Design and Development of a

Racing Motorcycle Engine

Introduction
World Superbike championship
Petronas designed engine when rules allowed 900cc I3 to
compete with 750cc I4
Rule changes meant that 900cc I3 must race 1000cc I4
Main engine development target was to maximise power
Baseline engine had rev limit of 14000 rpm
Engine speed limited by valve train dynamics
Target of 16000 rpm engine speed to increase power
Use of analysis to minimise testing
Valve Train Design and Development
Crankshaft Design and Development
2

Design and Development of the

Valve Train for a Racing
Motorcycle Engine
Phil Carden Ricardo UK
Ken Pendlebury Ricardo UK
Naji Zuhdi Petronas
Andrew Whitehead Del West USA

Contents
Introduction
Component development
Design Analysis
Failure modes
Conclusions

Introduction
World Superbike championship
Petronas designed engine when rules allowed 900cc I3 to
compete with 750cc I4
Rule changes meant that 900cc I3 must race 1000cc I4
Main engine development target was to maximise power
Baseline engine had rev limit of 14000 rpm
Engine speed limited by valve train dynamics
Target of 16000 rpm engine speed to increase power
Use of analysis to minimise testing
Every valve train component (except valve collets) was
redesigned
5

Camshaft development
Machined from 16CrNi11
Case hardened and carbonitrided
Inside diameter increased from 16 mm to 17 mm
Outside diameter maintained at 24.5 mm
Gear flange sculpted
Mass reduced by 8.4% - Inertia reduced by 10.3%
Marking of cam nose during run-in but no further problems

Valves development
Titanium 6242+0.2Si
Molybdenum alloy coating on stems, Cr2C3 CrNi coating on tip
Intake valve drilled with 2.25 mm diameter hole to save 1.4g
Intake valve head back angle reduced to save further 2.1g
Intake and exhaust collet groove changed to suit lash disc
Final intake valve mass 20.8g
Final exhaust valve mass 21.4g
Parameter
Intake valve head diameter (mm)
Intake valve stem diameter (mm)
Intake valve length (mm)
Exhaust valve head diameter (mm)
Exhaust valve stem diameter (mm)
Exhaust valve length (mm)

Baseline
36.0
5.0
93.7
30.0
5.0
95.0

Final
36.0
5.0
93.2
30.0
5.0
94.5

Tappet development
Through hardened tool steel
Polished to 0.05Ra on contact face
3m thick DLC coating applied by plasma assisted CVD process
Slots introduced to save 4.9g
Thickness of crown and wall reduced to save further 2.6g
Durability adequate for racing
Final mass 26.0g
Parameter
Outside diameter (mm)
Overall length (mm)
Central crown thickness (mm)
Minimum wall thickness (mm)
Slots in skirt

Baseline
32.0
23.5
3.3
0.8
No

Final
32.0
23.5
3.0
0.7
Yes

Retainer, collet, shim development

Baseline engine used steel lash cap
This resulted in valve tip durability problems
Replaced by lash disc
Spring retainer was machined from maraging
steel and nitrided
Design was changed to suit revised spring
and lash disc
Collets were titanium and were unchanged
Combined mass was reduced by 1.0g
9

Valve spring development

Baseline spring pack designed for 13500 rpm
Problems soon emerged at higher speed (see later slides)
but actual spring failures were very rare
Alternative springs were procured from several suppliers
Final springs designed for peak lift of 11 mm and 16000 rpm
Effective combined spring mass reduced by 3.3g
Stresses were increased
Spring wire quality was improved
Parameter
Wire diameter (mm)
Mean coil
diameter (mm)
Fitted length (mm)
Fitted force (N)

Baseline
inner
outer
2.90
3.95
17.9
24.75

Final
inner
outer
2.90
3.7
17.1
23.7

32.6
153

28.75
130

32.6
298

28.75
279

10

Design Analysis Cam profile design

Ricardo Software VALKIN
Focus mainly on intake valve train
Changes made to enable high speed
Peak intake lift reduced by 1 mm
Intake period increased by 2.8 deg

11

Parameter
Peak kinematic valve lift L
(mm)
Inner seat diameter D (mm)
L/D
Lift area integral
Period top of ramp (deg)
Ramp height (mm)
Ramp velocity (m/s)
Valve acceleration on
opening flank (m/s2)
Valve acceleration on cam
nose (m/s2)
Valve acceleration on
closing flank (m/s2)
Opening side acceleration
ratio
Closing side acceleration
ratio

Baseline
12.0

Final
11.0

35.0
0.343
0.555
307.2
0.20
0.432 @
14000 rpm
29818 @
14000 rpm
11530 @
14000 rpm
36962 @
14000 rpm
2.51

35.0
0.314
0.557
310.0
0.20
0.500 @
16000 rpm
33404 @
16000 rpm
13305 @
16000 rpm
41554 @
16000 rpm
2.51

3.21

3.12

Design Analysis Cam/tappet

Ricardo Software VALKIN
High speed contact stress increased
Low speed contact stress reduced
Film thickness at nose reduced slightly
Film thickness at transition improved
Tappet edge clearance increased

12

Parameter
Peak cam tappet contact
stress at idle (N/mm2)
Peak cam tappet contact
stress at rated speed
(N/mm2)
Lubricant film thickness at
peak cam lift (m)
Deschler and Wittman
number at peak lift
Maximum number of
consecutive crank degrees
at which oil film thickness is
less than 0.1 m at rated
speed
Minimum tappet edge
clearance (mm)

Baseline
831 @
3500 rpm
400 @
14000 rpm

Final
764 @
3500 rpm
436 @
16000 rpm

0.295

0.278

0.207

0.272

8.26

7.86

0.30

1.90

Design Analysis Dynamics

Ricardo Software VALDYN
Node representing effective
local mass of camshaft

Camshaft bending stiffness

and support stiffness
Cam/tappet stiffness
(dependent on eccentricity)

Node representing tappet mass

Stiffness of valve stem

between tip and centre
of mass

Valve spring models

- 8 masses per coil
- Connected by stiffness
- Coil clashing model
- Spring interference damping
uncertainty so modelled
with 2 levels for each run

Combined stiffness to
represent valve seat
contact and valve head
bending
13

Design Analysis Valve seating

Baseline design
Loss of control from ~14000 rpm
Sharp transition to high velocity
at ~14800 rpm
Large valve bounce at 15000 rpm
Failures of valve stem
Final design
Loss of control from ~16000 rpm
Below 4 m/s even at 17000 rpm
No failures
Conclusions not dependent on
spring damping assumption
14

Design Analysis Valve jump

Baseline design
Sudden transition at
~14600 rpm
Final design
Progressive increase
in separation from
~16000 rpm with
high damping
Less than 0.2 mm
peak separation at
17000 rpm
Results sensitive to spring interference damping assumption
15

Design Analysis Spring surge (1)

Baseline design
High surge amplitude on both
springs
+/- 1mm normal target for
passenger car engines
Final design
Significant reduction in surge
across speed range
Results moderately sensitive to
spring interference damping
assumption
16

Design Analysis Spring surge (2)

Spring seat hammering
On baseline design surge led to loss
of contact between spring and seat
at high speed just after valve
closing
High force when contact reestablished
Some
failures of
spring end
tangs
resulted

17

Design Analysis Spring stress

Stress at worst case location in spring
increases as baseline valve train loses
control at high speed
Final design
Pseudostatic spring stress levels
were increased but the spring
strength was also improved
Dynamic stresses were controlled
to similar level as baseline design
Parameter
Pseudostatic spring cover factor at
rated speed
Shear stress in outer spring pseudostatic (N/mm2)
Shear stress in inner spring pseudostatic (N/mm2)
1

Baseline
1.21 @ 14000
rpm
298 fitted
1065 max
356 fitted
1066 max

Final
1.39 @ 16000
rpm
410 fitted
1294 max
291 fitted
1199 max

18

Design Analysis Whole engine model

VALDYN model
was extended
and used to
calculate
effect of
timing drive
on valve train
motion
dynamic loads
at gears and
fasteners

19

Tappet bore failure

Several failures of cylinder head structure at tappet bore
Cracks in cylinder head at machined slot for cam clearance
VALDYN used to calculate moment on tappet
Reaction forces calculated and applied to local FE model
Low safety factors confirmed and alternative designs addressed
Small change in fillet radius gave desired improvement

20

Valve tip wear

The use of a lash cap resulted in sporadic failures
Attributed to
poor control of diametral fit leading to tipping of lash cap
also poor surface finish inside lash cap
Problem solved by change to lash disc design

21

Valve stem failures (1)

As engine speed was increased several intake valve stem
failures occurred
Stems broke just below the bottom of the
spring retainer
Failure surfaces indicated tensile fatigue so
loads must have been very high but
analysis indicated that valve train should
be under control

22

Valve stem failures (2)

Spring relaxation is known to be dependent on stress, time
and temperature and typically results in 5% loss of preload
Up to 40% loss of preload was
measured on baseline springs
Graph shows consequence of a
20% reduction in baseline
spring preload on the valve
seating velocity at high speed
300 rpm reduction in
transition to damaging level

23

Valve stem failures (3)

140

Testing of revised springs was

performed on a cylinder head rig

Averrage relaxation force (N)

Periods of constant speed operation

were followed by measurement of
spring force

Competitor

120

NHK

100
80
60
40
20
0
Fitted

Compressed

100

18000

90
16000

Spring relaxation force (N)

70

14000

Engine speed (rpm)

The best springs relaxed by just 3%

following prolonged operation at
16000 rpm

80

period 1
period 2

12000

60
50

period 3
Fitted relaxation

40

10000

This rose to 10% after 50000 cycles

at 16500 rpm

30
20

8000

10
6000

0
0

24

100000

200000 300000 400000

Spring cycles

500000

600000

Valve spring end tang failures

Springs from several suppliers
were tested during the project
One of the alternative spring
designs suffered from failure of
the end tangs
This was attributed to impacts
between spring and seat

25

Conclusions
The final intake valve train
had effective mass reduced by 15.3g (18%)
had exceptional durability with rev limiter set to 16000 rpm
was able to survive overspeed events at up to 17000 rpm
without failure
Success was achieved by
making extensive use of dynamic simulation
combined with minimal rig testing
The contribution of world class component suppliers to the
success of the project was invaluable
26

Thank you for your attention

Any questions ?
27

Design and Analysis of a

Lightweight Crankshaft for a
Racing Motorcycle Engine
Naji Zuhdi, PETRONAS
Phil Carden, Ricardo UK
David Bell, Ricardo UK

Contents
Introduction
Design overview
Engine balance
Main bearing analysis
Torsional vibration
Stress analysis
Conclusions

29

Introduction
World Superbike championship
Petronas designed engine when rules allowed 900cc I3 to
compete with 750cc I4
Rule changes meant that 900cc I3 must race 1000cc I4
Main engine development target was to maximise power
Baseline engine had rev limit of 14000 rpm but changes to
valve train enabled 16000 rpm engine speed
Piston and rod were also redesigned to enable operation at
higher speed
This presentation covers design and analysis of crankshaft
30

Crankshaft design objectives

Main objectives
Reduce crankshaft mass
Reduce rotating inertia
Reduce friction
Reduce windage
Maintain adequate
crankshaft strength
Maintain adequate bearing
durability
Maintain acceptable engine
balance
31

Crankshaft design overview

Fully machined crank
Integral drive gear
Double vacuum remelted
steel 31CrMoV9
Gas nitrided to 800Hv to
depth of 0.3 mm
Polished bearing journal
surfaces
Full circumferential grooves in
main bearings
Big end bearings supplied
from main bearings via
drillings
32

Crankshaft Design Iterations

Pictures show the design
evolution of the crankshaft
The drive gear was moved
from web 3 to web 5 to avoid
transmitting power through
the balancer shaft
Piston and rod were
lightened during the project
Component
Piston assembly mass (kg)
Connecting rod mass (kg)

Baseline
0.292
0.278

Final
0.249
0.245

33

Baseline
Design

Intermediate
Design

Final
Design

Reducing Mass and Rotating Inertia

Smaller counterweights used for final
design as engine was no longer fully
balanced (see later section)
Reduce the mass of upper portion of the
crankshaft
Drill through the crank pin
Use heavy metal inserts in
counterweights
30% mass reduction
35% inertia reduction

34

Baseline

Final

Minimising Friction Losses

Windage loss reduction
Thinner webs with chamfered edges
Shrouded balance shaft
Bearing friction loss reduction
Journal diameters were not changed
due to cost and lead time implication

35

Increasing Strength
Final design had piston guided rod
This eliminated the need for thrust faces
on big end journals
Thus permitting use of a large fillet radius
in the critical area of crankshaft overlap
region
Baseline

36

Final

In-Line 3 Cylinder Engine Balance

In-line 3 cylinder engine has
Balanced primary and
secondary reciprocating forces
Unbalanced primary and
secondary reciprocating
moments
Baseline FP1 engine had crank
counterweights and balancer
shaft arranged to give complete
balance of primary moment
But is complete moment balance
necessary ?
37

In-Line 3 Cylinder Engine Balance

Frecip = mrecip 2 R cos

Fb =

mrecip

1
Fa = (mrot + mrecip ) 2 R
2

1
mrecip 2 R
2

1
mrecip
2

Frot = mrot 2 R

Fa

mrot

1
mrot + mrecip
2

Fa cos 30

Cyl#2

Fa cos 30

Fa cos 30
30

30

Fa sin 30
Fa

Fa cos 30

Fa sin 30
Cyl#1

38

Cyl#3

Fa

Crank/Balancer Design Iterations

Removing the counterweights on
the balance shaft was tried
Level of vibration was acceptable to
riders but a frame failure occurred
A compromise was adopted for the
final design

Baseline
Design

Experimental
Design

39

Component

Baseline
Design
50%
100%
100%
100%

Fa force balance factor

Primary forces
Primary moment, Ma
Primary moment, Mb

Final
Design

Final
Design
30%
100%
90%
60%

Residual out-of-balance moments

Numerical values of residual out-ofbalance moments are shown
For engine with no counterweights
For baseline design
For final design
Parameter
Primary shaking moment
pitch at 14000 rpm (Nm)
Primary shaking moment
pitch at 16000 rpm (Nm)
Secondary shaking moment
pitch at 14000 rpm (Nm)
Secondary shaking moment
pitch at 16000 rpm (Nm)
Primary shaking moment
yaw at 14000 rpm (Nm)
Primary shaking moment
yaw at 16000 rpm (Nm)

No c/w
9158

Baseline
0

Final
1031

1345

803

803

683

892

5816

215

281

40

Main bearing analysis

ENGDYN bearing analysis shows
Reduced peak specific load at
worst case speed (peak torque)
Slight reduction in minimum oil
film thickness at high speed
Slight increase in hydrodynamic
power loss at 14000 rpm
Parameter
Baseline
Maximum peak
56.3 @ Main No.4
specific main bearing
12000 rpm
2
load (N/mm )
Minimum oil film
0.59 @ Main No.4
14000 rpm
thickness (m)
Maximum predicted
159.3
@
14000
rpm
oil temperature (C)
Total power loss at all
3.022
main bearings (kW)
@ 14000 rpm

Final
54.9 @ Main No.4
12000 rpm
0.53 @ Main No.2
16000 rpm
160.2 @ 14000 rpm
165.5 @ 16000 rpm
3.144 @ 14000 rpm
4.001 @ 16000 rpm

41

Torsional vibration analysis

VALDYN
linear
frequency
domain
analysis
Reduction
in inertia
results in
more
crank
motion at
low speed
42

Torsional vibration analysis

ENGDYN 3D crankshaft
dynamics analysis shows
significant increase in
crankshaft twist for final
design
Baseline crank natural
frequency of 1317 Hz
Final crank natural
frequency of 971 Hz

43

4.5 order peak

Stress analysis
Finite element analysis was
performed on the baseline and
final crankshafts
ENGDYN used to
Calculate boundary conditions
Combine FE models
Solve equations of motion
Calculate combined stresses at
5 degree intervals for each
engine speed
Calculate Goodman safety
factors at fillets and oil holes
44

Stress analysis
UTS 1050 N/mm2
Yield strength 900 N/mm2
Fatigue strength estimated
accounting for influence of
nitriding and size effect
At pin fillets 745 N/mm2
At main fillets 747 N/mm2
At pin oil holes 745 N/mm2
Baseline results indicate that
lowest safety factor occurred at
crank pin fillet on web No.1
45

Stress analysis
Results compared for pin fillet
at web No. 1
Lower safety factor for
intermediate design
Lowest value for final design at
4.5 order resonance at ~13500
rpm
Intermediate crank design did
fail at pin fillet on Web No.1

46

Practical experience
Track testing was performed with engines having various
degrees of unbalance
Riders preferred low inertia of final design
Riders were prepared to tolerate increased vibration
Crankshaft was very durable
No failures of baseline or final design on test or during
racing
Crankshaft was usually replaced after 4 million cycles
some baseline cranks experienced 7 million cycles
some final cranks experienced 6 million cycles
No significant wear of main bearings
47

Conclusions
The final crankshaft design
had exceptional durability even when rev limiter was
set to 16000 rpm despite considerable increase in twist
due to torsional vibration
had partially balanced primary reciprocating moment
was guided by analysis using Ricardo Software

48

Thank you for your attention

Any questions ?
49