible. laboratory receiver or spectral analysis
unit. Similar to a radio tuner therefore, the
Spectrum tester thus measures the signal amplitude
Pulsation within on-board electrical sys- only within a specific narrow frequency
tems is usually based on the analysis of band (receiver bandwidth). The test pro-
current and voltage progressions as func- gresses through the entire relevant fre-
tion of time (Fig. 5a). Amplitudes at specific quency range using a procedure analogous
frequencies are generally viewed as the to the station search function in standard
prime criterion for evaluating interference radios: while maintaining the initial band-
signals affecting radio reception (Fig. 5b). width, the laboratory receiver proceeds
Under standard conditions, the interfer- through a range of frequencies, either at
ence signals encountered in the automotive graduated intervals or in a continuous
environment are rarely isolated sinusoidal progression.
waves with single amplitudes recurring at Interference signals recurring at fre-
consistent frequencies. Much more com- quencies lower than the test bandwidth
mon are superimpositions consisting of produce the continuous signal pattern
numerous oscillation components with indicative of wide-band interference.
a variety of amplitudes and frequencies. If the frequency is higher than the test
The “spectrum” for an interference signal bandwidth, the test monitor will also pick
is a portrayal of amplitude versus fre- up gaps in the spectrum, and the result
quency designed to facilitate evaluation of will be the line spectrum that indicates
interference potential in individual wave- narrow-band interference.
bands (Figs. 6a and 6b). Table 4 (refer to Electric motors are a typical source of
section "Interference-suppression catego- wide-band interference. Commutation
ries") lists the most significant wavebands frequencies in electric motors are located
for automotive applications. around just a few 100 Hz, with the exact
Interference test technology divides figure depending on the motor’s number
interference signals into wide-band and of poles and its speed. At a test bandwidth
narrow-band interference: a spectrum with of 120 kHz (bandwidth corresponding to an
a continuous, progressive curve (Fig. 6a) is FM radio tuner receiver bandwidth) this
indicative of wide-band interference, and produces a continuous spectrum. At the
identifies the corresponding radio-inter- same test bandwidth, a 2 MHz cycle fre-
ference source as a wide-band interferer. quency (of the kind that might be encoun-
A contrasting pattern composed of isolated tered in a microprocessor-equipped con-
spikes forming a so-called line spectrum trol unit) will generate a completely differ-
indicates narrow-band interference stem- ent spectrum, this time with the spiked line
ming from narrow-band emitters (Fig. 6b). spectrum typical of narrow-band interfer-
Initially, the classifications in either cat- ence (interference signal recurs at fre-
egory are arbitrary: among the factors that quency greater than test bandwidth).
determine whether interference is emanat- While all electric motors – fans, wind-
ing from a “wide-band” or “narrow-band” shield wipers, servo units and fuel-supply
radio-interference source are the recep- pump, etc. – join the alternator as typical
tion properties of the potentially suscepti- emitters of wide-band interference, yet an-
ble devices as reflected in the characteris- other potential emitter is the high-tension
tics of the test equipment employed to ignition circuit. In addition, low-frequency
monitor the emissions. clock signals stemming from devices such
as switching elements can produce wide-
band interference. The list of narrow-band Measuring incident interference

interference emitters includes the micro- Test techniques (DIN 57 879/VDE 0879,
processors in control units as well as all Sections 2 and 3; CISPR 25) are stipulated
other devices using high-frequency con- for monitoring interference transmitted
trol signals. through wiring and antennas. Research on
individual components is carried out using
Interference signals picked up by radio
receivers may also be of the so-called con-
Schematic diagram of simulated automotive
ducted interference type, that is they may 7 electrical circuit
propagate through wiring and conductors
5 H
(such as the radio’s power supply) or they
A P
can enter the receiver’s I/O ports via in-
ductive and capacitive radiation in the wir-
ing harness. However, interference usually
0,1 F
Fig. 7 enters through the antenna, either through
Terminals: coupling in the antenna cable, or because
P-B Test
the antenna picks up the electromagnetic
A-B Power supply M
fields generated by radio-interference 1 F S
M-B Radio-interference
monitor sources. One particularly effective trans-
S Switch mission antenna is the wiring harness.
B Reference ground Other factors influencing the amplitude 50
UAE0686-1Y
(metal sheet, of the interference signals picked-up by
shielding for
the receiver are the vehicle-body structure
simulated vehicle
and the type and location of the antenna. B B
electrical system)
2 Interference-suppression levels: permissible radio-interference voltage limits for individual frequency ranges in
dBµV for wide-band (B) and narrow-band (S) interference as defined in CISPR 25 (DIN/VDE 0879-2)
Interference- Interference-suppression levels
suppression 0.15 to 0.53 to 5.9 to 30 to 70 to
levels 0.3 MHz (LW) 2.0 MHz (MW) 6.2 MHz (SW) 54 MHz 108 MHz (VHF)
B S B S B S B S B S
1 100 90 82 66 64 57 64 52 48 42
2 90 80 74 58 58 51 58 46 42 36
3 80 70 66 50 52 45 52 40 36 30
4 70 60 58 42 46 39 46 34 30 24
Table 2 5 60 50 50 34 40 33 40 28 24 18
3 Permissible radio interference-field strength in dBµV/m for interference-suppression levels in individual frequency ranges
according to DIN/VDE 0879, Section 2, or CISPR 25 for wide-band interferers, measured with quasi-peak detector (B), and
narrow-band interference, measured with peak detector (S).
Interfer- Interference-field strength level
ence- 0.15 to 0.53 to 5.9 to 30 to 68 to 76 to 142 to 380 to 820 to
sup- 0.3 MHz 2.0 MHz 6.2 MHz 54 MHz 87 MHz 108 MHz 175 MHz 512 MHz 960 MHz
pression (LW) (MW) (SW) (VHF)
level B S B S B S B S B S B S B S B S B S
1 83 61 70 50 47 46 47 46 36 36 36 42 36 36 43 43 49 49
2 73 51 62 42 41 40 41 40 30 30 30 36 30 30 37 37 43 43
3 63 41 54 34 35 34 35 34 24 24 24 30 24 24 31 31 37 37
4 53 31 46 26 29 28 29 28 18 18 18 24 18 18 25 25 31 31
Table 3 5 43 21 38 18 23 22 23 22 12 12 12 18 12 12 19 19 25 25
laboratory setups and standardized test units are usually switched on along with
setups in shielded chambers. Interference the ignition). Also of significance is the fact
is monitored with a laboratory receiver. that the radio receiver cannot distinguish
Line lengths and other geometric dimen- these spurious signals from useful signals,
sions are specified. The power supply making reception of weak transmissions
must be provided by a precisely defined impossible. This situation is reflected in
vehicle electrical system. Voltage sources the definition of interference-suppression
are also precisely defined, with laboratory categories. In a given interference-sup-
specimens being powered from simulated pression category, narrow-band interfer-
vehicle electrical systems (Fig. 7). ences are assigned lower permissible in-
terference levels than wide-band interfer-
Interference-suppression categories ences.
Similar to pulsation in the vehicle electri-
cal system, narrow- and wide-band inter- Because individual vehicle configuration
ference factors are classified by interfer- also has a substantial effect on the quality
ence-suppression category in order to of broadcast reception, the interference-
facilitate selection and design for specific suppression data derived from laboratory
applications. Within this classification testing must be confirmed in a practical
system, demands on sporadically active automotive environment. The test involves
radio-interference sources are less strin- measuring antenna voltage at the end of
gent than the requirements placed on the antenna cable to which the radio re-
components in continuous operation, ceiver will subsequently be connected.
such as the alternator. The permissible CISPR 25 also prescribes limits for radio
CISPR 25 (DIN/VDE 0879-2) radio interfer- interference voltage as determined using
ence levels are listed in Table 2. Table 3 in- this procedure (Table 4). It contains volt-
dicates the permissible interference-field age levels defined to reflect the unfavor-
strengths for radiation measurements with able reception conditions as encountered
antennas. in motor vehicles, where the strengths of
useful signals not only fail to exceed just
Narrow-band interference of the kind a few mV, but also suffer from substantial
generated by CPU clock signals in control fluctuation owing to vehicle motion and
units poses an especially acute problem the multiple reception paths generated
for radio reception. These interference by signal reflections. Table 4
signals are always present (the control QP-B Quasi-peak
detector relays
aural impression
4 Limits defined for permissible radio interference voltage at vehicle antenna in dBµV produced by
Frequency range Frequency Continuous wide-band Sporadic wide-band Narrow-band interference
interference interference interference B Wide-band
MHz QP-B b QP-B b S interference emitter

with peak detector
LW 0.14 to 0.30 9 22 15 28 6
indicates maximum
MW 0.53 to 2.0 6 19 15 28 0
level
SW 5.9 to 6.2 6 19 6 19 0 S Narrow-band
2-way transceivers 30 to 54 6 (15*) 28 15 28 0 interference emitter
2-way transceivers 70 to 87 6 (15*) 28 15 28 0 with peak detector
VHF 87 to 108 6 (15*) 28 15 28 6 indicates maximum

level
2-way transceivers 144 to 172 6 (15*) 28 15 28 0
* Limit values for
C-network car phone 420 to 512 6 (15*) 28 15 28 0
high-voltage
D-network car phone 800 to 1000 6 (15*) 28 15 28 0 ignition
504 Electromagnetic compatibility | EMC between the vehicle and its surroundings
Interference from electrostatic charges EMC between the vehicle

ESD, or electrostatic discharge, can dam- and its surroundings
age or even destroy electronic compo-
nents. The voltages produced in the event Since early 1996, an official regulation
of this discharging can be anywhere up to has placed mandatory limits on radiation
several volts, which in turn also generate by the vehicle of interference affecting
very high, pulsed currents. For this reason, stationary radio reception, while at the
suitable measures are required to prevent same time defining suppression require-
the destructive effect or, better still, the ments during exposure to external electro-
discharges themselves. Electronic compo- magnetic fields. This ordinance (European
nents that come into contact with vehicle Directive 95/54/EC) superseded an earlier
passengers are particularly endangered. directive of more limited scope, which
only limited interference emissions while
ISO TR 10605 defines test procedures for defining the homologation certification
examining the effects of electrostatic dis- procedure governing EMC in motor vehi-
charge and the immunity to interference cles.
that components exhibit both in the labo-
ratory and when installed in the vehicle. Emitted interference
The test is conducted with a suitable ESD Electromagnetic emissions radiated by mo-
test-pulse generator, usually in the shape tor vehicles must remain within the limits
of a pistol. The generator produces the for narrow and wide-band signals (Fig. 1)
high-voltage pulses that are then coupled to ensure that they do not interfere with
into the test specimen. radio and television broadcasts or private
transmissions. These limits are specified in
the Directive 95/54/EC cited above as well
as in VDE 0879, Section 1 and the CISPR 12
standard. These regulations contain de-
tailed descriptions of the test procedure,
which employs antennas set up at defined
distances from the vehicle (10 meters,
3 meters).
The specifics of the test procedure are
1 Wide- and narrow-band limit values for vehicles described in the regulations quoted.
as monitored at a distance of 10 m
In practice, it is the ignition system that is
usually responsible for the greatest amount
dB V/m of interference radiation. However, since a
comprehensive range of measures have
Interference field strength
been taken to ensure good radio reception

Wide-band
40 limit value in the vehicle, the emitted interference is
limited to such an extent that compliance
with the legal limit values is usually com-
Narrow-band fortably achieved.
30 limit value
Incident radiation
As a vehicle travels through the high-in-
tensity electromagnetic field (near field)
UAE0681-3E
20
30 75 400 1,000 immediately surrounding a high-intensity
Frequency MHZ emissions source, radiation penetrates
through gaps and apertures in the vehicle
Electromagnetic compatibility | EMC between the vehicle and its surroundings 505
body and interferes with on-board electri- The anechoic chamber is used for investi-
cal systems. The intensity of this effect (in- gating the overall performance of vehicle
cident radiation) varies according to com- electrical systems under real-world oper-
ponent locations, body configuration and ating conditions. The Bosch anechoic
the wiring harness. chamber (Fig. 2) is equipped with devices
capable of generating high-frequency
Vehicle measurements fields within a frequency range extending
At one time the procedure for verifying that from 10 kHz to 18 GHz, with a maximum
various electronic systems would continue field strength of Emax = 200 V/m. Since
to provide satisfactory performance under these field strengths represent a health
exposure to high-intensity fields involved hazard, the test vehicle is operated via
driving the motor vehicle to a number of remote control from within a shielded
different broadcast transmitter locations. control room, while testing is monitored
Now, special test chambers are available by video cameras. The chamber is screened
for this purpose. with metallic sheeting, and non-conduc-
These chambers must be surrounded tive materials (wood and plastic) line the
(shielded) with a metallic casing to prevent interior, as metallic substances could
the electromagnetic field generated during foster spurious readings. To prevent reflec-
testing from radiating to the outside. In ad- tions and inhibit the formation of standing
dition, the chambers must be equipped with waves, walls and ceiling are covered by pyr-
absorption materials to inhibit formation amid-shaped absorption elements made of
of standing waves (nodes and antinodes), graphite-filled polyurethane foam.
which would lead to major variations in Vehicle testing proceeds on a chassis
field strength between the various measur- dynamometer (rolling road) capable of
ing points. accommodating simulated speeds of up to
2 Measuring resistance to incident radiation of vehicle electrical systems in the EMC anechoic chamber
UAE0689Y
506 Electromagnetic compatibility | EMC between the vehicle and its surroundings
200 km/h. A fan is also present to direct up a different procedure is used for this
to 40,000 m3/h (corresponding to a wind range. The alternate procedure relies on
speed of 80 km/h) of air over the vehicle. antennas to project electromagnetic fields
Among the advantages that distinguish into standardized bench setups.
indoor testing from the old open-air tests The specifics of the methods outlined
near transmitters is that the former method below are defined in DIN 40 839, Section 4,
allows latitude for considerable variations and ISO 11 452, Sections 1 – 7 (which also
in both frequency and field strength. This contains additional process descriptions
facilitates evaluation of vehicular resis- detailing less widespread methods).
tance to incident radiation under a wide All these methods furnish a precise pic-
range of conditions, and not at just a few ture of a system’s resistance to incident
available points. The ability to gradually radiation. This can then be used to imple-
increase field strengths right up to each ment improvements while the develop-
electronic device’s operational limits fur- ment phase is still in progress. Thanks to
nishes information on safety margins. this inestimable advantage it is now im-
Specifics defining the test procedures possible to imagine development projects
used in determining radiation resistance without these laboratory measurements
in the overall vehicle are described in as flanking measures.
DIN 40 839, Section 4. ISO 11452, Sec- It must be taken into consideration that
tions 1 – 4 contain similar descriptions as the design of the electrical system in itself
elements within a more extensive compila- is not the only important factor. Because
tion of special-purpose test procedures. installation conditions in the vehicle and
the routing of the wiring harness can also
Laboratory measurements exercise a decisive influence on ultimate
Although the data garnered from assess- levels of interference resistance, all earlier
ments of incident radiation on the vehicle results still need to be confirmed on a pro-
as a whole are invaluable, the disadvan- duction vehicle in the anechoic chamber.
tage of this type of testing is that it can
only be performed at an advanced stage Stripline method (Figs. 3 and 4)
in the vehicle and electronic system design The designation, “stripline”, refers to a
process. Only very limited latitude is avail- conductor in the form of a strip. This con-
able for responding to indications of inad- ductor has a length of 4.1 m and a width of
equate resistance to incident radiation un- 0.74 m. It is arranged with a clearance of
covered at this stage. 0.15 m above the conductive sheet (coun-
This explains the necessity to determine ter-electrode). A high-frequency alterna-
at an early stage how an electronic system tor serves as the source of a transverse
will operate during subsequent use in its magnetic wave generated between the
automotive environment: prompt informa- stripline and the counter-electrode, con-
tion allows effective recourse to remedial tinuing until it reaches the terminal resis-
action when needed. Various test proce- tor. The stripline dimensions have been
dures have emerged for obtaining the re- selected to minimize the likelihood of re-
quired information. flections occurring during wave propaga-
In the first of the three tests described tion, thus ensuring constant field-strength
below, interference waves propagating amplitudes relative to frequency.
through a conductor are coupled into the
wiring harness of the system under inves- A typical system setup might include a
tigation. As the suitability of these test ar- control unit, wiring harness and peri-
rangements for evaluating the frequency pheral devices (sensors and actuators).
range beyond > 400 MHz is restricted, These are set up at a height half way be-
Electromagnetic compatibility | EMC between the vehicle and its surroundings 507
tween the base plate and the stripline, Bulk current injection method (Fig. 5)
with the wiring harness in alignment with Bulk current injection (BCI) is an immunity
the waves’ direction of propagation. test commonly used in the automotive
and aircraft industry. As with the stripline
The field strength between the planes at method, the system being tested is ar-
a given frequency is then increased until ranged above a conductive sheet (counter-
the system malfunctions or until a maxi- electrode). A current clamp attached to the
mum specified level is reached. If the in- wiring harness injects transformer current
crements are small enough, repeating this into its individual wires. The vectorial sum
process at progressively altered frequen- of these currents corresponds to the total
cies will produce a detailed diagram por- current radiated by the clamp. While test-
traying resistance to incident radiation as ing using the stripline method proceeds
a function of frequency (Fig. 4). through a range of increasing field
strengths, BCI relies on progressively
3 Stripline method higher current flows as it continues until
the system malfunctions or until current
3 reaches the specified terminal value.
1
TEM cell (Fig. 6)
As with the stripline method, a transverse
G
electromagnetic field (TEM) is generated
Fig. 3
here between a strip conductor and a 1 High-frequency
5
counter-electrode. The TEM counter-elec- alternator
7 6 trode, however, is an enclosed housing 2 Resistor
instead of a metal plate. This leads to an- 3 Stripline
other distinguishing feature: TEM test 4 Counter-electrode

2 2
(conductive sheet
benches do not need to be set up within
or cell)
specially shielded chambers.
UAE0246-1Y
5 System under test

4 6 Wiring harness
7 Peripherals
(sensors, actuators)
4 Resistance to incident radiation 5 BCI Method
Fig. 4
Determined using
-1 1 2
V·m G stripline, BCI or
TEM cell method
Resistance to incident radiation
Fig. 5
7
1 High-frequency
4
alternator
2 Resistor
6 5 3 Counter-electrode
(conductive sheet
or cell)
4 System under test
UAE0249-1E
UAE0247-1Y
5 Wiring harness
3
MHz 6 Peripherals
Frequency (sensors, actuators)
7 Current clamp
508 Electromagnetic compatibility | Guarantee of immunity and interference suppression
A further difference lies in the fact that Guarantee of immunity and

only the test specimen per se (for example: interference suppression
a control unit) is subject to electromag-
netic exposure. The peripheral equipment As early as the planning and conception
is located outside the TEM cell and is con- phases, EMC requirements for immunity
to and suppression of radio interference
wiring harness arranged at right angles have to be considered as part of every
to the direction of propagation of the development project focusing on any elec-
electromagnetic waves. tronic system or component. The finished
The measurement procedure is consis- products must incorporate EMC measures
tent with that of the stripline method. as an integral design element.
Here, too, the field strength is increased
until the system malfunctions or until a EMC in the electronic control unit (ECU)
maximum specified level is reached. The first EMC requirement reflected in
any ECU is to select microprocessor cycle
Incident radiation via antenna frequencies that are as low as possible
Once again similarities with the stripline while simultaneously making every effort
method are encountered, as the test speci- to keep the signal transit curves as flat as
men, including control unit, wiring har- possible. EMC characteristics join stan-
ness and peripherals, is set up on a base dard operational aspects as an essential
plate. The wiring harness is laid out at criterion governing the selection of com-
a defined distance from this base plate. ponents (integrated circuits). Whereas the
An antenna located at a specified distance components’ resistance to interference
from this assembly generates an electro- must be maximal, its interference emis-
magnetic field for projection into the en- sions should be minimal. On printed-
tire test setup. Again, field strength is circuit boards designed for optimal EMC
raised until the system malfunctions or performance, all components with pro-
until field strength reaches the maximum nounced sensitivity to interference, and
value specified for testing. all of those with the potential to cause it,
6 TEM-cell test setup
3
6 5
Fig. 6 1
1 High-frequency
G
alternator
2 Resistor
3 Stripline
5
4 Counter-electrode
2 2
(conductive sheet
UAE0248-1Y
or cell)
5 System under test
6 Peripherals 4 4
(sensors, actuators)
Electromagnetic compatibility | Guarantee of immunity and interference suppression 509
are decoupled from the connected wiring High-voltage ignition

harness. This is done by locating the com- The ignition system’s high-voltage circuit
ponents as far as possible from the plug represents a potential source of substan-
connections. tial interference to radio reception.
Suppression components, usually ca- One standard solution is to use spark plugs
pacitors designed for high-frequency ap- and cable sockets with noise-suppression
plications, place the required restrictions resistors. While the resistor element is
on potentially disruptive interference. usually located at the end of the ignition
These interference-suppression capacitors cables, in modern (multiple-coil) systems
are located either directly within the inte- it can be integrated within the dedicated
grated circuits or in the plug-connector ignition coils plugged directly onto the
area. The suppression elements in the spark plugs for the individual cylinders.
plug-connector area combine with hous- In using resistors, the object is to find a
ings featuring optimal conductive proper- balance between the demand for ignition
ties (shielded casing) to prevent high-fre- energy and the required level of interfer-
quency radiation in the immediate vicinity ence suppression.
from reaching the inside of the ECU. This
ensures that signals produced outside the Retrofit interference suppression
unit do not cause interference inside the As made clear in the above sections,
unit. On the other side of the coin, high- EMC design considerations must be care-
frequency signals produced inside the unit fully aligned and balanced against opera-
do not cause undesirable interference in tional requirements. Subsequent remedial
the surrounding environment. measures superimposed on the original
design at a later date usually entail sub-
Electric motors and other stantial effort, and should be avoided in
electromechanical components production vehicles.
Standard practice in the design of electric Should the interference-suppression
motors reflects the procedures used for measures incorporated in electrical com-
ECUs and sensors by incorporating inter- ponents prove inadequate for dealing with
ference suppression in the design process the conditions encountered in specific ap-
right from the initial development stages. plications (in special-purpose emergency
A sample problem is the interference pro- vehicles, etc.), supplementary action can
duced by the brush arcing during current be implemented to deal with the problem.
reversal in motors with commutator Options include installation of filters as
control. This can severely disrupt radio well as supplementary shielding for com-
reception. Suitable suppression elements ponents and wiring.
(capacitors and throttles) limit the inter- These options should be exercised with
ference. Motors are designed with optimal great care, as field modifications to elec-
operation of these interference suppres- tronic components can lead to operational
sors in mind. malfunctions.
The voltage pulses produced during
switching operations in electromagnetic
actuators are kept to an acceptable level
by suitable elements, such as quench
resistors, in the circuit.
510 Fault diagnostics Monitoring during vehicle operation (on-board diagnosis)
Fault diagnostics
The rise in the sheer amount of electronics Monitoring during
in the automobile, the use of software to con-
trol the vehicle, and the increased complexity
vehicle operation
of modern fuel-injection systems place (on-board diagnosis)
high demands on the diagnostic concept, Overview
monitoring during vehicle operation ECU-integrated diagnostics belong to the ba-
(on-board diagnosis), and workshop diag- sic scope of electronic engine-management
nostics (Fig. 1). The workshop diagnostics systems. Besides a self-test of the control unit,
is based on a guided troubleshooting proce- input and output signals, and control-unit
dure that links the many possibilities of on- intercommunication are monitored.
board and offboard test procedures and test On-board diagnosis of an electronic system
equipment. As emission-control legislation is the capability of a control unit to interpret
becomes more and more stringent and con- and perform self-monitoring using “software
tinuous monitoring is now called for, law- intelligence”, i.e. detect, store, and diagnosti-
makers have now acknowledged on-board cally interpret errors and faults. On-board
diagnosis as an aid to monitoring exhaust- diagnosis runs without the use of any addi-
gas emissions, and have produced manufac- tional equipment.
turer-independent standardization. This Monitoring algorithms check input and
additional system is termed the on-board output signals during vehicle operation, and
diagnostic system. check the entire system and all its functions
for malfunctions and disturbances. Any er-
rors or faults detected are stored in the con-
trol-unit fault memory. Stored fault informa-
tion can be read out via a serial interface.
1 Diagnostic system
Diagnostic
tester
Offboard
tester
Te
tl
s
ine
s
æ UWT0104E
d d
ar ar
fbo bo
Of On Control units

Fault diagnostics Monitoring during vehicle operation (on-board diagnosis) 511
Input-signal monitoring Output-signal monitoring

Sensors, plug connectors, and connecting Actuators triggered by a control unit via out-
lines (signal path) to the control unit (Fig. 2) put stages (Fig. 2) are monitored. The moni-
are monitored by evaluating the input signal. toring functions detects line breaks and
This monitoring strategy is capable of detect- short-circuits in addition to actuator faults.
ing sensor errors, short-circuits in the bat- The following methods are applied:
tery-power circuit UBatt and vehicle-ground 쐌 Monitoring an output signal by the output
circuit, and line breaks. The following stage. The electric circuit is monitored for
methods are applied: short-circuits to battery voltage UBatt, to
쐌 Monitoring sensor supply voltage (if vehicle ground, and for open circuit.
applicable). 쐌 Impacts on the system by the actuator
쐌 Monitoring detected values for permissible are detected directly or indirectly by a
value ranges (e.g. 0.5....4.5 V). function or plausibility monitor. System
쐌 If additional information is available, a actuators, e.g. exhaust-gas recirculation
plausibility check is conducted using the valves, throttle valves, or swirl flaps, are
detected value (e.g. comparison of crank- monitored indirectly via closed-control
shaft speed and camshaft speed). loops (e.g. continuous control variance),
쐌 Critical sensors (e.g. pedal-travel sensor) and also partly by means of position sen-
are fitted in redundant configuration, sors (e.g. position of turbine geometry in
which means that their signals can be the exhaust-gas turbocharger).
directly compared with each other.
2 System chart of an electronic system (example: common rail)
Sensors and setpoint generators ECU Actuators
Pedal-travel sensor Sensor evaluation

with low-idle switch Injectors
and kickdown switch
Wheel-speed sensors Rail-pressure control valve

(crankshaft, camshaft) Signal processing High-pressure pump
Exhaust-gas recirculation
actuator
Pressure sensors
(charge-air pressure, Boost-pressure actuator
rail pressure)
Throttle-valve actuator
Fault diagnosis
Temperature sensors Additional output stages
(charge-air, exhaust-gas,
and engine temperature)
Communication
K ISO interface
Air-mass meter Output stages (e.g. diagnostics)
L
Glow-plug control unit
æ SAE0750-2E
Switch inputs (clutch

switch, terminal 15, Diagnosis lamp
vehicle-speed controller)
CAN CAN interface
512 Fault diagnostics Monitoring during vehicle operation (on-board diagnosis)
Monitoring internal ECU functions Error handling

Monitoring functions are implemented in Error detection
control-unit hardware (e.g. “intelligent” out- A signal path is categorized at finally defective
put-stage modules) and software to ensure if an error occurs over a definite period of
that the control unit functions correctly at all time. Until the defect is categorized, the sys-
times. The monitoring functions check each tem uses the last valid value detected. When
of the control-unit components (e.g. micro- the defect is categorized, a standby function is
controller, flash EPROM, RAM). triggered (e.g. engine-temperature substitute
Many tests are conducted immediately after value T = 90°C).
startup. Other monitoring functions are Most errors can be rectified or detected
performed during normal operation and re- as intact during vehicle operation, provided
peated at regular intervals in order to detect the signal path remains intact for a definite
component failure during operation. Test period of time.
runs that require intensive CPU capacity, or
that cannot be performed during vehicle op- Fault storage
eration for other reasons, are conducted in af- Each fault is stored as a fault code in the
ter-run more when the engine is switched off. non-volatile area of the data memory.
This method ensures that the other functions The fault code also describes the fault type
are not interfered with. In the common-rail (e.g. short-circuit, line break, plausibility,
system for diesel engines, functions such as value range exceeded). Each fault-code input
the injector switchoff paths are tested during is accompanied by additional information,
engine runup or after-run. With a spark-igni- e.g. the operating and environmental con-
tion engine, functions such as the flash ditions (freeze frame) at the time of fault
EPROM are tested in engine after-run. occurrence (e.g. engine speed, engine temper-
ature).
Monitoring ECU communication
As a rule, communication with other ECUs Limp-home function
takes place over the CAN bus (Controller If a fault is detected, limp-home strategies can
Area Network). The CAN protocol contains be triggered in addition to substitute values
control mechanisms to detect malfunctions. (e.g. engine output power or speed limited).
As a result, transmission errors are even de- These strategies help to:
tectable at CAN-module level. A number of 쐌 Maintain driving safety
other checks are also performed in the ECU. 쐌 Avoid consequential damage
Since the majority of CAN messages are sent 쐌 Minimize exhaust-gas emissions
at regular intervals by the individual control
units, the failure of a CAN controller in a
control unit is detectable by testing at regular
intervals. In addition, when redundant infor-
mation is available in the ECU, the received
signals are checked in the same way as all in-
put signals.
Fault diagnostics On-board diagnosis system for passenger cars and light-duty trucks 513
On-board diagnosis system used to monitor emission-related compo-

nents or if they can affect the diagnosis re-
for passenger cars and sults.
light-duty trucks Normally, the diagnostic functions for all
The engine system and its components must components and systems under surveillance
be constantly monitored in order to comply must run at least once during the exhaust-gas
with exhaust-gas emission limits specified by test cycle (e.g. FTP 75). A further stipulation
law in everyday driving situations. For this is that all diagnostic functions must run with
reason, regulations have come into force to sufficient frequency during daily driving
monitor exhaust-gas systems and compo- mode. For many monitoring functions, the
nents, e.g. in California. This has standard- law defines a monitoring frequency (In
ized and expanded manufacturer-specific on- Use Monitor Performance Ratio) in daily
board diagnosis with respect to the monitor- operation starting model year 2005.
ing of emission-related components and sys- Since the introduction of OBD II, the law
tems. has been revised in several stages (updates).
The last update came into force in model year
Legislation 2004. Further updates have been announced.
OBD I (CARB)
1988 marked the coming into force of OBD I OBD (EPA)
in California, that is, the first stage of CARB Since 1994 the laws of the EPA (Environmen-
legislation (California Air Resources Board). tal Protection Agency) have been in force in
The first OBD stage makes the following re- the remaining U.S. states. The scope of these
quirements: diagnostics comply for the most part with
쐌 Monitoring emission-related electrical the CARB legislation (OBD II).
components (short-circuits, line breaks) The OBD regulations for CARB and
and storage of faults in the control-unit EPA apply to all passenger cars with up to
fault memory. 12 seats and to light-duty trucks weighing up
쐌 A Malfunction Indicator Lamp (MIL) that to 14,000 lbs (6.35 t).
alerts the driver to the malfunction.
쐌 The defective component must be dis- EOBD (EU)
played by means of on-board equipment The OBD attuned to European conditions is
(e.g. blink code using a diagnosis lamp). termed EOBD and is based on the EPA-OBD.
EOBD has been valid for all passenger cars
OBD II (CARB) and light-duty trucks equipped with gasoline
The second stage of the diagnosis legislation engines and weighing up to 3.5 t with up to
(OBD II) came into force in California in 9 seats since January 2000. Since January 2003
1994. OBD II became mandatory for diesel- the EOBD also applies to passenger cars and
engine cars with effect from 1996. In addition light-duty trucks with diesel engines.
to the scope of OBD I, system functionality
was now monitored (e.g. plausibility check of Other countries
sensor signals). A number of other countries have already
OBD II stipulates that all emission-related adopted or are planning to adopt EU or
systems and components must be monitored US-OBD legislation.
if they cause an increase in toxic exhaust-gas
emissions in the event of a malfunction (by
exceeding the OBD limits). Moreover, all
components must be monitored if they are
514 Fault diagnostics On-board diagnosis system for passenger cars and light-duty trucks
Requirements of the OBD system Communication with scan tool

The ECU must use suitable measures to mon- OBD legislation prescribes standardization of
itor all on-board systems and components the fault-memory information and access to
whose malfunction may cause a deterioration the information (connector, communication
in exhaust-gas test specifications stipulated by interface) compliant with ISO 15 031 and the
law. The Malfunction Indicator Lamp (MIL) corresponding SAE standards (Society of Au-
must alert the driver to a malfunction if the tomotive Engineers). This permits the read-
malfunction could cause an overshoot in out of the fault memory using standardized,
OBD emission limits. commercially available testers (scan tools, see
Fig. 1).
Emission limits Depending on their scope of application,
The U.S. OBD II (CARB and EPA) prescribes various communication protocols are used
thresholds that are defined relative to emis- throughout the world: The most important
sion limits. Accordingly, there are different are:
permissible OBD emission limits for the vari- 쐌 ISO 9141-2 for European passenger cars.
ous exhaust-gas categories that are applied 쐌 SAE J 1850 for U.S. passenger cars.
during vehicle certification (e.g. TIER, LEV, 쐌 ISO 14 230-4 (KWP 2000) for European
ULEV). Absolute limits apply in Europe passenger cars and commercial vehicles.
(Table 1). 쐌 SAE J 1708 for U.S. commercial vehicles.
Malfunction Indicator Lamp (MIL) These serial interfaces operate at a bit rate
The Malfunction Indicator Lamp (MIL) (baud rate) of 5 to 10 k baud. They are de-
alerts the driver that a component has mal- signed as a single-wire interface with a com-
functioned. When a malfunction is detected, mon wire for both transmission and recep-
CARB and EPA stipulate that the MIL must tion, or as a two-wire interface with a sepa-
light up no later than after two driving cycles rate data line (K-line) and initiate line
of its occurrence. Within the scope of EOBD, (L-line). Several electronic control units (such
the MIL must light up no later than in the as Motronic, ESP, or EDC, and transmission-
third driving cycle with a detected malfunc- shift control, etc.) can be combined on one
tion. diagnosis connector.
If the malfunction disappears (e.g. loose Communication between the tester and
contact), the malfunction remains entered in ECU is set up in three phases:
the fault memory for 40 trips (warmup cy- 쐌 Initiate the ECU.
cles). The MIL goes out after three fault-free 쐌 Detect and generate baud rate.
driving cycles. 쐌 Read key bytes which identify the transmis-
sion protocol.
1 OBD limits for passenger cars and light-duty trucks
Otto passenger cars Diesel passenger cars
CARB – Relative emission limits – Relative emission limits

– Mostly 1.5 times the limit of – Mostly 1.5 times the limit of
a specific exhaust-gas category a specific exhaust-gas category
EPA – Relative emission limits – Relative emission limits (U.S. Federal)
– Mostly 1.5 times the limit of – Mostly 1.5 times the limit of
a specific exhaust-gas category a specific exhaust-gas category
EOBD 2000 2005 (draft) 2003 2005 (draft)
CO: 3.2 g/km CO: 1.9 g/km CO: 3.2 g/km CO: 3.2 g/km
HC: 0.4 g/km HC: 0.3 g/km HC: 0.4 g/km HC: 0.4 g/km
NOX: 0.6 g/km NOX: 0.53 g/km NOX: 1.2 g/km NOX: 1.2 g/km
PM: 0.18 g/km PM: 0.18 g/km
Table 1
Evaluation is performed subsequently. Reading the fault information

The following functions are possible: Any workshop can use the scan tool to read
쐌 Identify the ECU out emission-relevant fault information from
쐌 Read the fault memory the ECU (Fig. 3). In this way, workshops not
쐌 Erase the fault memory franchized to a particular manufacturer are
쐌 Read the actual values also able to carry out repairs. Manufacturers
are obliged to make the required tools and
In future, the CAN bus (ISO 15 765-4) will be information available (on the internet) in
used increasingly to handle communication return for a reasonable payment, to allow
between ECUs and the tester. Starting 2008 this.
diagnostics will only be permitted over this
interface in the U.S.A. Vehicle recall
To make it easier to read out the ECU fault- If vehicles fail to comply with OBD require-
memory information, a standardized diagno- ments by law, the authorities may demand the
sis socket will be fitted at an easily accessible vehicle manufacturer to start a recall at his
place in every car (easy to reach from the own cost.
driver’s seat). The socket is used to connect
the scan tool (Fig. 2).
1 OBD system 3 Operating modes of the diagnostic tester
Service 1 (Mode 1)
MIL Read out current system actual values
ISO 15031 (e.g. engine speed and temperature).
SAE…
Sensors Service 2 (Mode 2)
Read out environment conditions (freeze frame) prevailing
Actuators Diagnostic when the fault occurred.
Systems manager
Service 3 (Mode 3)
Communication Generic
scan tool Read out fault memory The exhaust-gas-related and
confirmed fault codes are read out.
Service 4 (Mode 4)
Erase the fault code in the fault memory and reset the
æ UWT0105E
accompanying information.
OBD socket Repair
Onboard Offboard Service 5 (Mode 5)
Display measured values and thresholds of the λ oxygen
sensors.
2 Pin allocation of the OBD socket Service 6 (Mode 6)

Display the measured values of special functions
Pins 7 and 15: (e.g. exhaust-gas recirculation).
Data transmission to
DIN ISO 9141-2 Service 7 (Mode 7)
8 16
Pins 2 and 10: Read out fault memory In Service 7, fault codes that are
7 15 Data transmission to
6 14 not confirmed are read out.
SAE J 1850
5 13 Pins 1, 3, 8, 9, 11, 12, 13: Service 8 (Mode 8)
4 12 Not assigned by OBD Initiate test functions (specific to vehicle manufacturer).
3 11 Pin 4: Vehicle ground
2 10
æ UWT0106E
Pin 5: Signal ground Service 9 (Mode 9)

1 9 Pin 6: CAN high Read out vehicle information.
Pin 14: CAN low
Pin 16: Battery +
Functional requirements Temporary disabling of diagnostic functions

Overview Diagnostic capabilities may only be disabled
Just as for on-board diagnosis, all ECU input under certain conditions in order to prevent
and output signals, as well as the components false diagnosis. Examples include:
themselves, must be monitored. 쐌 Height too large.
Legislation demands the monitoring of 쐌 Low ambient temperature at engine
electrical functions (short-circuit, line switchon.
breaks), a plausibility check for sensors, 쐌 Low battery voltage.
and a function monitoring for actuators.
The pollutant concentration expected as Readiness code
the result of a component failure (empirical When the fault memory is checked, it is im-
values), and the monitoring mode partly portant to know that the diagnostic capabili-
required by law determine the type of ties ran at least once. This can be checked by
diagnostics. A simple functional test reading out the readiness code over the diag-
(black/white test) only checks system or com- nostic interface. After erasing the fault mem-
ponent operability (e.g. swirl flap opens and ory in service, the readiness codes must be
closes). The extensive functional test provides reset after checking the functions.
more detailed information about system op-
erability. As a result, the limits of adaption Diagnostic System Management (DSM)
must be monitored when monitoring adap- The diagnostic capabilities for all compo-
tive fuel-injection functions (e.g. zero deliv- nents and systems checked must normally
ery calibration for a diesel engine, lambda run in driving mode, but at least once during
adaption for a gasoline engine). the exhaust-gas test cycle (e.g. FTP 75,
Diagnostic complexity has constantly NEDC). Diagnostic System Management
increased as emission-control legislation (DSM) can change the sequence for running
has evolved. the diagnostic capabilities dynamically,
depending on driving conditions.
Switchon conditions The DSM comprises the following three
Diagnostic capabilities only run if the components (Fig. 4):
switchon conditions are satisfied.
This includes, for instance: Diagnosis Fault Path Management (DFPM)
쐌 Torque thresholds The primary role of DFPM is to store fault
쐌 Engine-temperature thresholds states that are detected in the system. In addi-
쐌 Engine-speed thresholds or limits tion to faults, it also stores other information,
such as environmental conditions (freeze
Inhibit conditions frame).
Diagnostic capabilities and engine functions
cannot always operate simultaneously.
There are inhibit conditions that prohibit
the performance of certain functions. In the
diesel system, the Hot-Film Air-Mass Meter
(HFM) can only be monitored satisfactorily
when the exhaust-gas recirculation valve is
closed. For instance, tank ventilation (evapo-
rative-emissions control system) in a gasoline
system cannot function when catalytic-con-
verter diagnosis is in operation.
Diagnostic Function SCHEDuler (DSCHED) OBD functions

DSCHED is responsible for coordinating Overview
assigned engine capabilities (MF) and diag- Whereas EOBD only contains detailed moni-
nostic capabilities (DF). It obtains informa- toring specifications for individual compo-
tion from DVAL and DFPM to carry this out. nents, the requirements in CARB OBD II are
In addition, it reports functions that require much more detailed. The list below shows
release by DSCHED to perform their readi- the current state of CARB requirements
ness, after which the current system state is for gasoline-engined and diesel-engined
checked. passenger cars. Requirements that are also
described in detail in the EOBD legislation
Diagnosis VALidator (DVAL) are marked by (E):
The DVAL (only installed in gasoline systems 쐌 Catalytic converter (E), heated catalytic
to date) uses current fault-memory entries converter
and additionally stored information to decide 쐌 Combustion (ignition) misfire
for each detected fault whether it is the actual (E, for diesel system not for EOBD)
cause, or a consequence of the fault. As a re- 쐌 Evaporation reduction system
sult, validation provides stored information (fuel-tank leak diagnosis, only for
to the diagnostic tester for use in reading out gasoline system)
the fault memory. 쐌 Secondary-air injection
In this way, diagnostic capabilities can 쐌 Fuel system
be released in any sequence. All released 쐌 Lambda oxygen sensors (E)
diagnoses and their results are evaluated sub- 쐌 Exhaust-gas recirculation
sequently. 쐌 Crankcase ventilation
쐌 Engine cooling system
쐌 Cold-start emission reduction system
(presently only for gasoline system)
쐌 Air conditioner (components)
쐌 Variable valve timing (presently only in
4 Diagnostic system management use with gasoline systems)
쐌 Direct ozone reduction system (presently
DSM only in use with gasoline systems)
DVAL
쐌 Particulate filter (particulate filter, only for
diesel system) (E)
쐌 Comprehensive components (E)
DFPM 쐌 Other emission-related
components/systems (E)
DSCHED “Other emission-related components/sys-

tems” refer to components and systems not
mentioned in this last and that may exceed
OBD emission limits, or block other diagnos-
tic functions if they malfunction.
MF DF
æ UAE0919E
MF DF
MF DF
Catalytic converter diagnosis Examples of this include zero-fuel-quantity

In the diesel system, carbon monoxide (CO) calibration, quantity-mean-value adaptation,
and unburned hydrocarbons (HC) are oxi- and the AS MOD observer function (air-sys-
dized in the oxidation-type catalytic con- tem model observer). The two last functions
verter. Work is ongoing on diagnostic capa- use information from the lambda oxygen sen-
bilities for monitoring the operation of oxi- sor as input signals. From the models, they
dation-type catalytic converters relating to calculate fluctuations between setpoint and
temperature and differential pressure. One actual quantities.
approach focuses on active secondary injec-
tion (intrusive operation). Here, heat is gen- Lambda-oxygen-sensor diagnosis
erated in the oxidation-type catalytic con- Modern diesel systems are fitted with broad-
verter by an exothermic HC reaction. The band oxygen sensors. They require a different
temperature is measured and compared with diagnostic procedure than two-stage sensors
calculated model values. These are used to since their settings may deviate from λ = 1.
derive the functionality of the catalytic con- They are monitored electrically (short-circuit,
verter. line interruption) and for plausibility. The
Equally, work is ongoing on monitoring heater element of the sensor heater is tested
functions for the storage and regeneration ca- electrically and for permanent governor
pabilities of NOx accumulator-type catalytic deviation.
converters that will also be installed in the
diesel system in future. The monitoring func- Exhaust-gas recirculation system diagnosis
tions run based on loading and regeneration In the exhaust-gas recirculation system, the
models, and the measured regeneration dura- EGR valve and – if fitted – the exhaust-gas
tion. This requires the use of lambda or NOx cooler are monitored.
sensors. The exhaust-gas recirculation valve is mon-
itored for its electrical and functional oper-
Combustion-miss detection ability. Functional monitoring is performed
Incorrect fuel injection or loss of compres- by air-mass regulators and position con-
sion result in impaired combustion, and thus trollers. They check for permanent control
to changes in emission values. The misfire de- variances.
tector evaluates the time expired (segment If the exhaust-gas recirculation system
time) from one combustion cycle to the next has a cooler fitted, its function must also be
for each cylinder. This time is derived from monitored, i.e. an additional temperature
the speed-sensor signal. A segment time that measurement takes place downstream of the
is longer than for the other cylinders indicates cooler. The temperature measured is com-
a misfire or loss of compression. pared with a setpoint value calculated from
In the diesel system, diagnosis of combus- a model. If a fault occurs, it is detected by
tion misses is only required and performed measuring the deviation between setpoint
when the engine is at idle speed. and actual values.
Fuel system diagnosis Crankcase ventilation diagnosis

In the common-rail system, a fuel-system di- Faults in crankcase ventilation are detected by
agnosis includes the electrical monitoring of the air-mass sensor, depending on the system.
injectors and rail-pressure control (high-pres- The legislation requires no monitoring if the
sure control). In the Unit Injector System, it crankcase ventilation has a “rugged” design.
also includes monitoring of the injector
switching time. Special functions of the fuel-
injection system that increase injected-fuel-
quantity precision are also monitored.
Engine cooling system diagnosis Sensors monitor the following faults (Fig. 5):
The cooling system comprises a thermostat 쐌 Electrical faults, i.e. short-circuits and line
and a coolant-termperature sensor. If the breaks (signal-range check).
thermostat is defective, for instance, the en- 쐌 Range faults (out-of-range check), i.e. un-
gine temperature can only rise slowly and, dercutting or exceeding voltage limits de-
consequently, the exhaust emission rates may fined by the physical measurement range of
increase. The diagnostic function for the ther- a sensor.
mostat uses the coolant-termperature sensor 쐌 Plausibility faults (rationality check); these
to check that a nominal temperature has been are faults that are inherent in the compo-
reached. A temperature model is also used for nents themselves (e.g. drift), or which may
monitoring. be caused by shunts, for instance. Monitor-
The coolant-termperature sensor is moni- ing is carried out by a plausibility check on
tored to ensure that a minimum temperature the sensor signals, either by using a model,
has been reached in addition to monitoring or directly using other sensors.
for electrical faults by means of a dynamic
plausibility function. Dynamic plausibility is Actuators must be monitored for electrical
performed as the engine cools down. These faults and – if technically possible – also for
functions can monitor the sensor for “stick- function. Functional monitoring means that,
ing” in both low and high temperature when a control command (setpoint value) is
ranges. given, it is monitored by observing or mea-
suring (e.g. by a position sensor) the system
Air-conditioner diagnosis reaction (actual value) in a suitable way by
The engine can be operated at a different using information from the system.
operating point in order to cover the air- Besides all output stages, the following
conditioner’s electrical load requirements. actuators are monitored:
The required diagnosis must therefore 쐌 throttle valve
monitor all electronic components in the 쐌 exhaust-gas recirculation valve
air conditioner that may cause an increase 쐌 variable turbine geometry of the
in emissions if they malfunction. exhaust-gas turbocharger
쐌 swirl flap
Particulate-filter diagnosis 쐌 glow plugs
The particulate filter is currently monitored
for filter breakage, removal, or blockage.
A differential-pressure sensor is used to
measure differential pressure (exhaust-gas
backpressure downstream and upstream
of the filter) at a specific volumetric flow. Fig. 5
The measured value can be used to decide 1 Upper threshold
for signal-range
whether the filter is defective.
check
2 Upper threshold
Comprehensive components for out-of-range
On-board diagnosis legislation requires that 5 Sensor monitoring
check
all sensors (e.g. air-mass meter, wheel-speed V
1
3 Lower threshold
sensors, temperature sensors) and actuators 2 for out-of-range
(e.g. throttle valve, high-pressure pump, glow check
5 4 Lower threshold
æ UWT0107E
plugs) must be monitored if they either have

Voltage
for signal-range
an impact on emissions, or are used to moni- check
3
tor other components or systems (and conse- 4 5 Plausibility range
quently, may block other diagnoses). rationality check
520 Fault diagnostics On-board diagnosis system for heavy-duty trucks
On-board diagnosis system EOBD for trucks and buses > 3.5 t
European OBD legislation provides for
for heavy-duty trucks a two-stage introduction. Stage 1 (2005)
In Europe and the U.S.A., there exist draft requires monitoring:
laws that have not yet been adopted; they are 쐌 of the fuel-injection system for closed
based closely on legislation for passenger cars. electrical circuit and total failure.
쐌 of emission-related engine components
Legislation or systems for compliance with OBD
In the EU, there are plans to introduce new emission limits (Table 1).
type approvals in October 2005 (coinciding 쐌 of the exhaust-gas treatment system for
with Euro 4 emission-control legislation). major functional faults (e.g. damaged
With effect from October 2006 an OBD sys- catalytic converter, urea deficit in the
tem will become obligatory for every com- SCR system).
mercial vehicle. In the U.S.A., the draft of the
Californian CARB provides for the introduc- Stage 2 (2008) requires:
tion of an OBD system for Model Year (MY) 쐌 Monitoring of the exhaust-gas treatment
2007. It is probable that EPA (U.S. Federal) system for emission limits.
will also follow with a draft in 2004 for subse- 쐌 The OBD emission limits must be adapted
quent introduction in MY 2007. Besides that, to the prevailing state of the art (availabil-
there are initiatives to promote worldwide ity of exhaust-gas sensors).
harmonization (World Wide Harmonized
(WWH) OBD). However, this is not expected Protocols for scan-tool communication
until 2012. Japan is planning to introduce an over CAN have been approved using either
OBD system in 2005. ISO 15 765 or SAE J1939.
CARB OBD for HD trucks > 14,000 lb.

(6.35 t)
The present draft law is very close to passen-
ger-car legislation in its function require-
ments, and also provides for a two-stage
introduction:
쐌 MY 2007: Monitoring for functional faults.
쐌 MY 2010: Monitoring for OBD emission
limits (Table 1).
1 OBD emission limits for heavy-duty trucks (draft) The main changes compared with present
passenger-car legislation are as follows:
CARB 2007 2010
– Functional check – Relative limit 쐌 Erasing the OBD fault memory by scan
no limits – 1.5 times the value tool is no longer possible. This is only pos-
of each exhaust-gas
category
sible by self-healing (e.g. after repair).
– Exception: 쐌 SAE J1939 has also been approved as an al-
catalytic converter, ternative to CAN diagnostic communica-
factor 1.75
tion to ISO 15 765 (as for passenger cars).
EPA – to be defined – to be defined
EU 2005 2008
– Absolute limit – Absolute limit
NOX: 7.0 g/kWh NOX: 7.0 g/kWh
PM: 0.1 g/kWh PM: 0.1 g/kWh
– Functional check – Subject to review
for exhaust-gas by EU Commission
treatment system
Table 1
Global service 521
왘 Global service
“Once you have driven an automobile, you creased rapidly. In the 1920s Robert Bosch
will soon realize that there is something started to systematically create a nationwide
unbelievably tiresome about horses (. . .). customer-service organization. In 1926 all the
But you do require a conscientious me- repair centers were uniformly named “Bosch
chanic for the automobile (. . .)”. Service” and the name was registered as a
Robert Bosch wrote these words to his trademark.
friend Paul Reusch in 1906. In those days, it Today’s Bosch Service agencies retain the
was indeed the case that breakdowns could same name. They are equipped with the latest
be repaired on the road or at home by an em- electronic equipment in order to meet the de-
ployed chauffeur or mechanic. However, with mands of 21st-century automotive technology
the growing number of motorists driving their and the quality expectations of the customers.
own cars after the First World War, the need
for workshops offering repair services in-
1 A repair shop in 1925 (photo: Bosch)
æ UWT0079Y
2 A Bosch car service workshop in the 21st century carried out with the very latest electronic testing equipment
æ UWT0080Y
522
 Index
ABS 354 air-conditioner diagnosis 519

ABS control loop 358 air-conditioning control 88
ABS controller 392 airflow rate 485
ABS hydraulic modulator 405, 408 air-gap fluctuations 197
ABS systems 411 air-mass flow 224
ABS versions 407 air-mass flow rate 224
absolute humidity 231 air-temperature sensor 263
absolute rotating-speed measurement 205 alternator control 477
A/C control unit 68 altitude compensation 465
acceleration sensors 207 AMA 93
acceleration/deceleration during cornering 386 AMR 184
accelerator-pedal sensors 264 AMR angular-position sensor 186
achievable shifting times 321 AMR basic principle 185
ACK check 78 AMR element 266
ACK field 77 AMR sensors 247
acoustic signaling devices 488 AMR speed sensor 203
action time 119 AMR steering-angle sensor 266
activation of converter and torque converter AMR travel sensor 187
lockup clutch 339 analog hall angular-position sensor 179
active member adress 93 analog input signals 18
active speed sensors 246 analog signals 153
active star topology 125 anemometers 224
active steering 478 angle measurement 222
active wheel-speed sensors 253 f. angle-measuring sensors 221
active-surge damping 464 angular measurement 189
actuator chain 290 angular positions 168
actuator module 440 angular-position sensor 173, 187 ff., 264
actuator selection 312 anisotropic etching 162
actuator triggering 474 anisotropic magnetoresistive sensors 184
actuator types 304 antiferromagnet 191
actuator unit 414 antilock braking system 354
actuators 304, 455 application data 122
adaptive frequency hopping 92 application layer 52, 74
adaptive functions 39 application protocols 138
adaptive pressure control 325 application-mode change 36
adaptive transmission control 327 arbitration 38, 141
addressing 48, 104 arbitration field 77
adjustment 145 arbitration phase 75
adjustment systems 484 architecture 8, 11
administrative functions 106 ARD 464
adressing 75, 142 array 243
adressing modes 48 ASC operation 327
AFH 92 ASCET-SD 32
aging 155 ASIC 346
air conditioner 477, 487 ASIC chips 346
air conditioning 485 AST 318, 321
air conditioning system 486 AST shifting operation 320
air gap 198, 200 asymmetric brake application 373
air quantity 224 asynchronous channel 105
airbag triggering 276

DOI 10.1007/978-3-658-03975-2, © Springer Fachmedien Wiesbaden 2015
Index 523
asynchronous machines 298 brake-by-wire 3

AT 119 brake-fluid pressure sensor 262, 451
ATC 327 braking control cycle 362, 364
automated shift transmission 318, 321 braking sequence without ABS 363
automatic brake functions 394 braking system 354
automatic steel folding roof 67 braking torque 373
automotive applications 147 bridge circuit 204
automotive control engineering 377 broadcast address 104
automotive networking 56 BS 143
automotive sensors 144 bulk current injection method 507
autosar 16 bulk silicon acceleration sensor 278
autosar initiative 16 bus access 86
average delivery adaption 466, 470 bus access method 48
axle sensors 271 bus coupling 72
bus driver 126
background program 29 bus guardians 113, 127
backup fault memory 42 bus states 129
bandwidth 501 bus system 22, 55, 70
barber pole structure 185 bus topology 45, 71, 115, 125
barber pole 186 bus voltage 129
basic CAN 80 bus-access diagram 116
basic CAN module 80
baud rates 142 calibration 146
BCI 507 calibration of a resistor 236
BCI method 507 calibration process 146
BDW 395, 402 camshaft sensors 248
Bernoulli’s equation 226 CAN 137, 142, 477
BFM 42 CAN bus 70
biometric features 494 CAN Calibration Protocol 138
biometric systems 494 CAN controller 80
bit arbitration 75 CAN extended data frame 38
bit rates 38 CAN interface 38
bit stream 51, 130 CAN message format 76
bit stuffing 79 CAN protocol 74
BlockSize 143 CAN standard data frame 38
bluetooth 90 CAN-B 70
bluetooth architecture 96 CAN-bus arbitration 38
bluetooth device addresses 96 CAN-C 70
bluetooth networking 90 capacitive coupling 498
bluetooth versions 91 capacitive pick-off 213, 277
bolometer 242 capacitive sensor wafer 232
bolometric sensor array 242 capacitors 278
bondzone microcontroller LTCC 344 CARB initialization 141
brake assist function 398 CARB 137
brake control unit 68 CAS 132
brake disk wiping 395, 402 cascaded star topologies 125
brake slip 357 catalytic converter diagnosis 518
brake torque 217 CCP 138
brake wiping 413 CCD principle 244
524 Index
CCD sensors 245 comering 367

CDD 401 comfort control unit 68
CDP 398 comfortable interior temperature 485
central bus guardian 127 commercial vehicles 427
central locking mechanism servo unit 489 common rail 450, 511
central locking system 489 common-rail fuel-injection system 416 f.
central processing unit 22 common-rail system 416, 422, 424, 427
central registry 106 common-rail system for commercial vehicles
chamber pressure model 315 457
changeover of register pump 305 common-rail system for passenger cars 456
characteristic curves 154 communication 20
characteristic curve type 152 communication controller 126
charge-coupled devices 244 communication cycle 116, 130 f., 133
chassis domain 69 communication layer 52
checksum 139 communication network interface 112
cioriolis acceleration 206 components of active steering 479
circuit diagram 138 composability 54, 123
circuit simulation 349 comprehensive components 519
classification of bus systems 59 concentration sensors 230
cleanliness quality 431 configuration status 106
cleanliness requirements 431 connection master 106
clique avoidance 120 contact-pressure control 338
ClkSyn flag 119 content-based addressing 74
clock control 132 continuity equation 226
clock control hierarchy 132 continuous linear curves 152
clock generator 22 continuous nonlinear curves 152
clock synchronization 119 continuous operation with intermittent loading
closed-loop and open-loop control 472 301
closed-loop control 472 continuous-running duty 301
closed-loop control system 388 control applications 472
cluster mode 121 control cycle 365
cluster startup 119 control cycle on slippery surfaces 364
clutch activation 339 control cycle on surfaces with good grip 362
clutch control 323 control field 77
clutch torque 329 control loop 377
CMOS imaging sensors 245 control mechanisms 52
CNI 112 control message 107
coding 71, 269 control of automated shift transmission AST
coefficient of friction 357 318
cold junctions 243 control of automatic transmission 322
collision avoidance symbol 132 control of continuously variable transmission
collision detection 141 338
collision prevention 141 control of hydraulic modulators 411
COM layer 117 control of injected-fuel-quantity compensation
comb electrodes 232 464
combined drivetrain ECU 343 control on the high-pressure side 418
comb-like structure 277 control unit software 26
combustion-miss detection 518 controlled deceleration for parking brake 398
combustion-synchronous interrupt 28 controlled deceleration for driver assistance sys-
Index 525
tems 395, 401 data packets 95

controlled system 359 data point program map 147
controlled torque converter lockup clutch 337 data processing 18, 458
controlled variables 360, 388 data rate 85
controller hierarchy 389 data transfer 103
controller state 120 data transfer rate 57
controlling bus access 75 data transfer sequence82
conventional braking-force distribution 396 data transmission channels 100
conventional shifting-point selection 326 data transmission system 71, 73, 85
conventional shifting-sequence control 323 DBC 348
coolant-temperature sensor 263 DD 395
cooperative scheduling 34 deactivation of limp-home function 339
cooperative task change 35 deferred pending mode change 121
coriolis acceleration 257 f. delivery calibration 439
coriolis vibrating gyros 258 detection of undesirable combustion 471
cornering 327, 381, 383 development methods 4
correction functions 420 development of hydraulic modulators 411
correction model 164 development process 30
correction parameters 146 development trends 10
countersteering 385 device model 101
coupling of networks 61 DFPM 516
CPU 22 diagnosis communication 137
CPU power distribution principle 26 diagnosis connection 136
crankcase ventilation diagnosis 518 diagnosis fault path management 516
crankshaft position calculation 27 diagnosis interfaces 136
crankshaft sensor 27 diagnosis legislation 136
CRC 134 diagnosis VALidator 517
CRC checksum 122 diagnostic functions 42
CRC field 77 diagnostic function SCHEDuler 517
critical obstacle-avoidance maneuver 383 diagnostic protocol 137
cross system functions 56 diagnostic system 510
cruise control 463 diagnostic system management 516 f.
C-state 120, 122 diagnostic tester 515
CTCC 337 diagraphm material 213
current regulator 346, 458 diaphragm 262
customer-specific adapter circuits 149 diagraphm-type sensors 212
CVT 338 diesel engine 421
cyclic redundancy check 78 diesel engine applications 421
cylinder shutoff 465 diesel rail-pressure sensor 451
differential hall sensor 178, 247
damping 208 differential lock 376
damping factor 208 differential sensor 197
data bytes 139 digital electronics 274
data exchange 476 digital hall angular-position sensor 183
data field 77, 87, 108 digital input signals 19
data frame 76, 103 digital modules 22
data length 143 direct bonded copper 348
data link layer 118 direct control 323
data memory 22 direct control 471
526 Index
direct pressure measurement 212 ECU 508

direct-contact sensors 235 ECU development 349
direct-current machines 297 ECU network 37
directional stability 403 ECUs in the drivetrain 335, 340
discrete output signal 153 EDC 417, 454, 459
displacement 207 EDC variants 460
displacement measurement 187 eddy currents 269
distance measurement 193 eddy-current principle 171, 270
DL 143 eddy-current sensors 171, 223
DMC 121 eddy-current torque sensor 223
dominant levels 50, 76 EDR 92
door control units 68 EDTC 392
double hall sensor 178 EE system architectures 13
downshift 324, 331, 334 EEPROM 25
downshift under load 324 effectiveness of control 361
DRAM 25 EHM 353
drift 155 electrodynamic principle 291
drive and adjustment systems 482 electric actuators 290
drive axle speed controller 372 f. electric fuel pump 422
drive mode 462 electric motors 509
drive slip 370 electrical EPROM 25
drive torque 373 electrical locking system 493
driven wheels 361 electrical machines 296 f.
driver-type determination 40 electrical power input 227
driver-type recognition 327 electrical signal evaluation 210
drivetrain 332 electric-motor automated shift transmission
drivetrain ECUs 342 AST 319
drivetrain management 316 electric-motor clutch management 318
driving program 339 electrodynamic converter 291
driving speed 64 electrodynamic immersion-coil actuator 294
driving stability systems 481 electrodynamic rotary actuator 294 f.
driving stabilization 480 electrodynamic short-stroke linear motor 294
driving-situation recognition 327 electrohydraulic control 323
DSCHED 517 electrohydraulic modules 353
DSM 516 electromagnetic actuators 292 f.
dual-hole version 238 electromagnetic compatibility 496
dual-port RAM 38 electromagnetic converter 291
duty-type ratings 301 electromagnetic principle 291
DVAL 517 electromagnetic rotary actuators 293
dynamic laterial response 403 electromagnetic single-winding rotary actuator
dynamic RAM 25 293
dynamic segment 131 electromechanical actuators 290
dynamic slots 131 electromechanical components 509
dynamics of a braked wheel 357 electromechanical converters 290
electromechanical door lock 493
EBD 396 f. electromechanical seat adjustment 484
EC motors 299 electron 443
ECM 318 electronic braking-force distribution 396 f.
economy mode 326 electronic control unit 18, 21, 455, 508
Index 527
electronic diesel control 417, 454, 459 error types 154

electronic distribution 396 ESP control loop 389
electronic heater control 485 ESP controller 390
electronic immobilizer 476 ESP controller functions 392
electronic modules 352 ESP hydraulic modulator 404, 406
electronic stability program 378, 388 ESP 378, 388
electronic system 511 etching rate 162
electronic transmission control 316 ETI 336
electronic transmission control unit 340 event control 52
electronic water-side control 487 exhaust-gas recirculation system diagnosis 518
electronically controllable differential lock 376 exhaust-gas temperature sensor 263
electronically controlled air conditioning system extended steering and countersteering sequence
485 384
electronic-control-unit network 138 external torque demands 474
electronics 443 external torque intervention 477
electrohydraulic actuators 302
elongation 213, 218, 442 failure quota 155
EM 352 failure rate 155 f.
EMC 496, 508 fanfare horns 488
EMC ranges 496 fast initialization 140
emergency braking 382 fault confinement 79
emission limits 73 fault diagnostics 510
emitted interference 504 fault information 515
end of frame 78 fault memory 42
engine control unit 417 fault storage 512
engine cooling system diagnosis 519 fault-tolerance strategy 111
engine intervention 328 fault-tolerant average algorithm 133
engine speed 64 FBlock 107
engine-brake function 465 fiber cable 103
engine-management sequence 474 f. field direction modulation 188
engine-oil temperature sensor 263 fill time 325
engines parameters 473 fill-time measurement 325
engine-speed sensors 246 filtering out interference 72
engine-temperature sensor 263 fingerprint sensor 495
engine-torque intervention 336 finger-type Lambda oxygen sensor 285
engine-torque specification 335 finger-type sensor 285
enginge drag-torque controller 392 5-baud initialization 140
enhanced brake assist function 413 fixed bolted connection 275
enhanced data rate 92 fixed limitation 463
EOBD 513 flash EPROM 22
EOL programming 20 flash EEPROM 25
EPROM 22, 25 flat-seat pressure regulator 308 f.
erasable PROM 25 flat-seat pressure regulator, falling characteristic
erasable ROM 25 310
error 154 flat-seat pressure regulator, rising characteristic
error detection 78, 512 311
error frame 76 FlexRay 69, 124
error graph 154 FlexRay application 68
error handling 79, 512 FlexRay controller 126
528 Index
FlexRay cycle 130 function simulation 313

FlexRay Group 135 functional structure 11, 13, 15
FlexRay protocol 128
flow profiles 225 GaAs 148
flow rate calculation315 gage factor 213 f., 218
flow simulation 313 gain error 154
flowmeters 224 gallium arsenide 148
fluid-mechanical actuators 295 f. galvanic coupling 497
fluid-thermal losses 295 galvanomagnetic effect 176
flux and voltage curve 199 GAP 98
force balance 314 gas measurement 230
force balance model 314 gas sensors 230
force generation 291 gas-analysis processes 230
force sensor 214 f., 220 gasoline rail-pressure sensor 451
forward voltage 241 gateway 68
4-layer microhybrid 347 gateway structures 61
four wheel drive vehicles 374 Gaussian frequency-shift keying 92
four-wheel-drive configurations 368 gear change 305
four-wheel-drive vehicle 368, 375 gear presupply pump 448
frame 86 gear pump 422
frame bit stream 130 gear selection 39 f.
frame check 78 gear wheel scanning 202
frame fields 122 generic access profile 98
frame format 121, 134 generic object exchange profile 98
frame ID 134 GFSK 92
frequency band 91 giant magnetoresistive elements 203
frequency hopping 95 giant magnetoresistive sensors 189
frequency-hopping method 91 global service 521
friction coefficient 359 glow control unit 476
frictional connection 332 GMA 366
FT-COM 112 GMA system 365
FT-COM layer 112, 117 f. GMR 189, 203
FTM 133 GMR angular-position sensor 191
fuel filtering 427 GMR element 267
fuel injection 419 f. GMR multilayers 190
fuel rail 450 GMR sensors 190, 204, 247
fuel system diagnosis 518 GMR sensor technology 189
fuel-consumption signal 476 GMR spin valve 190 f., 204
fuel-delivery control 419, 460 GOEP 98
fuel-injection process 461 GPS 193
fuel-injection system 425 GPS position measurement 193
fuel-quantity 435 gradient correction 133
fuel-temperature sensor 263 gradient deviation 154
full CAN 81 gradient sensors 202
full-CAN module 80 gray code 182
full-load smoke limitation 470 f. group addresses 104
fully-active control 410
function block 107 f. haldex coupling 376
function classes 109 half-differential short-circuiting ring sensor 173
Index 529
hall angular-position sensors 179, 265 hot-wire air-mass meter 228

hall ASICs 250 hot-wire/hot-film anemometers 227
hall effect 176 HRB 395, 402
hall IC 177, 201, 255 humidity sensors 232
hall phase sensors 248 humidity-measurement procedures 232
hall rod sensor 202, 248 hybrid integration 148
hall sensor 177, 250 hybrid topologies 47, 115, 125
hall sensor arrangement 180 hybrid wiring densities 344
hall switch 176, 183, 268 f. hydraulic brake assist 398
hall vane switch 177, 201 hydraulic control 322
hall voltage UR 176 hydraulic coupler 440
hardware 80, 126 hydraulic diagram 405 f.
hardware access 36 f. hydraulic fading compensation 395, 401
hardware-side protocols 96 hydraulic modulator 355, 404, 411
HBA 398 hydraulic modules 351
HDC 395, 400 hydraulic rear wheel boost 395, 402
header 86, 122, 139 hydraulic zero 438
header segment 134 hydraulic-control system 322
heat-dissipation model 347
heater element 286 ideal value 154
heating 485 identifier 87
heating current 227 identifier field 87
heavy-duty trucks 520 idle-speed control 462
hella sensor 174 I-Frame 122
HFC 395, 401 ignition interrupt 28
HHC 395, 399 IMA 439
high coefficient of friction 362 imaging sensors 243 f.
high-adhesion conditions 362 immunity 508
high-frequency signal feedback 500 impact-pressure air-mass meter 226
high-precision multisensor 187 impact-pressure flowmeter 226
high-pressure accumulator 450 improved protection against contamination 274
high-pressure components 430 incident radiation 504
high-pressure control 418, 422 incident radiation via antenna 508
high-pressure pump 445 ff. increasing reliability 157
high-pressure sensor 262, 451 increasing steering-wheel angles 385
high-pressure system 417 indicated value 154
high-speed CAN 70 individual intervention strategies 379
high-voltage ignition 509 indrect control 470
hill descent control 395, 400 inductance 173
hill hold control 395, 399, 413 inductive coupling 498
HM 351 inductive engine-speed sensor 246
holding-current phase 436 inductive speed sensor 198, 246
Hooke’s law 218 inductive wheel-speed sensor 252
host layer 118 inductive-type sensors 198 f.
host processor 112, 126 infrared light 244
host stack 97 inhibit conditions 516
hot junctions 242 initial braking phase 361
hot-film air-mass meters 272 f. initialization 139
hotspots 348 injected-fuel-quantity limit 465
530 Index
injection-quantity program map 439 laminar flow 225

injector 432 lane change 382
injector closed 433 f. laser diodes 103
injector delivery compensation 439, 466 f. lateral dynamic response 378 f.
injector fully open 433 lateral-force coefficient 357
injector opens 433 layer stack 204
injector variants 434 Lehr’s damping factor 208
in-line piston pump 449 light sensor 283
input and output devices 22 light-duty trucks 513
input signals 18 limp-home function 512
input/output 22 LIN bus 84
input-signal monitoring 511 LIN message format 86
instrument cluster 68 LIN network 88
intake tract 224 LIN networking 89
intelligent bolt sensor 280 LIN protocol 86
interference 501 linear bus topology 45, 84, 477
interference from electrostatic charges 504 linear solenoid 295
interference immunity 57 linearity deviation 154
interference suppression 496, 508 load interruption 332
interference-signal spectrum 500 load management 65
interference-suppression categories 503 load transfer 323
interference-suppression levels 502 local bus guardian 127
interframe space 78 local interconnect network 84
intermediate-speed control 463 locking mechanism 492
intermittent-periodic duty 301 locking systems 490
internal timer 27 logic bus states 71
internal torque demands 474 logic circuits 22
interpolation 147 logical addresses 104
interrupt control 26 logical data channels 95
interruption of tractive force 320 logical line interface 114
inventory 156 logical link control and adaptation protocol 97
inverse piezoelectric effect 442 longitudinal effect 209
I/O 22 longitudinal K factor 219
I/O ASIC 346 longitudinal resistance 184
ISIG signal characteristic 43 long-range distance measurement 192
ISO OSI reference model 50 long-range radar sensors 192
lost energy 329
key bytes 141 low 373
K line 136, 138 low coefficient of friction 364
knock sensor 275 low-adhesion conditions 364
knock-sensor signals 275 low-pressure stage 417
KWP 2000 137 low-speed CAN 70
KWP 2000 protocol 139 low-volume design 161
KWP 71 137 LTCC 344
LTCC-fine-line 344
lambda closed-loop control 468, 470 L2CAP 97
lambda oxygen sensor 284
lambda-based EGR control 469 macroticks 132
lambda-oxygen-sensor diagnosis 518 magnetic carriage 269
Index 531
magnetic flux 198 message scheduling 88

magnetic permeability 216 message transfer 143
magnetically inductive-type sensors 171 message types 142
magnetic-circuit calculation 314 message-oriented method 48
magnetoelastic anisotropic effect 216 metering unit 447
magnetoelastic effect 217 metering unit design 447
magnetoelastic tensile-force/compressive-force microbending effect 160
sensor 217 microcontroller 22 f.
magnetoresistive angular-position sensor 186, microhybrid circuit 148
188 microhybrid ECUs 344 f.
magnetoresistive differential sensor 183 microhybrid layout 349
magnetoresistive sensors 183, 282 micromechanical bulk silicon acceleration sen-
magnetoresistor 184 sors 278
magnetostatic sensor 176, 201 micromechanical hot-film air-mass meter 229
main control with hydraulic valves 322 micromechanical manufacturing 161
malfunction indicator lamp 514 micromechanical pressure sensors 260
maneuvers 380 micromechanical thermopile array 243
manufacturing costs 158 micromechanical yaw-rate sensor 4, 256
MAR 464 micromechanics 4, 161
mass flow rate 225 microsystem technology 161
master and slave nodes 84 microticks 132
master-slave 49 MIL 514
master-slave method 49 minislots 130 f.
master-slave principle 86 mixed scheduling 35
maturity model 33 mobile multimedia devices 90
maximum junction temperatures 348 mobile radio reception 500
maximum-rpm control 462 model library 4
McMess 138 modular system 351
mean service life 156 modulation method 92
mean temperature coefficient 237 modulation with ABS control 408
measuring incident interference 502 module types 351
measuring principles 208 modules for transmission control 350
measuring range 207 moisture measurement 230
mechanical door lock 490 monitoring 78
mechanical locking system 491 monitoring architecture 316
mechanical shifting sequence 329 monitoring during vehicle operation 510
mechanical stress 209 monitoring ECU communication 512
mechatronic modules 350 monitoring functions 42
mechatronic systems 2 monitoring internal ECU functions 512
mechatronics 2 monocrystalline silicon semiconductor resistors
media access test symbol 132 238
Media Oriented Systems Transport 100 MOST 100
MEDL 113 MOST 25 103
membership service 120 MOST application layer 107
mesh topology 46 f. MOST bus 100
message components 113 MOST cooperation 100
message descriptor list 113 MOST device 101
message filtering 75 MOST network 102
message format 76, 139 MOST standards 110
532 Index
MOST system 101 OBDI 513

motion equation 314, 331 object layer 74
movable magnet 180 object-oriented approach 39
MTS 132 offset correction 133
multilayered process 149 on/off solenoid values 304
multimaster 49 on-board diagnosis 510
multimaster principle 74 on-board diagnosis system 496, 513, 520
multimedia application 66 1D Simulation 314
multimedia data 63, 105 one-way clutch US 323
multimedia networking 60, 66 open and closed-loop control functions 338
multiplex applications 60 open-by-wire 492 f.
multi-pole alternating field 175 opening phase 436
multipole pulse generator 254 open-loop control 472
multi-slot packets 95 operating conditions 18, 159
mutual interference 497, 499 operating modes 128, 515
µ-split 365, 373 operating principle 269
operating system 34
natural frequency 332 optimization of magnetic force 314
near-range distance measurement 192 optoelectronic angular difference steering torque
needle lift 435 sensor 222
nested AMR bridges 188 orthogonally printed resistors 219
NetBlock 106 osciallation gyroscope 205
network architecture 15 oscillatory movement 206
network idle time 130, 132 OSI reference model 50
networking 44, 70 out-of-range check 519
network management 88 output signal 20, 153
network master 106 output-signal monitoring 511
network node 58, 71 over and understeering behavior 381
network node architecture 126 overlap control US 325
network organization 48 overload frame 76
network size 114 oversteering 383
network topologies 44, 62, 115 overview of diesel fuel-injection systems 426
N-Frame 122
NIT 130, 132 packet data 105
NMK 439 parallel connection 297
non-contacting temperature measurement 242 parking-aid assistant 68
non-driven wheels 360 partially-active control 410
non-return to zero 129 particle-analysis system 431
nozzle needle 434, 438 particulate-filter diagnosis 519
NRZ method 129 passenger cars 422, 513
NTC resistors 236 passenger-car diesel engines 468
NTC temperature sensor 263 passive nodes 116
passive star topology 125
OBD 513 passive wheel-speed sensor 252 f.
OBD emission limits 520 payload segment 135
OBD functions 517 PCS 103, 348
OBD II 513 pedal travel simulation 414
OBD limits 514 permissible radio interference-field strength
OBD system 514 f. 502
Index 533
PFM 42 power-output 473

physical addresses 104 PRC 43
physical connections 95 preemptive scheduling 35
physical data channel 94 preemptive task change 35
physical effects for sensors 165 pressure calculation 315
physical layer 50, 74, 117, 129 pressure correction 326
physical layout 138, 142 pressure control 418
pickup-current phase 436 pressure-control valve 452
piconet 93 pressure drop 226
piezoelectric 442 pressure generation 409, 418 f.
piezoelectric acceleration sensors 279 pressure maintenance 408
piezoelectric bimorph 210 pressure measurement 212 f.
piezoelectric coefficient 442 pressure modulation 408 f.
piezoelectric components 209 f. pressure regulator 308
piezoelectric effect 233, 442 pressure sensors 212, 260 f.
piezoelectric field strength 442 pressure-control valve 303, 446
piezoelectric knock sensors 275 pressure-regulator monitoring 43
piezoelectric sensors 210 pressure-regulator type 312
piezoelectric yaw-rate sensor 259 pressure-relief valve 453
piezo-inline injector 438, 441 pressure-wave correction 466
piezoresistive behavior 219 primary fault memory 42
pilot control 323 printed circuit board ECUs 341
planar hall sensors 181 priority assignments 76
planar Lambda oxygen sensor 286 priority distribution 35
planar wide-band lambda oxygen sensor 288 process 33
plastic optical fiber 102 processes and talk 35
POF 102 profile hierarchy 99
POF cable 102 profiles 98
point-to-point 125 program memory 22
poisson ratio 214 program-map variants 434
position control 208 programming 22
position detection 268 programming guidelines 31
position determination 193 program-run check 43
position sensor 168, 268 ff. PROM 22
potential hydraulic power 419 propulsion torque 474
potentiometer sensors 169 protective tube 285
potentiometer-type sensors 170 protocol layers 74, 128
potentiometric accelerator-pedal sensor 264 protocol services 119
potentiometric fuel-level sensor 170 protocol service layer 118
power 473 protocol stack 96, 117, 129
power chips 348 pseudo hall element 186
power classes 92 pseudo hall sensor 186
power loss 347 pseudo hall voltage 185
power sunroof drive unit 483 Pt resistor 238
power sunroofs 483 PTC 238
power supply 499 PTC thin-film/thick-film metallic resistors 237
power window 482 PTC/NTC 238
power window control 482 f. pulsating air-mass 224
power window motors 482 pulsating air-mass flow 229
534 Index
pulsation in the vehicle electrical system 499 relative speed 329

pulse wheels 254 relative switch-on duration 301
pulse-type input signals 19 reliability 155 f., 158
pulse-wheel teeth 248 reliability requirements 157
pulse-width modulated signals 255 remote frame 76
pump activation 339 requirements for bus systems 57
pump cell 289 requirements on vehicle sensors 158
pump current 289 requirements specification 6
pump plunger 445 resistance 185
PWM signals 20 resistance/voltage conversion 235
PWM switching valves 303 resistive sensors 235
PWM Valves 306 resolution 63
pyrometric sensor 242 resonance curves 209
resonant frequency 208
quadruple hall sensor 182 resonant system 208
quality assessment 30 response time 335
response to braking 367
radial-field sensor 197 restrictor 315
radial-piston pump 445, 447 f. restrictor flow model 315
Radio Fequency Communication 97 retrofit interference suppression 509
rail-pressure sensor 262 return force 207
rain sensor 283 reverse-gear lock 339
rain/light sensor 283 Reynolds number 225
RAM 22, 25 RFCOMM 97
random pixel access 245 right-left cornering sequence 381
random-access memory 25 ring structure 101
rapid steering 385 ring topology 46
rapid steering and countersteering 380 ROM 22, 25
ratio control 339 roof control unit 67
ratio criteria 41 rotary motion 205
rationality check 519 rotatable alternating fields 174
readiness code 516 rotating field speeds 299
read-only memory 25 rotation measurement 205
real-time applications 59 rotational position determination 269
real-time capability 26, 58 rotational speed 195
rear door variant 491 rotors 197
receiver 496 rpm measurement 195
reception 82 rpm sensors 195
recessive level 50, 76
recharging the step-up chopper 436 safety functions 305
recursions 6 SBC components 412
recursion method 6 SBC functional modules 415
reference vacuum 260 SBC sensors 414
reflection-free termination 73 SBC 412
refrigerant circuit 486 scan tool 514
reintegration of a node 121 scanning electron microscpope 431
relative acceleration 332 scatternet 93 f.
relative humidity 231 SDAP 98
relative inventory 155 SDP 97
Index 535
seat and steering column adjustment 484 signal level 50

secondary fault memory 42 signal processing 19
seebeck effect 239 signal quality 255
segmented messages 142 f. signal shapes 153
SEM 431 signal transmission 61
semiconductor depletion layers 241 signal types 61
semiconductor memories 24 signal-range check 519
sensitivity 207 f. signals 153
sensor classification 152 simplified V-model 30
sensor cluster 258 simulated automotive electrical circuit 502
sensor function 145 simulation 4
sensor integration levels 164 simulation tools 349
sensor market 150 simulations in development 313
sensor measuring cell 277 simultaneous engineering 7
sensor measuring principles 168 single differential barber pol 187
sensor model 144 single-axle driven vehicle 370
sensor plates 226 single-phase alternating-current machines 300
sensor prototyping phases 152 single-phase asynchronous motors 300
sensor shapes 196 single-pixel sensor 243
sensor symbol 145 single-point sensors 243
sensor technologies 167 single-track model 391
sensor types 246 single-wire line 72
sensors 455 sintered-ceramic 236
sensors of the SBC systems 414 sleep mode 81
sensotronic brake control 412 slip 359
sequence control 333 f. slip time 326, 329
serial data transmission 477 slip-ring rotor 299
serial port profile 98 slip-time measurement 325
series connection 297 slot counter 131
series-wound machine 297 slot timing 117
service discovery application profile 98 smallest replaceable unit 111
service discovery protocol 97 smart sensor 145 f., 164
service lives 156 SMM acceleration sensors 276
servo valve 439 smooth-running control 464
setpoint generators 455 snapshot memory 42
SFM 42 soft stop assist 413
shared vehicle electrical system 497 software 32
shear effect 210 software architecture 14
shifting operation and interruption of tractive software development 30
force 320 software layer model 34
shifting the ignition angle 335 software sharing 320
shifting with torque correction 319 software structure 26, 34
shifting without torque correction 318 solenoid-valve injector 433, 436
shifting-point selection 326 solenoid-valve monitoring 42
shifting-sequence control 323 spectrum 501
short-circuiting ring sensor 172 speed measurement 195
short-time-duty type 301 speed sensing 247
shunt-wound machine 297 speed sensors 195, 249
signal conditioning 19 spinning-current principle 177
536 Index
split 373 synchronization speed 326

spool-type pressure regulator 308 f. synchronous channel 105
sport mode 326 synchronous machines 299
SPP 98 synthetic antiferromagnets 191
spreading-resistance principle 238 system and control structure 388
spreading-resistance sensor 239 system events 104
spring element 279 system failure 415
squirrel-cage rotor 298, 300 system modules 455
SRAM 25 systematic error 155
SRC 464
SRU 111 tangential sensor 197, 203
standardization 83, 110, 123, 135 tangential-field sensor 203
standstill recognition 255 TCS 370, 374
star bus topology 47 TCS BIN 97
star ring topology 47 TCS controller concept 372, 375
star topology 45, 71, 115, 125 TCS hydraulic modulator 406
star-connected three-phase synchronous gene- TCS interventions 371
rator 299 TCS transversal differential lock controller 377
start of frame 77 TDMA 117
start quantity 462 TDMA 49
starter control 476 Telephony Control Protocol Specification-Binary
static RAM 25 97
static segment 131 TEM 507
static slots 130 TEM cell 507
steering train 479 TEM-cell test setup 508
steering-angle sensors 266 temperature 234
stochastic errors 155 temperature coefficient 237
storage 24 temperature measuring points 234
strain gage pick-off 213 temperature sensors 234, 263
strain measuring force sensors 216 temperature/humidity diagram 231
strain-gage force sensors 218 temporary disabling of diagnostic functions
strain-gage torque sensor 221 516
strain-measuring sensors 220 thermal acceleration sensors 211
stripline method 506 f. thermal conductivitiy 227
structure of traction control system 372 thermal image analysis 243
subscriber-oriented method 48 thermal resistance 348
suction side 419 thermal vias 347
surface micromechanical acceleration 276 thermocouple measuring set-up 240
surface micromechanical acceleration sensor thermocouples 239
148, 163 thermocouples connected in series 240
surface micromechanical pressure sensor 163 thermoelectric voltage 239
surface-micromechanical yaw-rate sensor 257 thermo-management 347
switching signals 20 thermophile pick-off 242
switching solenoid 292 thermopile sensor 242
switching valves 296 thermo-simulation 349
switchoff 436 thick-film resistors 238
switchon conditions 516 thin-film sensors 185
symbol window 132 three-phase asynchronous motor 298
synchronization 86, 119 three-phase machines 298
Index 537
3/2 PWM valve with high flow rate 307 transmission media 115
3/2 PWM valve with rising characteristic 307 transmission of control commands 104
3/2 PWM valve 306 transmission speed sensor 251
3/2 switching valve 305 transmission technology 91
3/2-way servo valve 438 transmission 82
throttle-valve sensor 171 transmission-oriented method 48
time control 116 transmitter 496
time division multiple access 49 transmitter initiative 78
time frame 29 transport layer 74
time synchronization 133 transversal differential lock controller 372 f.
time-controlled protocol 111 transverse effect 210
timer control 53 f. transverse electromagnetic field 507
time-triggered architecture 111 travel measuring force sensors 220
Time-Triggered Protocol 111 travel positions 168
timing of data transfer 94 travel-measuring systems 207
timing 132 f. triangulation 193
tiptronic shifting strategies 337 triggering phases 437
tolerance diagram 238 triggering sequence 436, 441
tolerance graph 155 triggering the interrupt 28
tolerance ranges 85 trimming 146
tolerance requirements 154 TSC controller 393
tools 32 TTA 111
tooth interrupt 27 TTA network 112
tooth space 200 TTA protocol stack 117
topology 61, 71, 92, 101, 114, 125 TTP/C 111
torque 473 TTP/C Controller 112
torque control 473 TTP/C node 113
torque converter lockup clutch 305, 336 TTP/C protocol 117
torque difference 361 tuning-fork 259
torque measurement 220, turbulent flow 225
torque motor 293 turn signaling 64 f.
torque sensor 215, 220, 223, 282 two-channel topologies 125
torque-controlled diesel injection 475 two-step curves 152
torque-controlled EDC systems 473 two-step lambda oxygen sensors 284
torsion angle 221 two-value capacitor motor 300
torsion-measuring sensors 221 two-wire current interface 250
total error 154 two-wire line 72
tow-actuator system 419 typical control cycles 362
traction control system 370, 374 typical control situations 373
traffic jam assist 413
trailer control unit 68 UART interface transmission frame 51
trailer segment 135 UART interface 51
transceiver 81 ultrasound 193
transfer function 209 ultrasound sensors 192
transformer pick-off 221 understeering 383
transmission acknowledgement 120 uninterrupted duty 301
transmission agent 72, 102, 126 universal motors 300
transmission control 249, 268 unsegmented messages 142
transmission functions 302 unsuccessful communication 141
538 Index
uphill driving 327 windshield-wiper operation 65

upshift 324, 329, 333 winter recognition 327
upshift under load 323 wiper potentiometer 169
wiper stage 65
variable acceleration 208 wrap principle 338
variable limitation 463 WUP 132
variable orifice plates 226 WUS 132
variable steering ratio 478
VCSEL 103 X-by-wire 2
vehicle control 39
vehicle dynamics sensors 414 yaw moment 365
vehicle measurements 505 yaw-moment build-up 365
vehicle motion domain 12 f. yaw-moment build up delay 365, 366
vehicle power supply control unit 68 yaw moment compensation 480
vehicle recall 515 yaw rate 205 f., 256
vehicle security systems 488
vehicle sensors 151 zener diodes 241
vehicle system architecture 11 zero delivery calibration 466
vehicle tracking 381, 385, 386 f. zero offset 154
vehicle travel directions 403 ZME 447
vehicle-speed controller 463
vehicle-speed limiter 463
ventilation 485
vernier principle 222
vertex 326
vertical halld evice 181
vibrating gyroscopes 258
video 244
virtual sample 7
visible light 244
V model 6 f.
voltage amplitude 500
voltage level 73
volume flow rate 225
wakeup pattern 132

wakeup sequence 128
wakeup symbol 132
watchdog ASIC 346
wave-propagation sensors 192
wheel bearing 255
wheel rotation recognition 255
wheel speed 373
wheel-speed sensors 252, 255, 355
wide- and narrow-band limit values 504
wide-band lambda oxygen sensor 288
widely differing adhesion 365
winding temperature 301
windshields 283

Automotive Mechatronics PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Automotive Mechatronics PDF

Hochgeladen von

Copyright:

Verfügbare Formate

Bosch Professional Automotive

Konrad Reif Ed.

ISBN 978-3-658-03974-5 ISBN 978-3-658-03975-2(eBook)

Library of Congress Control Number: 2014946887

Printed on acid-free paper

Springer Vieweg is part of Springer Science+Business Media

As the complexity of automotive vehicles increases this book presents operational

Complex technology of modern motor vehicles and increasing functions need a

2 Basics of mechatronics 165 Overview of the physical effects for sensors

302 Requirements 404 Hydraulic modulator

316 Electronic Transmission Control 412 Sensotronic brake control (SBC)

350 Modules for Transmission Control 430 High-pressure components of common-rail

488 Vehicle security systems 510 Fault diagnostics

Basics of mechatronics Dipl.-Ing. Christian Gerhardt,

Dipl.-Ing. Hans-Martin Heinkel, Dipl.-Ing. Klaus Miekley,

Dr.-Ing. Klaus-­Georg Bürger. Dipl.-Ing. Roger Frehoff,

Architecture Dipl.-Ing. (FH) Bernhard Bauer,

Dr. phil. nat. Dieter Kraft, Dr. Michael Harder,

Dipl.-Ing. Stefan Mischo. Dr.-Ing. Klaus Kasten,

Electronic control units Schwäbisch Gmünd,

Dipl.-Ing. Martin Kaiser, Dipl.-Ing. Frank Wolf,

Dr. rer. nat. Ulrich ­Schaefer, Dr.-Ing. Johann ­Riegel.

Dipl.-Ing. (FH) Gerhard Haaf.

Sensotronic brake control (SBC) Active steering

The term “mechatronics” came about as Mechatronic systems

Forces, travel, etc. Forces, travel, etc.

K. Reif (Ed.), Automotive Mechatronics, Bosch Professional Automotive Information,

A system that was implemented long ago is Examples at component level

In terms of comparable cost, this more Development methods

2 Model library for a micromechanical yaw-rate sensor

Rigid Elastic Comb-like Detection

From From Bending Segment Undivided Divided From From

The application of various domain-specific

V model depending on the technologies applied, for

A requirements specification (what) must Outlook

3 Recursion method at one level

challenges are emerging with regard to The design approaches of mechatronic

“Simultaneous engineering” across

4 V-model general overview

System design System test

Component design, Component test

Over the last three decades, tremendous Overview

1 Proportion of electrics/electronics in the motor vehicle

Automobile 2000 Automobile 2010

K. Reif (Ed.), Automotive Mechatronics, Bosch Professional Automotive Information,

strates just how networked the functions

tronic control units within different areas objectives simultaneously, development

cations for the subordinate subsystems. Vehicle system architecture

3 Levels in the vehicle motion domain

Environment Driving and assistance

Road layout Guidance level

Vehicle guidance data Vehicle

In the same way as Vehicle motion is Systematic creation of EE system

4 Example of a functional structure for the vehicle motion domain

Vehicle motion coordinator

Drive Brake Steering Chassis Stability intervention

CARTRONIC® concept Bosch introduced this concept to the Auto-

feature of mechanical development or pro- Network architecture

unnecessary the additional electronic con- Outlook

Digital technology furnishes an extensive Design

Operating conditions Hybrid versions combining compact

K. Reif (Ed.), Automotive Mechatronics, Bosch Professional Automotive Information,

of the control unit and used for calcula- Signal conditioning

Dipl.-Ing. Hans-Martin Heinkel,  Dipl.-Ing. Klaus Miekley,

Dr.-Ing. Klaus-Georg Bürger. Dipl.-Ing. Roger Frehoff,

Dr. phil. nat. Dieter Kraft,  Dr. Michael Harder,

Dr. rer. nat. Ulrich Schaefer, Dr.-Ing. Johann Riegel.