Beruflich Dokumente
Kultur Dokumente
AA/PDI-03.02-En The Bosch Yellow Jackets Edition 2001 Technical Instruction Automotive Electrical and Electronic Systems
2002 Automotive
The Program Order Number ISBN
Diesel-Engine Management
Diesel Fuel-Injection: an Overview 1 987 722 104 3-934584-35-7 Basic principles
Electronic Diesel Control EDC 1 987 722 135 3-934584-47-0 Components
Diesel Accumulator Fuel-Injection System
Development
Technical Instruction
Common Rail CR 1 987 722 175 3-934584-40-3
Diesel Fuel-Injection Systems and production
Unit Injector System/Unit Pump System 1 987 722 179 3-934584-41-1
Radial-Piston Distributor
Fuel-Injection Pumps Type VR 1 987 722 174 3-934584-39-X
Diesel Distributor-Type
Fuel-Injection Pumps VE 1 987 722 164 3-934584-38-1
Diesel In-Line Fuel-Injection Pumps PE 1 987 722 162 3-934584-36-5
Governors for Diesel In-Line
Fuel-Injection Pumps 1 987 722 163 3-934584-37-3
Gasoline-Engine Management
Emission Control (for Gasoline Engines) 1 987 722 102 3-934584-26-8
Gasoline Fuel-Injection System K-Jetronic 1 987 722 159 3-934584-27-6
Gasoline Fuel-Injection System KE-Jetronic 1 987 722 101 3-934584-28-4
Gasoline Fuel-Injection System L-Jetronic 1 987 722 160 3-934584-29-2
Gasoline Fuel-Injection
System Mono-Jetronic 1 987 722 105 3-934584-30-6
Spark Plugs 1 987 722 155 3-934584-32-2
Ignition 1 987 722 154 3-934584-31-4
M-Motronic Engine Management 1 987 722 161 3-934584-33-0
ME-Motronic Engine Management 1 987 722 178 3-934584-34-9
Gasoline-Engine Management:
Basics and Components 1 987 722 136 3-934584-48-9
Imprint
Published by: Unless otherwise stated, the above are all em-
Robert Bosch GmbH, 2002 ployees of Robert Bosch GmbH, Stuttgart.
Postfach 30 02 20,
D-70442 Stuttgart. Reproduction, duplication, and translation of this
Automotive Aftermarket Business Sector, publication, including excerpts therefrom, is only
Department AA/PDT5. to ensue with our previous written consent and
Product Marketing, with particulars of source.
Diagnostics, Test Equipment. Illustrations, descriptions, schematic diagrams,
and other data only serve for explanatory pur-
Editor-in-chief: poses, and for presentation of the text. They
Dipl.-Ing. (FH) Horst Bauer. cannot be used as the basis for design, installa-
tion, and scope of delivery. We undertake no lia-
Editorial team: bility for conformity of the contents with national
Dipl.-Ing. (BA) Jrgen Crepin, or local regulations.
Dipl.-Ing. Karl-Heinz Dietsche, Robert Bosch GmbH is exempt from liability,
Dipl.-Holzw. Folkhart Dinkler. and reserves the right to make changes at any
time.
Authors:
Dipl.-Ing. Adolf Fritz Printed in Germany.
(Control Unit Development), Imprim en Allemagne.
Dr. rer. nat. Ulrich Goebel
(Hybrid Circuit Board Production), 1st Edition, August 2001.
Armin Hess English translation of the German edition dated:
(Example Circuit), February 2001.
Karlheinz Isert (1.0)
(Printed Circuit Board Production),
Dr. rer. nat. Ulrich Schaefer
(Basic Principles, Components, Microcontroller)
Automotive
microelectronics
Contents
12 Electronic components
12 Passive components
13 Semiconductor components
14 Diodes
16 Semiconductor resistors
17 Transistors
20 Monolithic integrated circuits
24 Microcomputers
27 Semiconductor memories
30 Opto-electronic components
32 Micromechanical sensors
34 Microcontrollers
34 Microcontroller developments
36 Microcontroller components
37 Microcontroller design and
operating concept
41 Microcontroller memories
42 Microcontroller peripheral modules
44 Microcontroller busses
46 Circuit example
46 Overview
48 Design and construction
55 Operating concept
58 Programming
60 ECU development
60 Overview
64 Hardware development
68 Manufacture of semiconductor
components and circuits
68 Semiconductor components
72 Micromechanics
73 Conventional printed-circuit
boards
78 Film and hybrid circuits
Robert Bosch GmbH
Motor vehicles are now inconceivable without electronic control units and their associ-
ated sensors and actuators. Such components have revolutionised the automotive
world. All essential vehicle functions are electronically controlled by systems and com-
ponents that occupy only a tiny space. There are systems that control the function of
engine and gearbox, the safety and security systems and a large number of comfort and
convenience systems, and the number of intelligent vehicle systems is growing at a
rapid pace. Nor is there any foreseeable end to this trend.
Electronic control systems open the door to a vast array of possibilities. They im-
prove vehicle safety and ride comfort. At the same time, they make vehicles more
economical and kinder to the environment.
The triumphant advance of electronics has created a vocabulary of terms with
which we are bombarded every day. It is becoming more and more difficult to under-
stand what precisely those terms refer to.
This publication in the Bosch Yellow Jacket Technical Instruction series explains the
specialist terminology of microelectronics in detail.
It starts with a short introduction followed by an explanation of the principles of
semiconductor technology and how microelectronic components interact. A practical
example of a circuit helps to consolidate the theoretical concepts described. This is
then followed by a description of the development and production of control units.
Finally, the glossary provides a quick guide to the most important microelectronics
terms.
The functions of the individual electronic systems (e.g. MED, EDC, ESP) are described
in detail in separate publications in the Yellow Jacket Technical Instruction series.
Robert Bosch GmbH
Automotive microelectronics
Microelectronics have revolutionised motor Sensors and setpoint generators
vehicle technology. Initially, mechanical Sensors detect operating conditions (e.g. en-
components were replaced by electronic gine speed, wheel speed, temperature). They
devices in order to make systems more convert physical variables into electrical sig-
reliable. This was the case with the contact nals. Setpoint generators (e.g. controls oper-
breaker points in the conventional coil igni- ated by the driver) specify desired settings.
tion system, for example. Gradually, how-
ever, more and more new vehicle systems Control units (ECUs)
were introduced which simply would not Control units process the information re-
have been possible without the use of elec- ceived from the sensors and setpoint gener-
tronics. The impetus for these new develop- ators using specific mathematical calculation
ments was provided by increasingly de- sequences (control algorithms). They con-
manding requirements placed on the trol the actuators by means of electrical out-
exhaust-emission characteristics of the put signals. Control units also form the in-
internal-combustion engine (e.g. emission- terface with other systems and the vehicle
control systems), on comfort and con- diagnostics.
venience (e.g. climate control and naviga-
tion systems) and on safety (e.g. antilock Actuators
braking system [ABS] and airbags). Actuators convert the electrical output sig-
nals from the control unit into physical
variables. Examples of actuators are:
Overview gasoline-engine fuel injectors
diesel-engine fuel injectors
Cars and commercial vehicles equipped with electric motors (e.g. for driving the
the latest available technical improvements power-window regulator mechanism or as
are absolutely brimming with electronic sys- throttle-valve positioner on vehicles with
tems. Those systems can be subdivided into ETC (Electronic Throttle Control))
the following areas of application: fans
Engine and drivetrain
Safety Networks
Comfort and convenience As the number of electronic systems in vehi-
Communication and multimedia. cles grows, so does the amount of intercon-
necting wiring needed. The total length of
Fig. 1 provides an overview of the electronic the wiring in the wiring harness of a mid-
systems that can be found on modern motor range car is now roughly 1.6 km on average
vehicles. Many of the systems referred to are and incorporates up to 300 connectors with
now standard equipment on all new vehi- a total of around 2000 connector pins.
cles. By contrast, there are others that re-
main (as yet) the preserve of the most ex- Networking of the various systems reduces
pensive luxury models. the overall length of the wiring required. A
shared data bus consisting of only two wires
Electronic systems can be subdivided into (e.g. CAN bus) carries data that is read by
the following functional areas: all bus users. Depending on the specific re-
Sensors and setpoint generators quirements, a vehicle may also have separate
Control units (ECUs) busses for engine and drivetrain, comfort
Actuators and convenience systems, and communica-
ECU communication links (networks) tion systems.
and
Electronic diagnosis
Robert Bosch GmbH
Motor vehicles are now inconceivable without electronic control units and their associ-
ated sensors and actuators. Such components have revolutionised the automotive
world. All essential vehicle functions are electronically controlled by systems and com-
ponents that occupy only a tiny space. There are systems that control the function of
engine and gearbox, the safety and security systems and a large number of comfort and
convenience systems, and the number of intelligent vehicle systems is growing at a
rapid pace. Nor is there any foreseeable end to this trend.
Electronic control systems open the door to a vast array of possibilities. They im-
prove vehicle safety and ride comfort. At the same time, they make vehicles more
economical and kinder to the environment.
The triumphant advance of electronics has created a vocabulary of terms with
which we are bombarded every day. It is becoming more and more difficult to under-
stand what precisely those terms refer to.
This publication in the Bosch Yellow Jacket Technical Instruction series explains the
specialist terminology of microelectronics in detail.
It starts with a short introduction followed by an explanation of the principles of
semiconductor technology and how microelectronic components interact. A practical
example of a circuit helps to consolidate the theoretical concepts described. This is
then followed by a description of the development and production of control units.
Finally, the glossary provides a quick guide to the most important microelectronics
terms.
The functions of the individual electronic systems (e.g. MED, EDC, ESP) are described
in detail in separate publications in the Yellow Jacket Technical Instruction series.
1
Dri
vet
An rai
tilo n
Tra ck br Dig
ita
Ele ction aking ga l eng
ctro co sy Safe sol in
nic ntro stem ty Die ine e e ma
sta l sy
b s n n
s (A die el en gine agem
He ility p tem BS) sel gin : M e
ad rog (TC wit co e: otro nt:
lam ram S) h ntr Ele ni
p a (E ele ol ( ctr c
d S c ED on
an justm P) fue troni C) ic
d c en l in call
lea t j y
W ni ele ecti co
c o n
Ind ash-w Lit ng (ga troni n, trolle
d
iss
Elek
olle es ity ba ste Ne BD
r A sur sys r m two )
bar
rea e m tem dia rk
tronik
Ne oni s gn
osi
im
Sy twor toring s
ste k ( Co
Kfz
m CA mm
dia N) un
Co gn ica
osi Ele
mfo
r s ctr tion
t an Vo nico
dc ic v
onv (sp e co oice
ee ntro ou
c
eni
Ad enc Au h re l of tput
ap Cru e dio c fu
of vehicle service.
Se He tiv is
ec ec Vid equ ognit nctio
Robert Bosch GmbH
at a at rui on
dju ing s tr On
eo ipm ion) ns
ent
stm an (rada e con ol
d (ra
Electronic diagnosis
dia . C
faults that occur (e.g. short circuits in the
osi
face. The system tester can also be used to
s gn AN
trol unit continuously monitor the opera-
Electronic diagnosis functions on the con-
osi
s )
nected to the control units diagnosis inter-
tion of the system and its components. Any
UAE0856E
can then be read out in the course of service
3 Gasoline-engine ECUs
a b
Fig. 3
a 1979 Jetronic ECU
with 290 compo-
nents and a weight
of 1.14 kg
UAE0857Y
b 1996 Motronic
hybrid ECU with
82 components and
a weight of 0.25 kg
Robert Bosch GmbH
Semiconductors
The electrical conductivity of semiconduc-
tors is somewhere between that of conduc-
tors and insulators. Under normal conditions
they also have a very small number of free
charge carriers, but that number can be sub-
stantially increased by the application of en-
ergy from an external source. Their conduc-
tivity, therefore in contrast to that of metals
Robert Bosch GmbH
and insulators is heavily dependent on: jacent atoms. Every silicon atom has four
pressure (affects the mobility of charge outer electrons (Fig. 1). Each pair of adjacent
carriers), atoms is linked together by two shared elec-
temperature (affects the number and trons. In such a perfect crystal lattice, there-
mobility of charge carriers), fore, there are no free charge carriers, which
exposure to light (affects the number of means that the silicon is a non-conductor.
charge carriers), and This condition is changed fundamentally
added impurities (affects the number and by the introduction of suitable impurities
type of charge carriers). (doping), or energy from an external source.
Siemens
cm
102
Silicon
51022 Si atoms/cm3
Si Si 101
4+ 4+
Electrical conductivity
1
p-type
Si 10
n-type
4+
102
T= 300 K
Si Si 103
4+ 4+
104
1014 1016 1018 1020 cm3
SAE0048E
SAE0047Y
Doping concentration
Robert Bosch GmbH
ing gaps are referred to as holes. In silicon, Material Max. operating temperature
those holes are mobile, and within an electri-
Germanium 90 ... 100 C
cal field they move in the opposite direction
Silicon 150 ... 200 C
to electrons. Holes therefore act like free pos-
Gallium arsenide 300 ... 350 C
itive charge carriers. Each boron atom intro-
duced therefore supplies one free, positively
charged hole. The silicon becomes positively An n-type semiconductor always has some
conductive (Fig. 2) and is therefore referred holes and a p-type semiconductor always
to as p-type silicon. has some free electrons. Such minority
charge carriers are fundamental to the way
The conductivity of n-type silicon is around in which almost all semiconductor compo-
10 times as high as that of p-type silicon as- nents work (refer to section entitled Elec-
suming other factors are equal (e.g. number tronic Components).
of foreign atoms, temperature, pressure) be-
cause electrons can move about 10 times as p-n junction
fast as holes. The boundary between the p-type and
n-type regions of the same semiconductor
Intrinsic conductivity crystal is called the p-n junction. Its proper-
The application of heat or light can generate ties are fundamental to almost all semicon-
free charge carriers even in undoped silicon. ductor components.
They consist of electron-hole pairs and make
the semiconductor intrinsically conductive. p-n junction without external voltage
Such conductivity is generally low compared In the p-type region there are a large number
with that produced by doping. The number of holes () and few free electrons. In the
of electron-hole pairs increases exponentially n-type region, by contrast, there are ex-
with rising temperature and ultimately erases tremely few holes and a large number of free
the electrical differences between areas of electrons () (Fig. 3). Due to the concentra-
p-type and n-type silicon created by doping. tion differentials, the mobile charge carriers
Consequently there are maximum limits for in each region diffuse into the other region in
the operating temperatures of semiconductor each case (diffusion currents). As a result, the
components, as shown in the following table: p-type region is negatively charged and the
a b
Boundary zone Boundary zone
p-type region n-type region p-type region n-type region
Fig. 3
a Diffusion of mobile
charge carriers +
b Formation of
depletion layer Depletion layer
SAE0858E
n-type region positively charged. A potential broken down. When the diffusion voltage is
difference (diffusion voltage) is thus created exceeded, the charge carriers flood the
between the p-type region and the n-type p-n junction and a large current flows in
region which counteracts migration of the forward direction.
charge carriers. This brings the equalisation
of holes and electrons to a halt. As a result, a Breakdown voltage
region that is deficient in mobile charge car- The breakdown voltage is the reverse-direc-
riers, and therefore has poor electrical con- tion voltage above which a small increase
ductivity, is created at the p-n junction. It is in voltage brings about a steep rise in the
referred to as the depletion layer or space- reverse current.
charge region. Due to the diffusion voltage, The cause of this effect is the release of
the depletion layer has as strong electric field. bound electrons from the crystal lattice in
the space-charge region due to the high field
p-n junction with external voltage strength (Zener breakdown) or due to
If an external voltage is applied to a surges of accelerated electrons. The acceler-
p-n junction, it produces the effects ated electrons strike other electrons, break-
explained below (Fig. 4). ing them free of their bonds and starting an
avalanche-like increase in the number of
Reverse bias charge carriers (avalanche breakdown
If the negative terminal is connected to the or first breakdown). Both effects are re-
p-type region and the positive terminal to versible. The nature of the cause and the
the n-type region, the width of the space- level of the breakdown voltage are depen-
charge region increases. Consequently, cur- dent on the doping concentration profile.
rent flow is largely inhibited apart from a
very small residual current (reverse current) A second breakdown occurs if there is lo-
produced by minority charge carriers. calised heating of a semiconductor compo-
nent caused by current constriction so that
Forward bias the area concerned becomes more conduc-
If the positive terminal is connected to the tive. This results in a self-accelerating in-
p-type region and the negative terminal to crease in current and leads to the destruc-
the n-type region, the depletion layer is tion of the semiconductor component.
a b
+ +
+
Flow of holes
SAE0754E
Flow of electrons
Reverse voltage Forward voltage Fig. 4
a Reverse bias
b Forward bias
Robert Bosch GmbH
Electronic components
Electronic devices are made up a large num- substantially increased by the use of other in-
ber of components that can be subdivided sulating materials. The amplification factor
into two main categories: passive compo- is referred to as the relative permittivity or
nents and semiconductor components dielectric constant r. Since, in many cases,
(active components); the latter can be very large plates would be required for the
further subdivided into subcategories. frequently used capacitance levels, capacitors
are normally made by winding long strips
into a coil (wound capacitors) or packing
Passive components a large numbers of small plates together
(multi-layer capacitors). The electrolytic
Ohmic resistors, capacitors and inductances capacitor uses a thin layer of oxide as the
are classed as passive components. dielectric. This method makes it possible
to manufacture small capacitors with large
Ohmic resistors capacitances.
Ohmic resistors are generally made of ma-
terials with conductive properties similar When direct current is applied to a capaci-
to metals e. g. carbon (carbon-film resistors) tor, it stores up charge until the limit of its
or special metal alloys (metal-film resistors). capacitance is reached; at that point current
They are constructed so as to reduce to the can no longer flow. Thus for a certain
absolute minimum the effect of voltage, period at least the capacitor stores electri-
current and temperature on their electrical cal energy that is then available if the exter-
resistance. The conductor has a very small nal power supply fails. This effect is utilised
cross-section relative to its length, this being in the circuit for triggering the airbag, for
achieved either by the application of thin instance. Even if the wires connecting the
films to insulators or by winding wire into a airbag triggering unit to the vehicles battery
coil. are severed in a serious accident, the capaci-
tor still holds sufficient electrical energy to
In electronic circuits, resistors limit the cur- ensure that the airbag is deployed.
rent flow I or produce a voltage U propor-
tional to the current. The resulting energy In alternating-current circuits, a capacitor
loss is converted into heat. An extreme ex- has a similar effect to a resistor. Its resistance
ample of this is a light bulb, in which an is dependent on the capacitance and the fre-
ultra-fine tungsten filament is heated to quency of the alternating current. As the fre-
such a degree that it starts to glow. quency decreases, the resistance increases. If
the frequency is reduced to its lowest limit,
The resistance in a circuit is referred to by i.e. zero (direct current), the resistance is
the symbol R and its unit of measurement is equal to infinity and, consequently, no cur-
the ohm (). rent flow is possible. This relationship is
utilised, for example, by frequency filters in
Capacitors speaker systems in order to filter out the
The most simple type of capacitor consists high-frequency sounds from the lower-
of two parallel metal plates that are not in frequency ones (high-pass).
contact with one another. The area and sep-
aration of the plates as well as the medium The capacitance in a circuit is referred to by
separating them (dielectric) determine the the symbol C and its unit of measurement is
quantity of charge carriers that can be stored the farad (F).
by the capacitor (its capacitance). Using a
vacuum (or air) as the dielectric offers the
lowest capacitance. The capacitance can be
Robert Bosch GmbH
lithic literally meaning made from a single In this case too, the methods of production
stone from the Greek monolithos mean- and the type of use differ substantially from
ing single stone). conventional active components.
1 Diode characteristics
a b
reverse direction forward direction
4
Forward current
Current in
0 3
0.7 V
2
Fig. 1
a Characteristic of 1
diodes, e.g. rectifier
diode, Zener diode 0 UAE0886E
and Schottky diode 0 0 100 200 300 mV
Voltage in re- in forward
b Section of tunnel-
verse direction direction Forward voltage
diode characteristic
a b
UF +-
+
+
Fig. 2 U U1 RL Ug U U1
- + +
a Half-wave rectifier
- RL Ug
b Bridge rectifier
-+ - -
UF Diffusion voltage
(approx. 0.7 V)
U~ AC voltage UF U1 2 .UF
U1M U1M
U1 Transformed
AC voltage UgM Ug UgM Ug
Ug Rectified voltage
UAE0897Y
RL Load resistance t t
U U
U1M Amplitude of U1 U1
UgM Amplitude of Ug
t Time
Robert Bosch GmbH
Transistors
G Gate
* Bulk connection
B B S D S D S D S D S D S D
* * * * (connection desig-
nation not usual)
Robert Bosch GmbH
N P +++ - -
N
lar, voltage-controlled component.
UAE0458-1E
+++ ---
N n-type silicon +++ ---
-
P p-type silicon EB BC
E Emitter
Method of operation of a MOS FET
Emitter Base Collector
B Base (with reference to a MOS FET with p-type
C Collector enrichment): while no voltage is present at
Robert Bosch GmbH
N silicon
UGS Voltage between
If PMOS and NMOS transistors are cre- G and S
ated in pairs on the same silicon chip, they PMOS transistor NMOS transistor UDS Voltage between
are referred to as CMOS transistors (comple- D and S
mentary MOS transistors, Fig. 6 c). Their
particular advantages are very low power Thyristors
loss, high interference immunity, TTL-com- Thyristors are semiconductor components
patibility (transistor-transistor logic in elec- with at least three p-n junctions (one of
trical circuits), low power-supply voltage and which may also be replaced by a suitable
suitability for analog signal processing. metal-semiconductor contact) that can be
The low power consumption, particularly switched from a reverse-bias condition to a
for digital circuits, is due to the fact that forward-bias condition (or vice versa). The
current only has to flow when a digital cell term thyristor is used as the generic term for
changes ( e.g. switches from 0 to 1). Re- all types of component which conform to
tention of the information requires only that that definition. It is a contraction of the two
a voltage is applied. Since, with this configur- words thyratron (gas-filled tube triode) and
ation, the PMOS and NMOS transistors are resistor.
connected to one another, the leakage cur- Uses in power electronics: speed and fre-
rent is minimal. More than 80 % of ICs are quency control, rectification and conversion,
now manufactured using CMOS technology. switching.
Robert Bosch GmbH
4
I
II duction process. It involves the creation of
Pentium
1M 1M multiple p and n-type layers in a multi-stage
486DX
256 doping process using a pre-doped single-
64
386DX crystal silicon disc or wafer.
100 K
An integrated circuit IC (Integrated Circuit,
16 8088
4
see section entitled Manufacture of Semi-
10 K
8080 conductor Components and Circuits) does
1K
8008 not contain any separate (discrete) com-
1K
ponents but rather switching elements or
b 100
functional elements.
C
8008 Degree of integration
10 The degree of integration is defined as the
8080
number of functional elements, transistors
US cents per transistor
1
8088 or gates on a single chip. The following cate-
gories are defined based on the degree of
integration (and chip surface area):
386DX
0.1 SSI (small-scale integration): up to
486DX around 1000 elements per chip, average
Pentium I chip area 3 mm2 (varies considerably
0.01
II depending on level of power loss)
III MSI (medium-scale integration): up to
0.001 around 10,000 elements per chip, average
chip area 8 mm2
c 10
m
8008: 10 m Pentium I = 0.8 m LSI (large-scale integration): up to
II = 0.35 m
III = 0.18 m 100,000 elements per chip, average chip
8 area 20 mm2
VLSI (very-large-scale integration): more
than 100,000 elements per chip, average
6 8080: 6 m
chip area 30 mm2
Structure size
Fig. 7
a Increase in number of 4 The number of functional elements on
transistors per chip in 8088: 3 m
VLSI chips is constantly increasing. Micro-
memories and processors can now have up to 10 million
processors (Gordon 2
386DX: 1,5 m
transistor functions per chip. Memory mod-
Moores Law) I
486DX: 1 m II III ules (DRAM) can even contain several hun-
b Fall in cost of transis-
dred million transistors. The number of
UAE0759E
tors
0
c Decreasing size of
1970 1980 1990 2000 such functional elements is subject to an
structural units Year exponential increase over time. This fact
of new processors was first recognised by Gordon Moore
Robert Bosch GmbH
(co-founder of the company Intel) and is ex- Apart from a few exceptions, ICs are now
pressed by Gordon Moores Law (Fig. 7a): always based on MOS or combined technolo-
Elements per chip = 2(year 1956) 2/3 gies, and for that reason bipolar ICs are not
This law states that the number of func- described in any further detail at this point.
tions per chip will double every 18 months.
Since the structural units in which the chips Analog circuits
are manufactured are becoming smaller at Analog circuits are required wherever elec-
the same time as the number of functional tronic equipment has to communicate with
elements is growing, the size of the ships al- the outside world which is always analog.
ters very little and the costs per transistor are This applies in particular to the areas of
similarly dropping exponentially (Fig. 7b). preparation and processing of input and
Thus the structural units of new processors output signals. Examples of analog-input
have shrunk over a period of 28 years from signal generators are microphones (tele-
10 m in 1972 to 0.18 m in the year 2000 phone) and sensors. Examples of analog-
(Fig. 7c). output signal receivers are speakers or actua-
tors.
The degrees of integration LSI and VLSI Figure 8 illustrates the stages of develop-
demand methods such as CAD (computer- ment of an analog IC from conception to
aided design). VLSI circuits in particular can installation in an ECU.
only be created with the help of highly so-
phisticated programs which can convert en- 8 Development of an analog IC from conception
tire function blocks into corresponding cir- to installation
cuit subdivisions. Such program systems are
called HDL (Hardware Description Lan- 1 M
Vcc
On/off
IC classification 2
There are various systems for classifying ICs,
although there are invariably mixed cate-
gories within each system.
Classification by method of production:
3
Bipolar/Unipolar (MOS);
mixed category: e. g. BiCMOS, BCD
Classification by function:
Analog/Digital;
4
mixed category: e. g. mixed-signal IC
Classification by application:
Standard IC/ASIC
mixed category: e. g. ASSP
5
UAE0013Y
Robert Bosch GmbH
10 Mixed-signal IC
Sensing Acting
Analysing
UAE0763E
Robert Bosch GmbH
Microcomputers
(CPU) for variable data flash EPROM) (EEPROM)
For programs and
permanent data records
Arithmetic and Logic Unit Memory capacity
(ALU) Memory capacity Memory capacity 32 bytes
4-, 8-, 16-, 32-bit 64 bytes 32 kbytes 2 kbytes 512 kbytes 1 kbyte
(Timer, time SPI,
processing unit, CAN)
input capture,
Robert Bosch GmbH
Electronic components
output-compare
register)
Communication
Monitoring Resolution Resolution with external
circuit 50 ns 8 10 bit Data rate chips via
(watchdog) Counter Time range 4 32 8 32 200 bit/s address/
8 64 bit 50 ns 1s channels channels 1 Mbit /s data bus
UAE0454-1E
Semiconductor components
25
Robert Bosch GmbH
The chief components of a microcomputer ation. For that reason, it is also referred to as
are generally separate modules connected to a single-chip microcomputer.
one another on a printed-circuit board. For The microcontroller is used to control
simpler tasks such as are more and more fre- self-regulating systems such as an engine-
quently required for Internet access on wire- management system.
less communication devices, single-chip At present, there is a choice of 4-bit, 8-bit,
computers are becoming increasingly com- 16-bit or 32-bit microcontrollers for the
mon. They integrate the functions detailed various types of application. Depending on
above on a single silicon chip (System-on-a- application, they may also have expansion
Chip). The relatively small amount of RAM modules connected to them (e.g. additional
that can be accommodated on a chip at rea- memory for data and program code).
sonable cost limits the performance capabil- The program that is run by the CPU is
ities of such highly integrated systems. permanently fixed in the read-only memory
and is not changed for different appli-
Microprocessor cations. This distinguishes the single-chip
A microprocessor is a central processing microcontroller from the PC.
unit (CPU) in the form of an integrated
circuit on a chip. The concept of the micro- Transputer
processor avoids individualisation despite A transputer is a special type of micro-
the high degree of integration, and enables processor that is particularly suited to the
adaptation to the multiplicity of practical construction of parallel computer networks.
demands by reliance on programming. A In addition to the usual components of a
microprocessor is not capable of function- microprocessor, it has communication and
ing on its own; it is always part of a micro- process-handling hardware on a chip.
computer. At present, there are 16-bit, 32-bit
and 64-bit microprocessors. Processors are Programming
subdivided into two main categories: The only command form capable of direct
PCs use CISC (Complex Instruction-Set interpretation by a microprocessor is a bit
Computer) processors. These are ex- pattern, i.e. the binary representation of a
tremely versatile and permit unrestricted number. Since, however, this form of in-
programming. struction is not easy to work with for a pro-
Workstations normally use RISC grammer, and is therefore susceptible to
(Reduced Instruction-Set Computer) errors, easily memorable abbreviations
processors. These perform specific tasks (mnemonics) are used. These are automati-
such as are frequently required on work- cally translated by an assembler program
stations much faster, but are distinctly into bit patterns (machine code) that can be
slower for all other tasks. understood by the microprocessor. Micro-
computers for single-purpose applications
Microcontrollers are programmed in the assembler language
The microcontroller is a component that in- specific to the processor.
corporates the following elements on a chip: For more complex systems and programs,
CPU, high-level programming languages such as
random-access memory (RAM), C are needed, as otherwise it would be im-
peripheral modules (input/output, inter- possible to keep extensive programs man-
rupt, timer, serial interfaces) and ageable and free of errors. Such languages
an optional read-only memory (ROM). require sophisticated translation programs
(compilers) which convert the text of the
With these integrated components, the mi- high-level language into a form that can be
crocontroller is capable of stand-alone oper- processed by the processor.
Robert Bosch GmbH
Semiconductor memories
Programmable Programmable
on a program- Static Dynamic
in the circuit memories memories
ming device
UV Electrically
Read-only erasable erasable
UAE0465-1E
Memories can be organised on the basis of a with two transistors, of which either the one
variety of word lengths. The way the mem- (logical 1) or the other (logical 0) con-
ory is organised is generally indicated by a ducts at any one time. In SRAM, the infor-
multiplication sign, e.g. an 8 M 8 RAM or mation remains stored until the storage cell
8-megabyte RAM concerned is addressed and overwritten, or
8 million bytes the power supply is switched off.
= 64 million bits Static RAMs are currently available in
= 64 Mbit sizes up to 16 Mbit. Although considerably
more expensive than dynamic RAMs, be-
That means that the memory has eight data cause they can be written into and read sig-
inputs and eight data outputs at which the nificantly faster they are used as high-speed
eight bits of one of the eight million stored access computing memories (cache mem-
data words are simultaneously present. Its ory) for the CPUs in computers. Typical
word length is thus 8 bits. Since all memory read-write times are currently less than
specifications are based on the binary sys- 30 ns.
tem, the exact number of bits is as follows:
64 Mbit Dynamic RAM (DRAM)
= 8 Mbyte 8 bit The information in a DRAM is stored as an
= 223 8 bit electrical charge in the gate capacitor of a
= 226 bit CMOS transistor which either passes (logi-
= 67,108,864 bits cal 1) or does not pass (logical 0) a cur-
rent as a result. As such capacitors are sus-
Static RAM (SRAM) ceptible to leakage, the charge is gradually
Static RAMs use bistable switching elements lost. In order that the information (charge)
as the data storage cells. Their function is is retained, the memory has to be refreshed
similar to that of a flip-flop, a simple circuit at regular intervals (every few ms).
75,000 5,000
400
120
30
5 0.5 0.13 0.06 0.03
1973 1977 1981 1984 1987 1990 1995 1998 2001 2005
1 Local 1 Stick of 1 Paper
Stamp phone chewing clip
UAE0773E
DRAMs are currently available commer- possible with special equipment at relatively
cially in sizes up to 256 Mbit (on one chip), high cost.
and in laboratory conditions up to 1 Gbit
(gigabit = 109 bit). Since every storage cell EEPROM (Electrical EPROM)
consists of a transistor and a capacitor, this EEPROM is also referred to as E2PROM (E
requires more than a billion transistors on a squared PROM). This type of erasable read-
single chip. only memory can be electrically wiped and
DRAMs are volatile memories. This reprogrammed. The wiping and reprogram-
means that they lose their information when ming operation can be performed either in a
power is switched off. separate unit or in situ. Every storage cell of
DRAMs are very widely used nowadays as an EEPROM be individually overwritten.
system memories for all types of computer. For that reason, this type of memory mod-
The fall in price for 1 Mbit of memory over ule can also be used as non-volatile data
the last few decades has been so dramatic memory (e.g. for status information in
that ample memory can now be fitted in any engine-management systems).
PC without substantially affecting its price
(Fig. 13). Flash EEPROM
Yet another variation of EPROM and
Read-only memory EEPROM is the flash EEPROM. In this case,
Read-only memory (ROM) is permanent- electrical flash pulses are used to erase spe-
storage memory that allows any memory cific storage areas or the entire contents of
location to be accessed directly but as the the memory. The erased areas can subse-
name indicates allows the information quently be reprogrammed.
only to be read and not altered. The flash memory can be reprogrammed
A ROM is a non-volatile memory, i.e. the on a programming station. However, the ad-
information it contains is retained even vantage of flash EEPROM is that is can also
when the power supply is switched off. It is be reprogrammed while still inside the
usually used to store program codes (control sealed control unit. When this is done, the
programs) and fixed data (function tables, memory area which contains the program-
encoding rules, character generators, engine ming routines must not be erased, or alter-
characteristic-data maps) that need to be re- natively, the programming routines must be
trievable at any time. The information may transferred to the RAM before the memory
be indelibly entered in the memory either by is wiped. The microcontroller then works
the manufacturer (in one of the last stages with the RAM as the program memory.
of production) or by the user by appropriate Flash EEPROM is used wherever relatively
programming of specially prepared memo- large volumes of data need to be stored but
ries (PROMs or programmable ROMs). also have to be changed from time to time
(e.g. in mobile phones, digital cameras and
Erasable ROM as program memories in electronic control
There are also ROMs whose contents can be units in motor vehicles).
erased and reprogrammed as outlined
below.
k 0.5
mA Light intensity E
12 5,000 lx
0.4
4,000 lx
Photoelectric current I
10
Resistance R
0.3 3,000 lx
8
6 0.2 2,000 lx
4
1,000 lx
0.1
2 500 lx
UAE0761E
UAE0917E
0
200 400 600 800 1,000 lx -0.2 -0.1 0 20 40 60 80 V
Light intensity E Reverse voltage U
Robert Bosch GmbH
of colours available (infrared, red, yellow, expensive, flat screens are manufactured us-
green, blue). ing thin-film techniques (vapour deposition
through a screen). The advantages of flat
LEDs are used for numerical and alphabeti- screens over cathode-ray tubes are their
cal displays as well as for warning lamps. slimmer dimensions, lower power consump-
Since it has been possible more recently to tion and the absence of flicker.
substantially increase the amount of light
generated, LEDs are now also used in motor Charge-coupled device (CCD)
vehicles to illuminate the instrument panel CCDs are integrated circuits arranged as a
or as a third (high-level) brake light. They charge-coupled array. They are used to
have the advantage of a relatively high light record images in video and digital cameras
yield and, as a result of their very long ser- as well as in scanners. Thousands or even
vice life, do not have to be replaced. millions of individual elements laid out in a
matrix pattern (array) are created on a chip
Laser diode (Fig. 16). Digital cameras can record up to
With the appropriate mechanical/optical 3.5 million pixels with a colour resolution of
design, LEDs can also be made to emit laser 16 million colours. CCDs are not only capa-
light, i.e. a parallel beam of monochromatic ble of converting the visual information into
(of a single wavelength) and coherent (the electrical signals, they can also store the data
waves are in phase) light. Such properties are until it can be transferred to an external
required for the transmission of data by storage medium. To do this, the output
fiber-optic cables. Laser diodes were there- register (5) is scanned at a high rate.
fore a necessary requirement for the high
data flow rates such as occur in computer
networks with access to the Internet or the
transmission of television signals by cable.
Phototransistor
A similar effect to that used by the photo-
diode also occurs in the phototransistor. In-
cident light striking the base alters the con-
ductivity between the emitter and the collec-
tor in such a way that a current proportional
to the intensity of the light is able to flow
through the transistor. This type of compo- 16 CCD, image-sensor structure
B2 3 CCD array
screen raster in a cathode ray tube). As the 5
4 Row clock pulse
areas required are generally substantially B1/B2
larger than the silicon slices available, and 5 Output register
such slices would, in any case, be much too 6 Video output
Robert Bosch GmbH
IH Hall-effect current
RH Hall-effect constant IH
UH Hall-effect voltage IV Supply current
d Thickness of the B Flux density of the magnetic field
plate d Thickness of the plate
Robert Bosch GmbH
UAE0817Y
there are resistors which change their resis-
tance in response to physical deformation
(piezoresistive effect). By measuring the re-
sistance, the pressure can be calculated with
the aid of the integrated bridge circuit. Acceleration sensor
Finger-like structures are etched into the
Pressure sensors are used in motor vehicles surface of a silicon ship (Fig. 19). Every
for such tasks as measuring the intake mani- other finger is rigidly attached to the chip,
fold pressure and the atmospheric pressure. while the alternate fingers are only attached
at one end so that they can oscillate freely at
the other. These fingers are connected up as
a multi-layer capacitor. If the sensor is accel-
erated in a particular direction, the distance
18 Pressure-sensor cell with reference vacuum on the between the fixed and the movable fingers
structural side
changes, thus altering the capacitance. That
change can be measured as an electrical sig-
nal from which the acceleration can then be
calculated.
UM Measurement
voltage
Strain resistors R1
(compressed) and R2
(stretched)
Robert Bosch GmbH
Microcontrollers
Microcontrollers are single-chip computers of such control systems was that their relia-
that are used for event-driven control sys- bility was heavily influenced by manufactur-
tems or for controlling technical processes. ing tolerances, environmental conditions
Modern-day life is no longer conceivable and wear.
without them and they are to be found
wherever control systems for electrically In the 1960s, Intel, followed shortly after-
operated devices are required. wards by Motorola, developed the first
microprocessors initially for use in cash
registers. Following the development of inte-
Microcontroller grated circuits and the first microprocessors,
developments the inexorable progress of digital technology
began in earnest in the 1970s. For the first
Five billion microcontrollers were produced time, it was now possible to eliminate a large
in 1999, that is roughly equivalent to one for number of interference factors.
every person on the Earth. Its triumphant
march into every aspect of our day-to-day The end of the 1970s saw the dawn of the
lives is clearly demonstrated by the number age of the single-chip microcomputer.
of microcontrollers produced in the last ten Prominent representatives of that period
years (Fig. 1). were the Intel 8048 and Motorola 6800 mi-
A microcontroller consists of the inte- crocontrollers, the structure of which is still
gration of a CPU (central processing unit) the basis for many of todays microcon-
with memories and peripheral circuits on a trollers. In the succeeding period, advances
single chip. These are referred to as embed- in high-integration technology allowed
ded systems. Frequently such systems incor- more and more peripheral circuits to be in-
porate a microprocessor that was originally corporated in the chip. These were the be-
developed for computer applications and
subsequently adapted and modified so as to 1 Unit production figures for microcontrollers
5
data exchange with their environment, low
power consumption, high data density and 4
low production costs.
3
What then is the actual origin of microcon-
trollers? Even before the transistor was in-
2
vented, most of the devices in which micro-
controllers are now used were already in ex-
1
istence cars, telephones, radios, televisions
and household appliances. The difference
was that for most applications the control 0
1990 1992 1994 1996 1998 2000
systems were mechanical and only in very
SAE0879E
Year
rare and very costly exceptions were func-
tions controlled by electron tubes and relays.
In addition to their high cost, the problem
Robert Bosch GmbH
ginnings of the microcontroller as we know tinually growing, simple 4-bit designs are
it today. still used in large quantities (Fig. 2).
At the end of the 1980s, the term embedded To make it worthwhile to develop a stand-
control system was introduced to distin- alone microcontroller design, the produc-
guish them from infinitely programmable tion figures should run to at least 1 million
systems such as the PC. Embedded systems units a year. The unit price will then be
are designed exclusively for a specific appli- between 0.5 and 15 Euro.
cation so that the entire program together Applications that hold the promise of
with the vehicle-specific data in a ROM or such production volumes include control
EPROM are a permanent part of the prod- systems for automotive applications, com-
uct that cannot be altered by the user. munications devices, computer peripherals,
domestic appliances and entertainment sys-
Most microcontrollers sold today have an tems. In addition, microcontrollers are also
8-bit CPU. Although the numbers of more popular in industrial electronics, automa-
powerful 16 and 32-bit controllers are con- tion systems and testing and measuring
equipment, although recourse is generally
2 Proportion of microcontroller production volume made to standard products.
by processor type (1999 figures)
70%
60%
50%
40%
30%
20%
10%
0%
SAE0880E
Microcontrollers
Robert Bosch GmbH
Input and
Bus Arithmetic Bus
output units,
logic unit ADC/DAC,
(ALU) timer,
serial interfaces
Data and
program Bus
memory
Bus Bus
Program
Interrupt
monitoring
system
unit
SAE0881E
Robert Bosch GmbH
Peripherals
Instruction Arithmetic (I/O)
register logic unit (ALU)
ory-register architecture is derived from the Number of bits available for encoding
location of the operands. Different operations require very widely
varying numbers of bits for encoding in-
Register-register architecture structions of different levels of complexity.
This last architecture category also called For example, the null operation NOP (No
load-store architecture addresses both Operation) requires no operators (no ad-
operands of an operation explicitly in the dresses) nor any additional bits for defining
registers. Before any mathematical opera- the precise function. The operation MAC
tion, therefore, both operands first have to (Multiply & Accumulate: A = A + (B C)),
be loaded into a register. The result is then on the other hand, requires three operators
copied back to the memory. and additional bits for encoding the func-
tion, such as the behaviour in the event of a
Operand addresses mathematical overflow (saturation addi-
Another important distinguishing feature tion), identifying the data format (Q for-
among instruction-set architectures is the mat), etc.
possible number of implicitly and explicitly
encoded addresses. The operation C = A + B Decoding complexity
requires three addresses: Instructions of differing lengths are difficult
the address of operand A, to decode. First of all, the length of the cur-
the address of operand B, and rent instruction has to be established. Then
the address of the result operand C. there is a check as to whether the instruction
has been completely read. Both operations
Instruction-set architectures which allow generally require several clock cycles and
free choice of these three addresses (possi- limit the possible processing speed to a sub-
bility of encoding three addresses) are re- stantial degree.
ferred to as non-destructive instruction-set
architectures. But as three addresses nor- Length of instructions
mally occupy too many bits in the instruc- There are currently three different concepts
tion code, many architectures use an implicit for the length of instructions.
method of addressing.
With implicit addressing, one of the ad- Fixed-length instruction set: every instruc-
dresses of the two source operands is also tion is the same length, e.g. always 16 bits.
used as the destination address. Thus, the Multiple instruction sets: it is possible to
address of one of the source operands is alternate between two different fixed-
used to store the result of the operation, and length instruction sets, e.g. 16 bits and
that means that the operand in question is 32 bits. The first bit of each instruction
overwritten, i.e. destroyed. That destruc- then indicates the format.
tion has led to the adoption of the term Multiple instruction length: the length of
destructive instruction-set architecture. the instructions varies according to com-
plexity, e.g. the NOP command can have a
Requirements placed on instructions length of 16 bits and the intersegment
Instruction-set architectures differ from one jump a length of 32 bits.
another not only by virtue of their operand
memories and operand addresses, but also Instruction set
on the basis of the length of their instruc- The complete instruction set of a microcon-
tions. The following two very different re- troller enables it to execute any expression of
quirements must be met in this connection: a higher-level programming language (e.g.
C) by a sequence of instructions. The in-
structions are classified as follows:
Robert Bosch GmbH
Data processing operations: mathematical, CPU. This involves first calculating its
logical and conversion instructions. address, which is required in order to load
Control instructions: jumps, comparisons, the instruction into the CPU.
etc. Decode-instruction phase (2): the instruc-
Input/output instructions: instructions for tion transferred to the CPU is analysed.
inputting and outputting data. The relevant function is then initiated.
Memory instructions: instructions for Fetch-operand phase (3): the operands are
reading and writing data from/to the fetched from the registers or memories to
memory. the executing unit (e.g. ALU).
Execute phase (data operations) (4): once
The bits available in an instruction for ad- the operands have been transferred to the
dressing the operands/result are limited. executing unit, the data operation speci-
Embedded applications, however, require fied by the instruction is executed.
ever larger programs and data volumes, Store-operand phase (5): the result of the
which leads to larger and larger numbers of data operation is written back to a register
addresses. As a result, the addresses become or memory in this phase (which is there-
longer and longer and encoding them be- fore also called the write-back phase).
comes more and more complex. There are a
number of different methods for solving this Execution of instructions on older
problem. microcontrollers
On older microcontrollers, instructions are
Instruction execution processed purely sequentially (CISC archi-
Execution of instructions on modern tecture: Complex Instruction-Set Com-
microcontrollers puter). This requires a complete clock cycle
The various phases of instruction execution for each phase:
are illustrated in Figure 3. Phase 1: calculating the next instruction
Fetch-instruction phase(1): before an in- address and reading the instruction
struction can be executed, it has to be Phase 2: decoding the instruction
fetched from the program memory to the
Indirect Indirect
operand result
1 3 addressing 5 addressing
Fetch Fetch Store
instruction operand operand
2 4
Calculate Decode Calculate Data Calculate
address instruction address operation address
SAE0883E
Next Vector
Fig. 3 instruction operation
The numbers indicate
the five phases of execu-
tion of an instruction
Robert Bosch GmbH
Memory-mapped peripherals
(memory-mapped I/O)
Peripherals and memory modules share a
common address area. The advantage of this
is that the large number of instructions for
addressing memories can also be used for
peripheral modules. However, this uses up
valuable address space which can be disad-
vantageous with 4-bit and 8-bit microcon-
trollers.
Peripheral modules: Operating mode fahrzeug [Open Systems and their Interfaces
Another characteristic for distinguishing be- for Electronics in Motor Vehicles]).
tween I/O modules is the mode of opera-
tion, which can be one of the following: Interrupt-driven I/O
The peripheral module processes all input
Programmed I/O and output operations independently and
The I/O module is controlled directly by the signals to the CPU via a dedicated line
CPU. The microcontroller controls all func- (interrupt line) when new information is
tions by means of a program. In this case, received or when an action by the CPU is
the CPU has to wait while the I/O unit per- required. The basic advantage of this is that
forms an operation. This mode of operation the CPU and the peripheral modules can
is therefore only used with microcontrollers work in parallel. The CPU program only has
that exclusively perform input and output to be interrupted when the peripherals need
operations (I/O processors). assistance. Checking of the programs for er-
rors, however, is extremely time-consuming
Polled I/O because all interrupt constellations have to
The I/O module is capable of performing be checked. This architecture supports
independent operations and the input/out- event-driven operating systems such as
put data is temporarily stored in special Windows CE, OSEK or OS9.
buffer memories. The CPU periodically
checks the module status and transfers new DMA (Direct Memory Access) I/O
data as required. This mode of operation is With this mode of operation, the I/O mod-
primarily suited to RISC microcontrollers as ules can exchange data directly with the
they only have a software interrupt system. memory, without the involvement of the
Although the periodic scanning of the CPU. This mode is primarily used for high-
I/O module demands a large proportion end microcontroller systems. As with inter-
of the CPU resources, polled I/O is ideally rupt-driven I/O, this mode of operation re-
suited to time-based operating systems quires hardware that prioritises all pending
(time-triggered architecture or TTA), requests.
e.g. OSEK+ (Offene Systeme und deren
Schnittstellen fr die Elektronik im Kraft-
Data register
error detection)
Data storage
environment
Sensors
Status register
Address bus
Dual-ported
data memory
Harvard architecture
Where a microcontroller has two busses
(for data communication and transfer of
instructions), it is said to have a Harvard
architecture (Fig. 5b).
Robert Bosch GmbH
Circuit example
The interaction between a circuit and mi- Development environment
crocontrollers (C) is best explained by Program development
means of an example. The circuit used in The user program is created on a PC using a
the example chosen is referred to as a devel- text editor (source code, Fig. 2a). A compiler
opment circuit because it is used to test user then translates the source code from the
programs developed on a PC. Such circuits programming language in which it is written
are sometimes also called evaluation (e.g. C++) into an assembly language based
boards. on mnemonics. The assembly language is
the preliminary stage to the machine lan-
guage and is still easy to read. An assembler
Overview is then used to convert the assembly lan-
guage into a digital code that can be under-
A development circuit is part of a develop- stood by the development circuits micro-
ment system used in conjunction with a PC. controller. The precise memory addresses
Figure 1 shows the overall layout of a devel- are specified by a linker. The machine code
opment workstation with the circuit de- or object code can then be transferred to
scribed in this example. the circuit by the PC.
The advantage of a development circuit is
that all components are easily accessible. In Assembler programs can also generate list-
addition, the signals of the various inputs ing files in which each line of the original
and outputs can be tapped at the terminals. source code is shown alongside the corre-
In that way individual signals can easily be sponding object code. This is useful for
checked. locating errors.
8
1
Fig. 1
5 6
1 Circuit signal on the 2
oscilloscope
3
2 Output indicators
4
(LEDs)
3 Development circuit
4 Connection to PC
(connector X2)
5 19-inch module rack
system with power
UAE0846Y
supply unit
6 Circuit layout
7 Circuit diagram
8 Example program
Robert Bosch GmbH
Hardware configuration
Following download, the user program is
stored in the external RAM module. Nor-
mally this type of memory module is used
only for data storage. In development cir-
cuits, however, the components are con- b Transfer 2
nected in such a way that the RAM can op- 3
Object 4
erate as a program memory as well. This code 5
makes it possible to download the user pro- 7
6 Fig. 2
gram from the PC and store it on the devel- a Program creation
opment circuit without having to alter the b Transfer of object
actual program memory (the EPROM). c code
g 8000 6
c Starting and running
Starting the user program Start the program
command d Stopping the
To start the user program, the monitoring 8
6 program
program that the microcontroller runs has
to branch to the RAM. The start command 1 PC
and start address are entered on the PC and d 2 Development circuit
transferred to the development circuit (g 3 Microcontroller
address; e.g. g 8000, i.e. go to address 8000). 9 4 EPROM with
monitoring program
The monitoring program reads this infor-
5 RAM
mation and then executes the instruction
SAE0847E
6 Interface module
to go to the user program (Fig. 2c). 7 Connector X2
8 Connector X1
9 Reset button
Robert Bosch GmbH
The EPROM can have its contents erased Output modules (IC8, IC9)
by irradiation with UV light and can then be Addresses and data are outputted via the
reprogrammed using a programming 74HCT245 bi-directional power drivers.
device. They are thus available at connector X1 in a
19-inch module rack system. Other periph-
RAM (IC4) eral devices can be connected via these
The 62256 32-kbyte RAM (Random-Access outputs.
Memory) acts as supplementary working On IC8, the lead for the direction of
memory in addition to the microcontrollers transmission (pin DIR) is set to logical 1.
internal RAM for storing variable data (vari- Consequently, this module operates unidi-
ables such as calculated values and signals) rectionally. It sends the addresses A0 to A7.
and as a program memory. The RAM is IC9 transmits and receives data (bi-direc-
connected to the microcontroller via the tional transfer). Therefore, the direction of
address/data bus. transmission has to be controlled by IC1 via
When the power supply is switched off, the DIR pin.
the RAM loses all the information placed in
it (volatile memory). For that reason, this Control of ports 4 and 5
circuit incorporates a back-up battery (see These ports perform typical microcontroller
section headed Back-up battery overleaf). functions. Both ports are buffered by a
74LS245 bi-directional 8-bit driver. It sup-
Level converter (IC5) plies sufficient power for controlling the
For serial data transmission, the PC has an consumer units (e.g. LEDs) (cf. bipolar tran-
interface which enables it to communicate sistor control: low base current, high collec-
with other devices. That interface conforms tor current).
to the RS232 standard. The interface cable Control of the two ports is illustrated in
can carry TxD (Transmit Data) and RxD an example program at the end of this
(Receive Data) signals. However, the signal chapter.
levels of these transmissions are unsuitable
for use with the microcontroller. The
MAX232 level converter or interface module Discrete components
adjusts the signal levels used by the micro- Clock pulse generator
controllers serial interface (0 ... 5 V) to those The quartz crystal Q1 operates as a clock
used by the PC RS232 interface. The capaci- pulse generator for the oscillator integrated
tors C11 ... C14 function as voltage stores. in the microcontroller. The capacitors C15
and C16 are also required. The clock fre-
Logic modules (IC6, IC7) quency is 12 MHz.
The NAND (negative AND) gates of the
74HCT132 (IC6) process the control signals Reset button
from the microcontroller. Those gates form As its name suggests, the reset button S1 re-
a logic circuit that permits access to the fol- sets the microcontroller. After a reset, the
lowing physical addresses on the two mem- microcontroller starts running programs
ory modules IC3 and IC4: from address 0000H. That is why the moni-
EPROM addresses 0000H ... 7FFFH and toring program has to start at that address in
RAM addresses 8000H ... FFFFH. our example.
The controller board has a straightfor-
The two gates on the 74HCT125 (IC7) ward power-up reset. When the power is
process the control signals RD\ and WR\ switched on, capacitor C17 first has to be
from the microcontroller and output them charged via resistor R1. The voltage at C17
via the connector X1. and therefore at the microcontrollers reset
Robert Bosch GmbH
ICs
IC1 80C535 CMOS microcontroller 1
IC2 74HCT573 8-bit D flip-flop 1
IC3 27C256 32-kbyte EPROM 1
IC4 62256 32-kbyte RAM 1
IC5 MAX232 RS232 interface module 1
IC6 74HCT132 NAND Schmitt trigger 1
IC7 74HCT125 unidirectional 8-bit bus driver 1
IC8 74HCT245 bi-directional 8-bit bus driver 1
IC9 74HCT245 bi-directional 8-bit bus driver 1
Capacitors
C1 100 nF (SMD) SMD interference-suppression capacitor 1
C2...C9 100 nF (ceramic) interference-suppression capacitor 8
C10 1000 F (ELKO) Smoothing capacitor 1
C11...C14 10 F (Tantal-ELKO) Storage capacitor 4
C15, C16 22 pF (ceramic) Capacitor for quartz crystal 2
C17 10 F (ELKO) Power-up capacitor 1
Diodes
D1 LED (red) Power supply on indicator 1
D2 1N4001 Blocking diode for battery 1
D3 1N4001 Polarity-reversal protection diode for X3 1
Resistors
R1 68 k (0.25 W) Power-up resistor 1
R2 820 (0.25 W) Series resistor for LED D1 1
Quartz crystal and battery
Q1 12-MHz quartz crystal Quartz crystal for clock pulse generator 1
BAT 3.6-V battery Back-up battery for the RAM 1
Switches and buttons
S1 Button Reset button 1
S2 Switch On/Off switch 1
Connectors and sockets
X1 64-pin male connector DIN 41 612 type C, to module rack 1
X2 9-pin SUB-D socket Connection to PC (RS232 interface) 1
X3 2-pin connector For external power supply (optional) 1
J1...J6 8-pin strip connector Connections for ports 6
B1 PLCC68 68-pin IC socket for IC1 1
B3, B4 DIL28 28-pin IC sockets for IC3 and IC4 2
B2, B8, B9 DIL20 20-pin IC sockets for IC2, IC8 and IC9 3
B5 DIL16 16-pin IC socket for IC5 1
B6, B7 DIL14 14-pin IC sockets for IC6 and IC7 2 Table 3
52
1
VCC Positive power supply voltage C13 X1
VCC 8 A12
GND Ground 8 C12
8
IC2 A13
Multiple conductor showing R1 VCC IC4
Circuit example
GND
GND GND Q1 ... Q8 DIR A4
OE I/O ... I/O 7 GND C4
G A5
CE VCC C5
IC1 8 A6
Data
8 WE B ... B7
C6
ALE P. ... P.7 7 A7
vcc
PSEN J2 C7
P5. ... P5.7 J5 D2 BAT C4
IC8
RESET
VCC C16
P2. ... P2.6 GND GND A...A7 VCC
XTAL1 IC3 A17
Q1 P2.7 C8
VCC VCC GND C17
XTAL2 DIR
IC6 A ... A7 VCC A18
8 GND Addresses
C15 C16 EA C3 C18
P1. ... P1.7 IC6A IC6C G A19
1 A8 ... A14 GND 8
PE 8 B ... B7 C19
2 & & GND A2
J1
J6 GND GND 3 9 OE
VCC
8 VCC 10 A1
VCC IC6B CE O ... O7 5V
P6. ... P6.7 4 IC6D11 GND C1
VCC VCC D3
VPP
Robert Bosch GmbH
5 & 12
J4 C1 2
6 & 13 A32
GND X3 1 C32
Ground
8 VCC GND
P4. ... P4.7 GND VCC GND
C6 IC7
S2 A1 IOR\
J3 GND GND IC7A 8
1 A9 IOW\
P3.7 (RD) VAREF VCC IC7C 9 VCC
2
IC5 10 C2
P3.6 (WR) VAGND 3 A21
4 C11 4 IC7B 11
P3.2 ... P3.5 P3 C1+ V+ + VCC C21
GND IC7D 12 D1
C13 5
Circuit diagram of development circuit with 80C535 microcontroller (circuit example)
+ C12 A22
P3.1 (T x D) C1 V 13 PORT 4
+ 6 C22
VCC +
P3. (R x D) C2+ VSS VCC VCC GND A23
C5 GND R2 C23
+ C14 C10
C2 GND C7 A24
GND X2
GND GND GND GND
T1 IN T1 OUT 2 Tx D C30
T2 IN T2 OUT A30
R1 OUT R1 IN 3 Rx D C29
5 A29 PORT 6
R2 OUT R2 IN
GND C28
8
A28
8 C27
A27
NAE0850E
Robert Bosch GmbH
2 Component layout and circuit layout of development circuit with 80C535 microcontroller
Component
layout J4 J5 D3 +
C10
This shows the S2 X3
+ C A
1
C
C
arrangement of 2 IC2 9
C1*
the components J6
IC1 IC
on the circuit D1
9
board. The black R R J2 C
3 IC3
2 1
dots represent J3 Q1
100 mm
J1
holes in the board. C8
S1
S1
C15 C16
They have tin-
IC IC C IC4 IC
plated sleeves + + C 6
C17 7 7 6 8
inserted in them
5
9
C5
C11 +
+
which provide the X2 +
C13 +
electrical connec-
6
+ IC
1
C14 + BAT C4
tion between the 5
NAE0851E
+
+ 32 X1
C12 + D2
conductor tracks + +
Circuit layout Component side (top side) * C1 is fitted on the solder side.
P4 0 0 P5
The circuit layout shows
7
+ 1
P6
+
position of the holes rela- +
tive to each other is also + 32
critical.
Originally, printed circuit boards were simply Even in the age of EDA (Electronic Design
bases for conductor tracks and components. The Automation), only a small part of the design
ever increasing density of circuits has seen them process can be truly automated. A controlled
develop into electronic components, however. design environment and a controlled design
Assisted by computer-aided systems, the circuit process make it possible to eliminate typical
board designer incorporates the requirements of errors and to achieve reproducible results.
all disciplines within the circuit board layout dur- Design quality is attained by a controlled design
ing the design process. The considerations in- process with automated and manual checks.
volved include circuit development (e.g. circuit
diagram), casing design (e.g. dimensions), manu-
facturing (e.g. testing points, production costs)
and last but not least the requirements of the ve-
hicle manufacturer (e.g. connector configuration).
a The circuit diagram and the list of components itemising all circuit
elements are the starting point for the circuit-board layout (a).
The component details (e. g. dimensions, connector-pin assign-
ment) are entered in a CAD* library (b). Those component specifi-
cations define the decisive parameters for circuit-board design,
production documents, CAM processing and manufacture. That
information and the circuit-diagram information is entered in a com-
ponent connection schedule on the circuit-layout design system.
The component connection schedule details which connections
are connected to which.
b c During this phase, the physical design of the circuit board
(c) also has to be considered (e.g. dimensions, cut-outs,
connector positions, restricted zones in which no compo-
nents, or only very small ones, may be placed).
The component positions are then decided upon on the
basis of thermal, EMC, production system and physical
considerations (d). On the CAE* system, the electrical
connections are represented by straight lines (referred to
as rubber bands) drawn directly between two connected
points. Those rubber bands then have to be rearranged as
d conductor tracks on the circuit board. This involves untan-
gling them so as to produce the pattern of conductor
tracks. In this process, EMC and circuit design concepts
are taken into consideration. Multilayer circuit boards are
generally required. Any inconsistencies with design princi-
ples or specifications are indicated by the error marker.
Next, the production documents for manufacturing the
circuit board are produced (e), i.e. films of the conductor
tracks, print films, connection schedules for testing equip-
ment, component mounting diagrams and films for special
e Rubber prints (e.g. solder resist, solder paste).
band
NAE0854E
function is enabled by the logical combina- memory (latch). The latch is transparent
tion (IC6A and IC6B) of PSEN\ and RD\. when the level of the ALE signal is high.
Data can also be written to the RAM. That means that the full address is applied to
The user program is stored in this area of the program memory. The data outputs of
memory. The process of program develop- the program memory which are connected
ment must ensure that data is not stored in in parallel with the inputs of the latch are at
the area occupied by program code. Writing high resistance so as to prevent data
of data to the memory would otherwise collision.
overwrite and thereby destroy the program. The microcontroller subsequently sets the
level of the ALE signal to low. The address
Control lines low byte then remains stored at the latch
The microcontroller outputs control signals outputs. The full address remains applied to
via the control lines. They ensure that the the program memory for the entire com-
necessary components are selected at the mand cycle.
appropriate times. On the 80C535 (IC1) the The ALE signal is also generated for data
control signals are as follows: access to external data memories.
RAM
32 kbyte
Bit data Direct data Indirect data Application program
00FFH
128 byte 128 byte Program and
Bits in bit-addressable
Special Upper data memory
register of special
function register function RAM
register 8000H
(SFR)
(SFR) 7FFFH
0080H
007FH EPROM
32 kbyte
Bits in bit-addressable 128 byte Monitoring program
register of lower Lower RAM memory
RAM register banks
Program-only memory
SAE0848E
0000H 0000H
Robert Bosch GmbH
P0. The PSEN\ signal then switches back to ground. The microcontroller performs a re-
inactive status and the data lines of the pro- set and reverts to the start of the monitoring
gram memory revert to a high resistance program at address 0000H. After a certain
level. period, the capacitor is sufficiently charged
for the Reset\ input to adopt the high-level
RD\ status. The precise length of time depends
To read the data memory (external RAM), on the rating of the capacitor and the series
the microcontroller must first access the resistor R1.
module. This is done in the same way as In contrast to our example, there are also
when reading the program memory. In this microcontrollers on which the reset is active
case too, the address low byte is temporarily when the signal is high.
stored in the latch.
The data is placed on the data bus by the Serial interface
RAM when the OE\ pin switches to the low- The connection between the PC and the
level signal. The OE\ pin is controlled by the controller is provided by an RS232 serial
RD\ signal. During this phase, the micro- interface.
controller reads the data bus via port P0. The ports COM1 or COM2 can be used at
Theoretically, the OE\ pin can be con- the PC end, using either a 25-pin or a 9-pin
trolled directly by the RD\ signal from the SUB-D connector. That and the data trans-
microcontroller. In this development circuit, mission rate are set in the PC program dur-
however, the RD\ signal is combined with ing the programming process.
the PSEN\ signal via a logic circuit (IC6A, On the development circuit there is a
IC6B) so that the RAM can also operate as a 9-pin SUB-D socket (X2) which is con-
program memory. nected to the microcontroller serial interface
pins via the level converter MAX232 (IC5)
WR\ (see description of IC5).
The address to which data is to be written
is applied by the microcontroller to the ad-
dress bus. The ALE signal is used to control
copying of the data to the latch. The data is
subsequently placed on the address bus. The
RAM copies the data on the data bus when a
low-level signal is present at the WE\ (Write
Enable) pin. To that end, the microcon-
troller activates the WR\ signal.
Reset\
The Reset\ input (Pin 10) is required to
reset the microcontroller to a defined status
(input active at low signal level). When the
development circuit is switched on, this
takes place automatically through the con-
nection to capacitor C17 and its series resis-
tor R1.
At the moment the power is switched on,
capacitor C17 constitutes a short circuit.
This is a physical property of the uncharged
capacitor. As a result, the Reset\ input of the
microcontroller is briefly connected to
Robert Bosch GmbH
;Initialisation
8000: MOV R0,#01 ;Load register R0 with the constant 01H
8002: MOV R1,#80 ;Load register R1 with the constant 80H
;Main program
;1st moving light:
8004: MOV A,R0 ;Load contents of register R0 into accumulator A
8005: MOV P4,A ;Output the contents of accumulator A at port 4
8007: RL A ;Rotate the contents of A one place to the left
;(e.g. 000 0001B becomes 000 0010B)
8008: MOV R0,A ;Load contents of accumulator A into register R0
8009: LCALL 8100 ;Invoke the subroutine at address 8100H
;2nd moving light:
800C: MOV A,R1 ;Load contents of register R1 into accumulator A
800D: MOV P5,A ;Output the contents of accumulator A at port 5
800F: RR A ;Rotate the contents of A one place to the right
;(e.g. 1000 0000B becomes 0100 0000B)
8010: MOV R1,A ;Load contents of accumulator A into register R1
8011: LCALL 8100 ;Invoke the subroutine at address 8100H
8014: JMP 8004 ;Jump to address 8004H
;Subroutine time (approx. 0.13 s)
8100: MOV R3,#FF ;Load register R3 with the constant FFH
8102: MOV R2,#FF ;Load register R2 with the constant FFH
8104: DJNZ R2,8104 ;Decrement (reduce by 1) R2,
SAE0849E
ECU development
The control unit (ECU) is the central point which the ECU development stage plays
from which the functions of an electronic a decisive role. The following tasks are in-
system in a motor vehicle are controlled. volved in the development of the ECU
For that reason, extremely high demands in (Fig. 2):
respect of quality and reliability are placed hardware development,
on ECU development. function development,
software development, and
application.
Overview
Requirements
An electronic system consists of sensors and The product specifications and the develop-
setpoint generators, an ECU and actuators ment specifications document the require-
(Fig. 1). ments that a particular electronic system has
The sensors detect the operating parame- to meet. Those two documents form the
ters of the electronic system (e. g. wheel basis for the development process.
speed, engine temperature, ambient pres-
sure). The setpoint generators register the Product specifications
settings that the driver has specified with The product specifications define the re-
his/her operating controls (e.g. by means of quirements from the point of view of the
the air-conditioner switches). The sensors vehicle manufacturer. They describe the
and setpoint generators thus supply the in- functions that the product concerned must
put signals that are analysed and processed perform. They detail all requirements on the
by the ECU. part of the vehicle manufacturer with regard
Actuators (e.g. ignition coils, fuel injec- to the products and services to be supplied.
tors) convert the electrical output signals The requirements specified should be quan-
into physical variables. tifiable and measurable. The product specifi-
The process of developing an electronic sys- cations thus define what means of perform-
tem is made up of a number of stages, in ing what task is to be provided.
3 Actuators
4 Interface with other
during the course of the development
systems process.
5 Diagnosis interface
Robert Bosch GmbH
Quality Management
Product Specifications
Development Specifications
1.0
MAC
Parallel
PC interface
Serial
101101 ESU interface
K lead
(optional)
Hardware Software
Hardware is a generic term for the physical Software is a generic term for the total of the
(i.e. tangible) components of a system programs and data stored in the memory of
whether mechanical (e.g. heat sink, casing) a computer-controlled system.
or electronic (e.g. microcontroller, memory
modules, output modules). This applies to The central element of an ECU that per-
the components both individually and col- forms a function in a motor vehicle is the
lectively. microcontroller. It runs the program stored
in the program memory. That program rep-
The hardware of the ECU electrically pro- resents the functions of the ECU in the form
cesses the signals received from the sensors of program code.
and passes them onto the processor core of
the control unit for further processing. Out- The process of software development (Fig. 2)
put modules amplify the control signals so converts the requirements arising from the
that the actuators can be operated with the function descriptions into a program. The
required electrical power. machine code thus produced is entered in
the program memory of the ECU. To sim-
The task of hardware development (Fig. 2) is plify the process of writing a program, pro-
to design and produce an ECU that meets gramming languages are used. The increas-
the requirements arising from the develop- ing complexity of electronic systems makes
ment specifications. the use of high-level programming lan-
guages (such as the language C) absolutely
ECU functions essential. Software development is also
The control units job is to analyse the input assisted by simulation tools.
signals and to control the actuators so that
the system operates according to its in- Data adaptation
tended purpose. The overall behaviour of The behaviour of an electronic system does
the system can be broken down into a num- not depend solely on the ECU program. A
ber of separate functions (e.g. for gasoline- decisive role is also played by the data stored
engine management system: control of en- in the memory along with the program
gine idling speed, exhaust emission levels, code. An example of such data in the case of
radiator fan, etc.). And even with todays an engine ECU would be the ignition timing
state-of-the-art ECUs, the apparently map, which specifies the optimum ignition
straightforward control of the radiator fan timing from the point of view of fuel con-
is dependent upon a whole range of input sumption and emission levels for a range of
variables. It is not sufficient merely to switch engine operating conditions (engine speeds
the fan on when the engine is hot, and off and loads/torques). Such data is engine-spe-
again when it has cooled down. Further- cific. For that reason, it has to be calculated
more, every vehicle manufacturer has its and adapted during the process of develop-
own ideas as to exactly how this unit should ment of the vehicle. Adaptation of the data
operate. to the engine in this way is the task of the
application stage (Fig. 2).
Function development (Fig. 2) involves the
implementation of the engine manufac-
turers specifications, and the drawing up of
the function descriptions which describe the
ECU functions down to the very last detail.
Those documents form part of the develop-
ment specifications.
Robert Bosch GmbH
Hardware development test out the full range of functions and the
technical requirements. At this stage it is
The complexity of electronic systems has ready for continuous testing in prototype
seen a continuing increase in the past years. vehicles.
This tendency will continue in the future, The connection sizes and ECU dimen-
and it will only be possible to master such sions are as required for final production. It
developments by applying highly integrated may be, however, that not all of the vehicle
circuits. The call for small dimensions for all manufacturers specifications are satisfied at
system components places further exacting this stage, perhaps because different materi-
demands on the hardware development als have been used, for instance.
process.
Efficient and cost-effective hardware C prototype
development is now only possible by using The C prototype is the approval prototype
standard modules that are produced in large on which the vehicle manufacturers tests
numbers. for technical approval are performed. This
version of the ECU must reliably conform
Project starting point to all specifications. On successful approval
A schematic diagram sets out all the func- of the product, the development process is
tions that are to be performed by the ECU complete. As far as possible, C prototypes
being developed. From that point, the fol- are produced using full-production tools
lowing aspects can be clarified: and production methods as close as possible
definition of hardware required, to full series production.
cost estimate for the hardware,
extent of development work required, and D prototype
cost of tools. The D prototype is the pilot-series prototype
which also carries the full-production iden-
Hardware design prototypes tification plate showing the version number.
Once the project is underway, hardware de- D prototypes are fitted in pilot-series vehi-
sign prototypes are produced and subjected cles for large-scale vehicle trials. This version
to quality tests. These design prototypes may of the ECU is produced using volume-pro-
be one of four categories representing suc- duction methods and is fitted and tested un-
cessive stages along the road to the final pro- der volume-production conditions. It is the
duction ECU. Each prototype category is version with which the reliability of produc-
based on its predecessor and is designed for tion is verified.
a particular purpose in each case.
A prototype
The A prototype is derived from an existing
or modified ECU or a development circuit
board. Its range of functions is limited. Its
technical function is largely in place but the
A prototype is not suitable for continuous
testing. It is a function prototype that is used
for initial trials and to confirm the basic via-
bility of the design.
B prototype
The B prototype includes all circuit compo-
nents. It is a trial prototype that is used to
Robert Bosch GmbH
IC7B
ordering and developing new circuit
4 11
5 IC7D 12 D1
13
6
modules (function groups), VCC
VCC GND
+
R2
producing the circuit diagram, C7
C10
production
signment) the layout can then be produced. d Prototype
construction
e Testing
Robert Bosch GmbH
There are specific criteria that have to be Testing the finished circuit board
taken into account in the creation of the Electrical testing
layout. In connection with the positioning Once all components have been fitted and
of components these can include the all connections soldered, the circuit board
following considerations: must be tested. To this end, electrical testing
the power loss from specific components sequences are defined which run on a com-
(possibilities for heat dissipation), puter. These automatic tests check that all
EMC (electromagnetic compatibility) components are fitted and that the circuit
effects, functions properly.
convenient positioning of components in
relation to connectors, Thermographics
observance of restriction zones Thermographic images of the printed circuit
(component size), show the heat generated by the components
ease of fitting of components by auto- during operation (Fig. 2). Different temper-
mated production machinery, ature ranges appear on the film as different
accessibility of testing points, and colours. In this way, components that are
space requirements for testing adaptors. too hot can be identified. The information
obtained is incorporated in the list of modi-
Circuit board fications between the B and C prototypes.
The layout data can be used to produce films Changes to the circuit layout (e.g. heat-dissi-
for manufacture of the printed-circuit board pation through-contacting) can reduce the
(Fig. 1 c). The films are used to photographi- amount of heat generated.
cally expose the blanks (which are covered
with a light-sensitive coating) and then de-
velop and etch them. The individual layers
of multilayer printed-circuit boards are
placed on top of one another and hardened.
2
SAE0911Y
Robert Bosch GmbH
Encasing, sealing
passed on for final testing (Figs. 1 and 3).
Robert Bosch GmbH
Doping by diffusion
At high temperatures, the doping material
can diffuse into the silicon crystal. The effect
is achieved by creating a specific concen-
tration level of the impurity atoms at the
UAE0870Y
surface of the wafer. The difference in con-
centration causes the doping atoms to dif-
fuse into the silicon wafer.
Epitaxy
Wafer testing
This doping method creates a monocrys-
talline doped semiconductor layer a few
micrometres thick on a monocrystalline Chip assembly
substrate. If gaseous silicon tetrachloride
NAE0079E
ing a monocrystalline layer. If a defined The wafer is coated with a special lacquer
quantity of doping impurity is added to the and then exposed to light while covered with
gas flow, an epitaxial layer is created whose a metal screen. Following developing, the
electrical conductivity and conduction type areas of the lacquer coating previously cov-
differs fundamentally and abruptly from ered by the screen and the oxide layer below
that of the substrate. it can be removed by etching. The position,
size and shape of the holes thus created
Photolithography and the planar process match precisely the specified design. In the
In the photolithography process, the pattern subsequent doping process in the diffusion
of the component design is transferred to furnace (Fig. 4) or by ion implantation, im-
the wafer by using metal screens. When the purities which have a known electrical effect
design has been created, the data for pro- such as boron or phosphorus pass through
ducing the screen is saved onto a storage those holes in the oxide layer into the sili-
medium (e.g. magnetic tape or CD). It is con, thereby creating n-type or p-type re-
then used to control a photographic expo- gions in the desired locations. Afterwards,
sure device that transfers the design pattern the oxide layer is removed again and the
to photographic plates. This pattern is sub- wafer is ready for the next stage in the
sequently reduced by optical means to the manufacturing process.
actual size for the application and copied
onto metal screens on which it is repeated as The photolithographic process and the dop-
many times as will fit onto the size wafer ing process are repeated as many times as
being used. required to produce the desired number of
layers of different conductivity on the semi-
As this method can be used for structures conductor component. In the case of com-
that are many times smaller than the wave- plex integrated circuits, this may involve as
length of the light used, it will continue to many as 20 or more separate manufacturing
be used in the future. The size of the small- stages. In order to provide the electrical con-
est achievable structures depends on the nections between the resulting functional
wavelength of the light source. Using elements, the wafers are coated with alu-
UV lasers, photolithographic methods have minum or copper and then the metal con-
already produced structures as small as ductor track patterns are formed. This
0.08 m in the laboratory (by comparison, a process too may involve the formation of
human hair has a diameter of 40 ... 60 m). multiple layers of metal, one above the other.
Even smaller structures can be produced Completion of the wafer production process
using other methods such as X-ray or elec- is followed by electrical testing (preliminary
tron-beam lithography. However, those testing) of the individual chips on the wafer.
methods are substantially more expensive as Chips that do not meet the specifications are
they can only expose one IC at a time on marked with coloured dots. The wafers are
the wafer. For that reason, they are only used then cut up into individual chips using a
in special cases. diamond cutting tool (Fig. 5). The func-
tional chips are subsequently placed in metal
Silicon wafers are easily oxidised by oxygen or plastic casings and fitted with external
or water vapour. The oxide layer thus created connections (Fig. 6). After being hermeti-
prevents penetration by the doping atoms in cally sealed or encased in plastic, they go
the doping process. In the planar process, through a final testing stage.
holes are created in the oxide layer so that
when the wafer is doped, localised areas of
n-type and p-type material are created.
Robert Bosch GmbH
1 2 3 4
Fig. 6
1 Base
NAE0075Y
NAE0871
2 Chip
3 Bonded connection
wire
4 Connection pin
NAE0077Y
Robert Bosch GmbH
Micromechanics
c Production of bars
and ridges
c
1 Etching screen
2 Silicon
UAE0794Y
Fig. 3
1 Fixed electrode
D
2 Gap
3 Oscillating electrode
Robert Bosch GmbH
1 2 3 1 2 3 Fig. 1
Schematic diagram:
a b
conductor tracks are
shown as raised
surfaces
4 4
a One-sided
1 2 3 2 3 1 b Double-sided,
non-interconnected
c d c Double-sided,
interconnected
1. layer d Multilayer
2. layer
3. layer 1 Conductor track
SAE0872E
Photoresist application
(coating, exposing, developing).
This layer leaves the conductor
tracks and solder eyes uncovered.
Fig. 2
The base material is Protective coating application Protective coating application
(e.g. gold (Ag), tin-lead alloy (SnPb)) (e.g. gold (Ag), tin-lead alloy (SnPb))
coated on both sides
with a copper film
Robert Bosch GmbH
1 Wired component
volves a process of operations organised 5 2 2 Circuit board
according to the flow principle in a rigid 3 SMD
sequence, aligned to a particular spatial 4 Soldering point
arrangement and tied to a cyclic timing 5 Cement
Robert Bosch GmbH
4 Stages in pcb manufacture using a flow production process and reflow/wave soldering (example)
Turning the circuit board Cementing station SMD fitting Cement hardening
Turning the circuit board Wired-component fitting Wave soldering Position checking
Film and hybrid circuits using suitable material systems, resistors and
capacitors can also be integrated. The wiring
Film circuits densities of these circuits is considerably
In film integrated circuits, passive circuit el- greater than with thick-film circuits.
ements (usually conductor tracks, insulation
and resistors, but also capacitors and induc- Hybrid circuits
tors) are applied to a base (substrate) in lay- Hybrid circuits are integrated film circuits
ers. The advantages of these circuits are: with additional discrete components such as
fine structures (up to approx. 10 m) with capacitors and integrated semiconductor
high element density, and circuits (ICs) that are attached by soldering
good high-frequency characteristics. or cementing. The use of unpacked semi-
conductor chips, to which the connections
Those advantages are counterbalanced by are made by bonding, or SMDs makes high
relatively high production costs. component densities possible. Extremely
The thickness of the films originally gave small hybrid ECUs (microhybrids) can be
rise to the terms thin-film circuit and made by using ceramic multilayer sub-
thick-film circuit. Classification is now strates. The advantages of these circuits are:
based on differences in the method of pro- high permissible installation temperatures
duction. due to their good heat dissipation proper-
ties,
Thin-film circuits compactness of design combined with
With thin-film circuits, the films are applied good shock resistance, and
to glass or ceramic substrates using vacuum- good resistance to mediums.
coating processes.
Hybrid circuits are therefore particularly
Thick-film circuits suited to use in telecommunications systems
For thick-film circuits, the preferred method and for automotive applications where they
of manufacture involves application of the can be found in ABS, traction control, ESP,
films to ceramic substrates by screen print- transmission control and engine-manage-
ing processes followed by firing. ment ECUs (engine-mounted).
Ceramic multilayer substrates Figure 1 shows the basic stages in the pro-
Ceramic multilayer substrates are made up duction of a hybrid-circuit substrate. The
of unfired ceramic foils onto which conduc- tapes are punched with holes for the vias for
tor tracks are applied by screen printing each separate wiring layer and the holes are
methods. Several such foils are then com- then filled with silver paste (Fig. 2a). Screen-
bined to form a multilayer laminate which printing stations print the conductor tracks
is fired at 850 ... 1600 C so that it becomes a onto the film. The various layers are aligned
solid ceramic body with integral conductor with each other, laminated and then fired at
channels. A typical hybrid substrate consists 890 C. A specially controlled sintering pro-
of four or five layers. Particularly high cess keeps the ceramic firing to within toler-
wiring densities can be achieved with LTCC ance limits of roughly 0.03 %. This is impor-
(low-temperature cofired ceramic) line-line tant for the packing density. The circuits re-
substrates. sistors are printed on the reverse side of the
substrate and fired (Fig. 2b).
To make the electrical connections between For bonding on the top side, the surfaces
layers, holes are punched in the individual are finished using a plating process adapted
films and filled with metal paste. Those to the LTCC. The spacing of the microcon-
holes are referred to as so-called vias. By troller contacts (bond land grid on the
Robert Bosch GmbH
substrate) ranges from 450 to 260 m. The Method 2: the finished hybrid is cemented
component connections are bonded using to the aluminum casing using heat-conduct-
32-m gold wire and 200-m aluminum ing cement and connected to the plastic-en-
wire. cased connector pins using bonded gold or
Alongside the electrical vias, there are also 300-m aluminum wire. Before the cover is
thermal vias with a diameter of 300 m for cemented in place, a gel is applied to protect
optimum cooling of ICs with high power the circuit.
losses. The effective thermal conductivity of
the substrate can thus be increased from
approx. 3 W/mK to 20 W/mK.
All components are cemented with conduc-
tive cement. There are two methods used for
final assembly of the finished hybrid.
a b c
1 2 1 Fig. 2
3 a Inner layer
5 b Reverse side with
resistors
2 cm
2 cm
5 mm
1 Conductor track
NAE0876E
2 Via
3 Resistor
4 Microcontroller
5 Bonded wire
Robert Bosch GmbH
Analog circuit p Electronic circuit with Byte: Coherent group (word) of eight
which analog signals are processed as p bits.
opposed to p digital circuit. Analog signals
are infinitely variable within a specific range. C
Examples: battery voltage, speedometer
display. Capacitance: Property of a capacitor which
indicates how much electrical charge and
Avalanche breakdown: Sudden increase in therefore electrical energy it can store
the number of p charge carriers resulting when a specific voltage is applied.
from the release of bound electrons from the
crystal lattice of a p semiconductor in the CCD: Abbreviation for p charge-coupled
p space-charge zone caused by a high- device.
strength electric field (p avalanche effect).
Central processing unit (CPU): Main oper-
Avalanche effect: see p avalanche breakdown ating unit of a p microcontroller.
Circuit layout: Arrangement of conductor Digital circuit: Electronic circuit with which
tracks and components on a printed circuit. digital signals are processed as opposed to
p analog circuit. Digital signals are in the
Clock-pulse generator: Circuit that gener- form of a series of discrete pulses rather
ates a timing pulse of a fixed frequency for than infinitely variable characteristics. Ex-
synchronising all operations in a p micro- ample: digital clock.
computer.
Diode: Component with a p p-n junction
Closed-loop control: Method of control and two connections one to the p-type
characterised by a closed control loop. The region and one to the n-type region.
actual value of a variable is continually com-
pared with its specified value. As soon as a Discrete active component: Active compo-
difference is detected between the two, a nent with fewer than 100 individual func-
readjustment is made to the setting of the tions on a single p chip.
actuator (output variable).
Doping: Controlled introduction of impuri-
Communication: Exchange of data between ties with known electrical effects into high-
electronic components. purity semiconductor crystals with the aim
of producing p free electrons to create an
Conductivity: Reciprocal of resistance p n-type semiconductor or p holes to create
(1/ = A/V). a p p-type semiconductor.
Conductor: Material with high electrical DRAM: Abbreviation for p dynamic ran-
p conductivity due to its large number of dom-access memory.
free p charge carriers. Metals are the best
conductors. Dynamic random-access memory (DRAM):
Volatile short-term memory (DRAM =
CPU: Abbreviation for p central processing Dynamic p RAM) which allows direct ac-
unit. cess to any memory location and which
does not retain its contents when the power
D supply is switched off.
Lattice defect: Mobile electron gap and crys- Monitoring program: A p program by
tal lattice vacancy in p semiconductors the means of which the user can control the
atoms of which are held together by valency actual p user program.
bonds. It carries a positive charge. Lattice
defects are also called p holes. Monolithic integration: Highly integrated
arrangement of all components of a circuit,
LED: Abbreviation for p light-emitting diode. including the interconnecting conductors,
on a single silicon crystal (p chip).
Light-emitting diode (LED): p Diode that
emits light when a voltage is applied in for- Monolithic integrated circuit (IC): Active
ward direction. An LED produces the component with 100 individual functions
reverse effect to a photodiode. on a single p chip.
Register: A memory area integrated in the Space-charge zone: Poorly conductive re-
p CPU or a peripheral module that allows gion depleted of mobile p charge carriers at
fast reading and writing of data. a p p-n junction that is created by p diffu-
sion of free p charge carriers in either direc-
Reverse current: Small residual current aris- tion across the junction without the pres-
ing from minority charge carriers in the ex- ence of an external voltage.
tended p space-charge zone at a p p-n junc-
tion under external voltage. SRAM: Abbreviation for static random-
access memory.
ROM: Abbreviation for read-only memory.
Substrate: Base material in the manufacture
S of electronic and microelectronic modules
(e.g. silicon crystal, circuit board, ceramic).
Schottky diode: p Diode with a p p-n junc-
tion in the form of a metal-semiconductor T
junction. The electrons transfer more easily
from the n-type silicon to the metal film Thyristor: Component with at least three
than in the opposite direction, thereby creat- p p-n junctions that can be switched from a
ing an electron-depleted boundary layer reverse-bias condition to a forward-bias
known as the Schottky barrier. condition (or vice versa).
V Tables
Variable-capacitance diode: p Diode with a
p space-charge zone at the p p-n junction 1 Decimal multiples and sub-multiples of SI units
Glossary and tables for automotive microelectronics Where does the term electronics come from? 87
This term really originates from the ancient The first Electronic Engineer though goes
Greeks. They used the word electron for am- back to the 19th century. He was listed in the
ber whose forces of attraction for wool and 1888 Edition of a form of Whos Who, pub-
similar materials had already been described lished during the reign of Queen Victoria. The
by Thales von Milet 2,500 years ago. official title was Kellys Handbook of Titled,
Landed and Official Classes. The Electronic
The term electronics originates directly from Engineer is to be found under the heading
the word electrons. The electrons, and Royal Warrant Holders, that is the list of per-
therefore electronics as such, are extremely sons who had been awarded a Royal Warrant.
fast due to their very small mass and their And what was this Electronic Engineers
electrical charge. job? He was responsible for the correct func-
The mass of an electron has as little effect tion and cleanliness of the gas lamps at court.
on a gram of any given substance as a 5 gram And why did he have such a splendid title?
weight has on the total mass of our earth. Because he knew that Electrons in ancient
Greece stood for glitter, shine, and sparkle.
Incidentally, the word electronics is a prod-
uct of the 20th century. There is no evidence Source:
available as to when the word was used for Basic Electronic Terms (Grundbegriffe der
the first time. Sir John Ambrose Fleming, one Elektronik) Bosch publication (reprint from the
of the inventors of the electron tube could Bosch-Znder (Bosch Company Newspaper)).
have used it around 1902.
LAE0047Y
Robert Bosch GmbH
L Phototransistor, 31 T
Laser diode, 31 Photovoltaic cell, 30 Temperature coefficient, 14
Lattice defect p-type p doping Pipeline, 41 Tenting process
Layout (circuit example), 53 Planar process, 70 p printed-circuit board
Layout, 54, 65 PMOS transistor, 19 (manufacture)
LED, 30 p-n junction, 10f Thermographics, 66
Light-emitting diode, 30 Polled I/O, 43 Thick-film circuits (manufacture), 78
Linker, 46 Pressure sensor, 33 Thin-film circuits (manufacture), 78
Printed-circuit board Through-fitting method
M (manufacture), 73ff (printed circuit board), 75
Magnetoresistor, 16 Printed-circuit board (types), 73 Thyristor, 19
Memory, 27, 41 Printed-circuit board, 66 Transistor, 17ff
Memory-mapped peripherals, 42 Product specifications, 60 Transputer, 26
Memory-memory architecture, 38 Programmed I/O, 43 Tunnel diode, 15
Memory-register architecture, 38 Programming (circuit example), 58
Metal-film resistor, 12 Programming model, 37 U
Microcomputer system, 36 PROM, 29 Unipolar transistor, 18
Microcontroller, 34, 36 Prototype construction, 66 Units, 86
Microelectronics, 8 PTC thermistor, 16 User program, 47
Micromechanical sensors, 32 p-type doping p doping
Micromechanics, 72 V
Microprocessor, 24 Q Varistor, 16
Mnemonics, 46 Quality assessment, 63 Via (multilayer substrate), 78f
Monitoring program, 47 Quality management, 63 Von-Neumann architecture, 45
Multilayer (circuit board), 73
Multilayer capacitor, 12 R W
Multilayer substrate (manufacture), 78 RAM, 27 Wafer
Multiple instruction length, 40 Random-access memory, 27 (semiconductor manufacture), 68ff
Multiple instruction set, 39 Rectification, 15 Wiring harness, 4
Rectifier diode, 15 Wound capacitor, 12
N Register, 37 Write-back phase, 41
Networking, 4 Register-register architecture, 39
NMOS transistor, 19 Resistor, 12 Z
Non-conductor p conductivity Reverse current p p-n junction Zener breakdown p p-n junction
Non-destructive instruction-set Review, 63 Zener diode, 15
architecture, 39 RISC architecture, 41
NTC thermistor, 16 ROM, 29
n-type doping p doping
S
O Schottky diode, 15
Object code, 47 Semiconductor components
Ohm, 12 (manufacture), 68ff
Semiconductor components, 13
P Semiconductor, 8ff
Panel (printed circuit board), 75 Single-chip computer, 26, 34
Panel plating Software development, 62
p printed-circuit board Software, 62
(manufacture) Source code, 46f
Pattern plating Source, 18
p printed-circuit board Space-charge zone p p-n junction
(manufacture) Store operand phase, 41
Peripheral modules, 42 Subroutine, 59
Peripherals, 36 Subtractive process
Photodiode, 30 p printed-circuit board
Photolithography, 70 (manufacture)
Photoresistor, 30 Surface-mounting method, 75
Robert Bosch GmbH
Abbreviations A L
ABS: Antilock braking system LED: Light-emitting diode
ADC: Analog-digital converter LKS: Lagekontrollsystem,
ALE: Address Latch Enable i.e. position checking system
ALU: Arithmetic logic unit (p printed-circuit-board manufac-
ASIC: Application-specific integrated ture, production process)
circuit LSI: Large-scale integration
ASR: Antriebsschlupfregelung, LTCC: Low-temperature cofired
i.e. traction control ceramic
ASSP: Application-specific standard
product M
MOS: Metal-oxide semiconductor
B MSI: Medium-scale integration
BKS: Bauteilkontrollsystem,
i.e. component checking system P
(p printed-circuit-board manufac- PC: Personal computer
ture, production process) PROM: Programmable read-only
memory
C
CAD: Computer-aided design R
CAE: Computer-aided engineering RAM: Random-access memory
CAM: Computer-aided manufacturing RISC: Reduced instruction-set
CCD: Charge-coupled device computer
CISC: Complex instruction-set ROM: Read-only memory
computer
CPU: Central processing unit S
(of a p microcontroller) SiO2: Silicon dioxide
SMD: Surface-mounted device
D SMT: Surface-mounting technology
DAC: Digital-analog converter SoC: System on a chip
DMA: Direct-memory access SSI: Small-scale integration
E T
EEPROM: Electrically erasable TCS: Traction control system
programmable read-only memory
EMC: Electromagnetic compatibility U
EPROM: Erasable programmable UV light: Ultraviolet light
read-only memory
ESP: Electronic stability program V
VHDL: Visual Hardware Description
F Language
FET: Field-effect transistor VLSI: Very-large-scale integration
Flash-EPROM: Flash erasable
programmable read-only memory C: Microcontroller
H
HDL: Hardware Description Language
I
I/O ports: Input/output ports
IC: Integrated circuit
ICT: In-circuit test
(p printed-circuit-board manufac-
ture, production process)