Professional Documents
Culture Documents
Design of A Suspension For A Formula Student Race Car - Adam Theander
Design of A Suspension For A Formula Student Race Car - Adam Theander
VEHICLE DYNAMICS
AERONAUTICAL AND VEHICLE ENGINEERING
ROYAL INSTITUTE OF TECHNOLOGY
TRITA-AVE-2004-26
ISSN 1651-7660
Postal Address
Visiting Address
Internet
Telephone
Telefax
KTH
Vehicle Dynamics
SE-100 44 Stockholm
Sweden
Teknikringen 8
Stockholm
www.ave.kth.se
+46 8 7906000
+46 8 7909290
Abstract
In July of 2004 KTH Racing will attend at the Formula Student event in England. The
Formula Student event is a competition between schools that has built their own
formula style race cars according to the Formula SAE rules. In January of 2004 the
Formula Student project started at KTH involving over seventy students. The aim of
this thesis work is to design the suspension and steering geometry for the race car being
built. The design shall meet the demands caused by the different events in the
competition. The design presented here will then be implemented into the chassis being
built by students participating in the project. Results from this thesis work shows that
the most suitible design of the suspension is a classical unequal length double A-arm
design. This suspension type is easy to design and meets all demands. This thesis work
is written in such a way that it can be used as a guidebook when designing the
suspension and steering geometries of future Formula Student projects at KTH.
Acknowledgements
This master thesis has been conducted at the Division of Vehicle Dynamics, Department
of Aeronautical and Vehicle Engineering at the Royal Institute of Technology, KTH, in
Stockholm, Sweden. The work has been carried out from December 2003 to May 2004.
There are a few persons to whom I would like to especially express my gratitude.
Professor Annika Stensson, my examiner, who gave me the opportunity to carry out this
thesis work; research engineer Mats Beckman, my supervisor, for his engagement and
time spent helping me; Fredrik Westin, Ph.D student at Division of Internal Combustion
Engines, who has had a major role in KTH Racing as project leader and spent almost all
of his spare time working with the project; all students participating in the KTH
Formula Student project, without all of you there wouldnt have been any KTH Racing.
Finally I would like to take the opportunity to give a special thank to the students
known as Jrngnget, you know who you are. Without their effort the last couple of
weeks there would never have been a car to show on the 14th of May.
Stockholm, May 2004
Adam Theander
Table of Contents
1
Introduction ..................................................................................... 12
1.1
Background..................................................................................................12
1.2
Aim of the work...........................................................................................12
1.3
Competition Objective .................................................................................12
1.3.1 Vehicle Design Objectives .......................................................................13
1.4
Competition Events and Judging of the Cars ................................................13
1.4.1 Acceleration Event...................................................................................14
1.4.2 Skid-Pad Event ........................................................................................14
1.4.3 Autocross Event .......................................................................................15
1.4.4 Endurance and Fuel Economy Event ........................................................15
1.4.5 Judging of the cars ...................................................................................16
1.5
Rules Relevant to the Chassis Design...........................................................17
1.5.1 Wheelbase and Vehicle Configuration......................................................17
1.5.2 Vehicle Track Width ................................................................................17
1.5.3 Ground Clearance ....................................................................................17
1.5.4 Wheels and Tires......................................................................................17
1.5.5 Suspension ...............................................................................................17
Wheelbase ...................................................................................................18
Track Width.................................................................................................19
Kingpin and Scrub Radius............................................................................20
Caster and Trail............................................................................................21
Instant Centre and Roll Centre .....................................................................21
Tie Rod Location .........................................................................................22
Anti Features ...............................................................................................23
Ackerman steering .......................................................................................25
Camber ........................................................................................................26
Toe ..............................................................................................................27
Benchmark....................................................................................... 28
Methods............................................................................................ 30
4.1
Track width and wheelbase ..........................................................................30
4.2
Front Suspension Design..............................................................................33
4.2.1 The Rims .................................................................................................33
4.2.2 The Brakes...............................................................................................34
4.2.3 Front View Geometry...............................................................................34
4.2.4 Side View Geometry ................................................................................35
4.2.5 Control Arm Pivot Axis ...........................................................................36
4.2.6 Tie Rod Location and Ackermann Geometry............................................38
4.3
Rear Suspension Design...............................................................................38
Parameter Study.............................................................................. 46
6.1
The Taguchi Methods ..................................................................................46
6.2
Parameters of Interest...................................................................................47
6.2.1 Parameter Levels......................................................................................47
6.3
Results .........................................................................................................48
6.3.1 Parameter Study of Front Suspension .......................................................48
6.3.2 Parameter Study of Steering Geometry.....................................................53
6.3.3 Parameter Study of Rear Suspension ........................................................57
Nomenclature................................................................................... 72
10 References ........................................................................................ 74
10.1
10.2
Literature .....................................................................................................74
Oral References ...........................................................................................74
1 Introduction
1.1 Background
In the autumn of 2003 a group of students started a project at KTH. The objective of the
project was to build a race car according to the Formula Student rules and compete in
the event at Bruntingthorpe Proving Ground, Leicestershire, England, in July 2004.
There were three different courses given involved in the Formula Student project, one
project course in Internal Combustion Engines, one project course in Advanced
Machine Elements and a small project course in Machine Design. Soon there were over
70 students involved, either participating in one of the three courses or as volunteers.
The work carried out will be used by the students participating within the KTH Racing
project for year 2004 and hopefully be of much interest for upcoming years projects.
Presentation
Engineering Design
Cost Analysis
A presentation is held for the imaginary manufacturing firm who ordered the prototype.
The purpose of the presentation event is to evaluate the teams ability to sell their
product. The presentation judges evaluate the organization, content and delivery of the
presentation. An engineering design event is held to evaluate the effort put into the
design process and how the design meets the intent of the market. The purpose of the
cost analysis event is to teach the students participating that cost and budget are very
important and must be taken into account in every engineering process.
The dynamic events are:
Acceleration
Skid-Pad
Autocross
Endurance and Fuel Economy
3m
Finish
15
.2
5m
18.25m
Start
The procedure of the event is as follows: the cars will start by entering the right circle
completing one lap. Next lap will be timed and immediately after the left circle is
entered for the third lap. The fourth lap will be timed. Then the driver has the option to
make a second run immediately after the first. Each team will have two drivers who can
do two runs each. The design of the suspension and steering geometry will influence the
performance much.
Straights No longer than 60m with hairpins at both ends or no longer than 45m
with wide turns on the ends.
Constant Turns 23m to 45m in diameter.
Hairpin Turns Minimum of 9m outside diameter.
Slaloms Cones in a straight line with 7.62m to 12.19m spacing.
Miscellaneous Chicanes, multiple turns, decreasing radius turns, etc. The
minimum track width will be 3.5m
Length Approximately 0.805km.
Each team will have two drivers entering the event. Each driver will drive two timed
laps and the best time for each driver will stand as the time for that heat.
Straights No longer than 77m with hairpins at both ends or no longer than 61m
with wide turns on the ends. There will be passing zones at several locations.
Constant Turns 30m to 54m in diameter.
Hairpin Turns Minimum of 9m outside diameter.
Slaloms Cones in a straight line with 9m to 15m spacing.
Miscellaneous Chicanes, multiple turns, decreasing radius turns etc. The
minimum track width will be 4.5m.
In both the autocross event and the endurance event the suspension design and steering
geometry is of major importance. A well working design helps the drivers perform at
the edge of their capacity. The layout of the 2003 event endurance track can be viewed
in Figure 1.2.
75
150
100
75
50
150
50
350
1000
1.5.5 Suspension
Rule 3.2.3: The car must be equipped with a fully operational suspension system with
shock absorbers, front and rear, with usable wheel travel of at least 50.8mm (2 inches),
25.4mm (1 inch) jounce and 25.4mm (1 inch) rebound, with driver seated [1].
2.1 Wheelbase
The wheelbase, l, is the distance between the centre of the front axle and the centre of
the rear axle. The wheelbase has a big influence on the axle load distribution. A long
wheelbase will give less load transfers between the front and rear axles than a shorter
wheelbase during acceleration and braking according to Equation 2.1 and Figure 2.1.
Figure 2.1. Side view parameters for longitudinal load transfer calculations.
Fz1 = (1 ) mg + a x m
Fz 2 = mg + a x m
(2.1)
A longer wheelbase will therefore be able to be fitted with softer springs and will
increase the level of comfort for the driver. On the other hand a shorter wheelbase have
the advantages of smaller turning radius for the same steering input, see section 2.8 [3].
A car with too short wheelbase may act nervously on corner exits and in straight line
driving. Anti features can be built into a suspension and these will also affect the
longitudinal load transfer, see section 2.7.
Fz 2 =
lat hCG
tw2
(2.2)
A larger track width has the disadvantage that more lateral movement of the vehicle is
needed to avoid obstacles. According to the regulations the smallest section of the SkipPad may not be smaller than 3m and the Autocross and Endurance tracks no smaller
than 3.5m [1]. The amount of lateral load transfer wanted depends on tires fitted on the
car, see section 2.9. If the car has anti-roll bars these will also affect the load transfer.
Kingpin Inclination
Kingpin Axis
Wheel Offset
Spindle Length (+)
UBJ
UBJ
LBJ
LBJ
Caster (+)
Mechanical Trail
FORWARD
Figure 2.2. Kingpin geometry, side view and front view.
If the spindle length is positive the car will be raised up as the wheels are turned
and this results in a increase of the steering moment at the steering wheel. The
larger the kingpin inclination angle is the more the car will be raised regardless
of which way the front wheels are turned. If there is no caster present this effect
is symmetrical from side to side. The raise of the car has a self-aligning effect of
the steering at low speeds.
Kingpin inclination affects the Steer camber. When a wheel is steered it will
lean out at the top, towards positive camber if the kingpin inclination angle is
positive. The amount of this is small but not to neglect if the track includes tight
turns.
If the driving or braking force is different on the left and right side this will
introduce a steering torque proportional to the scrub radius, which will be felt by
the driver at the steering wheel.
Centre of Car
Instant Centre
Roll Centre
The camber change rate is a function only of the front view swing arm length, fvsa
length. Front view swing arm length is the length of the line from the wheel centre to
the instant centre when viewed from front. The amount of camber change achieved per
mm of ride travel would be as described in Equation 2.3 and Figure 2.3.
degrees
mm
= arctan
1
fvsa length
(2.3)
The camber change is not constant throughout the whole ride travel since the instant
centre also moves with wheel travel.
Braking Force
CG
+ FZ
FZ
Braking Force=
W(ax/g)
CG
svsa length
Moment=
W(ax/g)(% front braking)(svsa height)
IC
svsa
height
( h)
(2.4)
By substituting % front braking with % rear braking and tan F tan with R in Equation
2.4 the amount of anti lift can be calculated. The way that brake and drive torque is
reacted by the suspension alters the way to calculate the amount of anti present. If the
control arms react torque, either from the brakes or from drive torque, the antis are
calculated by the IC location relative to the ground contact point. If the suspension
doesnt react drive or brake torque, but only the forward or rearward force, then the
antis are calculated by the IC location relative to the wheel centre. For a rear-wheel
driven car there are 3 different types of anti features:
Anti dive, which reduces the bump deflection during forward braking.
Anti lift, which reduces the droop travel in forward braking.
Anti squat, which reduces the bump travel during forward acceleration.
Figure 2.5 shows the configuration for calculating the anti features for a car with
outboard front brakes and inboard rear brakes.
CG
l
IC
h
IC
R
svsa
heights
svsa length
Figure 2.5. Anti features during braking with outboard front brakes and inboard rear brakes.
a =
l
R
(2.5)
If both front wheels are tangents to concentric circles about the same turning centre,
which lays on a line trough the rear axle, the vehicle is said to have Ackermann steering.
This results in the outer wheel having a smaller steering angle than the inner. If both
wheels have the same steering angle the vehicle is said to have Parallel steer and if the
outer wheel has a larger steering angle than the inner it is called Reverse Ackermann.
Passenger cars have a steering geometry somewhere between Ackermann steering and
parallel steering while its common among race cars to use reverse Ackermann. By
using Ackermann steering on passenger cars, or other vehicles only exposed to low
lateral accelerations, it is ensured that all wheels roll freely with no slip angles because
the wheels are steered to track a common turn centre. Race cars are often operated at
high lateral accelerations and therefore all tires operate at significant slip angles and the
loads on the curve inner wheels are much less than the curve outer wheels due to the
lateral load transfer. Tires under low loads require less slip angle to reach the peak of
the cornering force. Using a low speed steering geometry on a race car would cause the
curve inner tire to be dragged along at much higher slip angles than needed and this
would only result in raises in tire temperature and slowing down the car due to the slip
angle induced drag. Therefore race cars often use parallel steer or even reverse
Ackermann. The different types of Ackerman are shown in Figure 2.6.
Ackermann
Parallel
Reverse
Ackermann
2.9 Camber
Camber angle is the angle between a tilted wheel plane and a thought vertical plane.
Positive camber is defined as when the wheel is tilted outwards at the top relative to the
car. The camber angle has influences on the tires ability to generate lateral forces. A
cambered rolling pneumatic wheel produces a lateral force in the direction of the tilt.
This force is referred to as Camber thrust when it occurs at zero slip angles. Camber
also affects the aligning torque due to distortion of the tire print. The effect of this is
rather small and tends to be cancelled with increasing slip angle. Cambering the wheel
also leads to a raise in the lateral force produced by the wheel when cornering. This is
true in the linear range of the tire. If the linear range is exceeded the additive effects of
the camber inclination decreases, this effect is called Roll-off. Therefore the difference
in lateral force when comparing a cambered wheel and a non-cambered wheel is small,
around 5-10% at maximum slip angle. The difference is much larger at zero degrees slip
angle due to the camber thrust. The effects of cambering the tyre are bigger for a bias
ply tyre than a radial ply tyre. For radial tyre the camber forces tends to fall of at camber
angles above 5 while the maximum force due to camber for a bias ply racing tyre
occurs at smaller angles.
Camber Thrust
Lateral Force
Negative Cambered
No Roll-Off
Negative Cambered
With Roll-Off
0 Camber
Slip Angle
2.10 Toe
Toe adjustment can be used to overcome handling difficulties in the car. Rear toe-out
can be used to improve the turn-in. As the car turns in the load transfer adds more load
to the outside wheel and the effect is in an oversteer direction. The amount of static toe
in the front will depend on factors such as Ackermann steering geometry, ride and roll
steer, compliance steer and camber. Minimum static toe is desirable to reduce rolling
resistance and unnecessary tyre heating and tyre wear caused by the tires working
against each other.
3 Benchmark
2004 is the first year the Formula Student project is held at KTH there are no previous
experience from cars to evolve from. To get a rough estimation about the dimensions
and weights of Formula Student race cars data was collected from the 2003-year event.
The data collected was wheelbase, track width and weight and can be viewed in Table
3.1.
Car Nr.
Track Width
Rear [mm]
1200
1360
1200
1240
1240
1060
1200
1194
1000
1160
1185,4
Result
2003
6
35
8
19
7
30
21
17
10
20
Table 3.1. Average dimensions and weights of Formula Student race cars at the 2003 event.
A major part of the cars participating in the 2003 event didnt have all data available so
the data listed in Table 3.1 are only from the cars where all dimensions and weights
where available.
Based on the literature survey, knowledge about the cars from the 2003 event and
discussions with persons with good knowledge in vehicle dynamics and racing a
guideline for the race car was set up. The purpose of this guideline is to have a defined
goal for the work. The guidelines set up was:
The kingpin inclination is kept below 8 since too much kingpin inclination causes a lot
of rising of the front axle when steering. Keeping scrub radius small will make the car
easier to handle at low speeds and reduces the risk that a sudden lost of traction for one
of the front wheels during braking causes the car to change direction and reduces the
steering moment disturbance. The caster angle has positive effects during cornering but
too much caster causes weight transfer that will have an oversteering effect. The
possibility to adjust the camber angle from 0 to about -4 will be very helpful during
the testing of the car. During the competition this also allows to set the camber to 0
during the acceleration event minimizing the rolling resistance. The camber gain is to
compensate for the lost of camber due to the roll angle during cornering. The reason for
having a much larger camber gain at the rear axle is to have as big contact patch
between the rear tyre and the ground during corner exits as possible. This will allow the
driver give more and earlier throttle. Having a slightly higher roll centre at the rear has
at least two advantages. The first is that softer springs can be used at the rear axle since
less rolling moment will appear here. The second is to keep the roll axle as parallel to
the cars main inertia axle.
4 Methods
4.1 Track width and wheelbase
The track width and wheelbase will have influences on the amount of load transfer
between the front and rear axle during acceleration and braking and the load transfer
from curve inner to curve outer wheels during cornering. To investigate the interactions
between longitudinal load transfer and wheelbase a MATLAB code was used. This code
doesnt take the alternation of the centre of gravitys height due to the pitch-attitude into
consideration since this is small and can be neglected. In Figure 4.1 the load transfer
from the front axle to the rear axle during acceleration is presented for two different
wheelbases, 1525mm and 1700mm as an example of the output from the MATLAB
code. The methods for determine the wheelbase is primary based on the packaging as it
will decide how small the wheelbase can be. The wheelbase should be as short as
possible, but not shorter than 1525mm, to optimize the ability to make sharp corners.
The load distrubation between curve inner and outer wheels during high speed cornering
is among other things a function of the front and rear trackwidth. Due to the lateral
acceleration there will be a load transfer from the curve inner to the curve outer wheels
present. In Figure 4.2 the MATLAB output for a steady state cornering simulation can
be viewed.The load on the curve inner wheels can be viewed as a function of trackwidth
and lateral acceleration. These plots are then compared with tyre data obtained from
Goodyear Racing showing how the vertical tyre loads influnces on the tyres ability to
produce lateral forces. An example of a tyre plot obtained from Goodyear Racing can be
viewed in Figure 4.3.
Figure 4.2. Lateral weight transfer as a function of track width and lateral acceleration.
Figure 4.3. Lateral force as a function of tire load and slip angle
for a 20x8-13 Goodyear racing tire for light vehicles [6].
Another factor that will determine the track width is the size of the race track.
According to the regulations no part of the race track may be smaller than 3.5m and the
tightest hairpin may not have an outside diamter smaller than 9m. The track will be 3m
wide in the Skid-Pad event but here the track width is of less importance. The turning
radius for a car having ackermann steering geometry is at low speed proportional to the
wheelbase and the steering angle. To make a hairpin that have an outside diameter of
9m the turning radius of the cars centerline need to be 9m minus half the trackwidth.
This results in an Ackermann angle of:
a =
l
r
tw
(4.1)
A larger trackwidth will have the disadvantage of a narrower A-arm angle to allow the
required wheel angle, causing the A-arm taking up more forces in the longitudinal
directions.
Front View
Side View
Rendered CAD-model
1 .
fvsa =
tw 2
tw 2
=
1 roll camber 1 wheel camber angle chassis roll angle
(4.2)
A line is projected from the ground contact patch through the wanted roll centre and to
the instant centre located at the desired fvsa length distance from the ground contact
patch. From the instant centre one line is projected back to the location of the lower ball
joint and one line to the location of the upper ball joint.
The length of the lower control arm shall be made as long as possible but is limited by
packaging. The drivers legs will have to be fitted in between the lower control arms in
order to keep the centre of gravity height as low as possible. The length of the upper
control arm will determine the curvature of the camber curve. If upper and lower control
arms are of the same length the camber curve will be a straight vertical line and if the
upper arm is shorter than the upper the curve will be concave toward negative camber
which is preferable. The shorter the upper arm is the more concave the camber curve
will be. It is possible to design geometry that will have progressive camber in bump
with much less in drop.
+
+
tw/2
fvsa length
Figure 4.5. Front view swing arm geometry for establishing locations of ball joints.
ax
CG
IC
svsa
height
svsa length
l
Figure 4.6. The side view swing arm geometry for establishing wanted amount of ant feature.
15
12
13
11,12
13
14
xis
a
LCA
UCA axis
LCA axis
14
UCA axis
Car Centreline
11
IC
IC
15
13
13
14
12
UCA axis
LCA axis
Car Centreline
Figure 4.7. Control arms pivot axis construction layout.
5
15
11
Tie Rod
UCA
axis
Forward
Control
Arms
5 Model Building
In order to evaluate the suspension system when designing according to the discussions
in chapter 4 a model is created to make it possible to simulate the cars dynamic
behaviour during different conditions.
All vehicle simulations are carried out in ADAMS. ADAMS is a simulation software
for dynamic simulation of mechanical systems. There are several subprograms included
in ADAMS for different simulation applications. One of these subprograms,
ADAMS/Car, is especially designed for use with vehicle simulations. In ADAMS/Car
all subsystems of a car can be modelled one at a time and then put together to a
complete vehicle. This also gives the advantage that it is very simply to change from
one model of a front suspension to another or between different tyre models. There are
also built in test rigs for testing suspensions or whole cars. All simulations can be
visualized, written to files or viewed as graphs with the included tools. A complete
vehicle is built up by subsystems. To simulate a vehicle the required subsystems are:
If a more detailed model is wanted, powertrains, brakes, anti-roll bars, and differentials
etc. can be added to the vehicle model. There are several pre-made subsystems included
and the opportunity to build your own subsystems. The subsystems interact with each
other via communicators. There are input communicators that read information into a
subsystem with data and output communicators that send information from a subsystem
to another.
5.5 Body
The body subsystem used consists only of a point mass in the centre of gravity.
Communicators act between the body subsystem and steering subsystem, front
suspension subsystem and rear suspension subsystem.
5.6 Simulation
A simulation is very easily done with ADAMS. There are a number of different pre-set
simulation modes that can be used such as a suspension analyser where different kinds
of wheel travel and steering simulations can be run. There are also full vehicle
simulation models available such as ISO lane change and steady state cornering. A
problem when simulating with ADAMS is the sensitivity for equilibrium problems with
ill-defined models.
6 Parameter Study
Studies were made to investigate the influences of different settings of system
parameters such as kingpin inclination angle, caster angle, roll centre height. Knowing
how these parameters influences and interacts made it possible to improve the models
until the guidelines were fulfilled. The parameter study also reveals the interactions
between parameters. This information is lost when performing tests using the one
parameter at the time approach. This kind of information is very useful when changing
one parameter as this can lead to unexpected effects on other parameters.
V: All main effects are isolated and two factor interactions are isolated from
other two factor interactions.
IV: All main effects are isolated from two factor interactions, but two factor
interactions can be mixed with other two factor interactions.
III: Main effects and two factor interactions can be mixed.
The size of the orthogonal matrix to obtain a certain resolution can be viewed in Table
6.1.
Resolution
Matrix
III
IV
V
7
4
3
L8
15
8
5
L16
31
16
6
L32
Table 6.1. Maximum of two level factors as a function of resolution and matrix size.
Trial no.
1
2
3
4
5
6
7
8
A
-1
1
-1
1
-1
1
-1
1
B
-1
-1
1
1
-1
-1
1
1
Parameters
AxB
C
1
-1
-1
-1
-1
-1
1
-1
1
1
-1
1
-1
1
1
1
AxC
1
-1
1
-1
-1
1
-1
1
BxC
1
1
-1
-1
-1
-1
1
1
The orthogonal matrix should be filled with low levels replacing the minus signs and
high levels replacing the plus signs. Trials are then performed with accurate settings for
the parameters of each row in the orthogonal matrix. The desired quantify is measured
for every attempt. When all trials are performed the column for each parameter is
multiplied with the result column and then divided by the number of plus signs. This
gives the influence of the parameter. The influence is a measure of the product of the
selected parameter interval and the quantity measured in the trials. Table 6.2 shows the
L8 matrix with resolution V which were used for most of parameter studies conducted.
The Taguchi methods are often used in tests generating one result per trial. It is possible
to use the Taguchi method in tests generating more than one result per trial such as time
dependent tests. This gives the possibility to study how the parameters influence the
sought variable over a specified interval. This approach is presented by Beckman and
Agebro in their thesis work, A Study of Vehicle Related Parameters Influencing the
Initial Phase of Ramp Rollovers, and is called the Continuous Taguchi Method [5].
By using the Taguchi method the design parameters are investigated to find important
interactions and which parameters that influences on the behaviour of the system
analyzed.
6.3 Results
The parameter study was an iterative process. After each iteration the results were
analysed and the models used improved until the guidelines were fulfilled. The results
presented below, Figure 6.2 to Figure 6.24, are from the final design and the adjustment
levels are the wanted. The upper part of each figure shows how the adjustment levels
affect the studied parameter and the lower part of the figure shows the unchanged
design. By adding the results presented in the upper part of each figure to the unchanged
set-up in the lower part the same figure the characteristics after made changes is
received.
Explanation
Height to the inner joints for the upper control arms [mm]
Height to the inner joints for the lower control arms [mm]
Anti-Dive adjustment of the rear lower control arm [mm]
Figure 6.1. Parameters used in the parameter study of the front suspension.
In Table 6.4 the chosen low and high levels of the parameters are shown. The mounts
on the frame will also have levels between the low and high levels used in the parameter
study.
Parameter
A
B
C
Low Level
300.0
119.0
0.0
High Level
320.0
139.0
20.0
Unit
[mm]
[mm]
[mm]
Table 6.4. Levels used in the parameter study of the front suspension.
Figure 6.2. Results from parameter study of Anti Dive characteristics during roll.
Figure 6.3. Results from parameter study of the camber angle of left wheel during roll.
Figure 6.4. Results from parameter study of roll centre vertical travel during roll.
Figure 6.5. Results from parameter study of the lateral travel of roll centre during roll.
Figure 6.6. Results from the parameter study of toe angle variation during roll.
Figure 6.7.Results from parameter study of lateral wheel travel at track during roll.
Figure 6.8. Results from parameter study of scrub radius variation during roll.
A
C
Explanation
Location of outer tie rod end in Y-direction [mm]
Location of inner tie rod end in X-direction [mm]
Location of inner tie rod end in Z-direction [mm]
Parameter
A
B
C
Low Level
572.0
-40.0
325.0
High Level
592.0
-60.0
335.0
Unit
[mm]
[mm]
[mm]
Table 6.6. Levels used in the parameter study of the front suspension.
Figure 6.10. Results from parameter study showing camber as function of wheel travel.
Figure 6.11. Results from parameter study of toe angle variations as function of wheel travel.
Figure 6.12. Results from parameter study of camber characteristics for curve outer wheel as function
of steering rack displacement.
Figure 6.13. Results from parameter study of camber characteristics for curve inner wheel as function
of steering rack displacement.
Figure 6.14. Results from parameter study of outside turn diameter as function of steering rack
displacement.
Figure 6.15. Result from parameter study of percent Ackerman as function of steering rack
displacement.
Explanation
Height to the inner joints for the upper control arms [mm]
Height to the inner joints for the lower control arms [mm]
Anti-Dive adjustment of the rear lower control arm [mm]
Parameter
A
B
C
Low Level
300.0
119.0
0.0
High Level
320.0
139.0
20.0
Unit
[mm]
[mm]
[mm]
Table 6.8. Levels used in the parameter study of the rear suspension.
Figure 6.17. Result from parameter study of anti-squat as function of roll angle.
Figure 6.18. Result from parameter study of camber for curve inner wheel during roll.
Figure 6.19. Results of parameter study of camber for curve outer wheel during roll.
Figure 6.20. Result from parameter study of roll centre vertical travel during roll.
Figure 6.21. Result from parameter study of roll centre lateral travel during roll.
Figure 6.22. Result from parameter study of toe angle variation of curve inner wheel during roll.
Figure 6.23. Result from parameter study of toe angle variation of curve outer wheel during roll.
Figure 6.24. Result from parameter study of lateral wheel travel at wheel centres.
1700
= 0.439 rad = 25.14
r 2 tw 4500 12 1250
l
1700
=
=
= 0.444 rad = 25.46
1
r 2 tw 4500 12 1350
(7.1)
During a 1G turn the change in vertical force on the curve inner wheel is 275N with a
trackwidth of 1250 mm and 254N with a trackwidth of 1350mm, in kilograms 28kg and
26kg. The difference may not be big but since a tyres ability to produce lateral force is,
among other parameters, a function of the tyre load a few kilos can be enough to pass
beyond the peak of the tyre force curve, the tyre will be overloaded. Since the very low
weight of a formula student car there are no tyres avivible developt especially for this
kind of light weight vehicles, therefore the risk of overloading a tyre due to lateral load
transfer is very low. The problem for this kind of light weight vehicle is the other way
around, the load transfer may lead to loss in lateral force on the curve inner wheel due to
underload.
In Figure 3.3 it can be seen that an increase in tyre load from 125lbs to 250lbs, an
increase of 100%, results in an 115% increase of the lateral force at 10 slip angle for a
20x8-13 Goodyear racing slick. The reason to use a wider trackwidth will not be to
eliminate the risks for a tyre overload, it is to prevent a tyre underload to happen.
The trackwidth is choosen to 1250mm in the front and 1200mm at the rear. The main
reason for having a smaller rear trackwidth is that wider tyres are fitted in the rear than
the front, the front tyres will be 20x6.2-13 and the rear tyres 20x7.2-13, a difference
of 1.0 or 25.4mm. Having the same trackwidth front and rear would cause the rear
tyres inner line to be closer to the inside of the curve than the front tyres. This could
lead to the driver knocking down the cones that marks the track with his rear tyres
trying to take the shortest way possible only looking on his front wheels. Having a
bigger front trackwidth will also have the advantage of letting the frontaxle taking up a
bigger part of the rolling moment. The effect of this is that softer springs can be used at
the rear optimizing the rear traction and allowing more and earlier throttle on corner
exits.
7.2
The final design of the front uprights was set in cooperation with the MME-students
designing the uprights and the Vehicle Dynamics-students designing the brake system
due to the packaging issues. The locations of the joints at the uprights can be found in
Table 7.1 and joints on the frame in Table 7.2. Figure 7.4 shows a model with the joints
marked.
Name
Wheel Centre
LCA outer
UCA outer
Tie Rod outer
X-location
0.0
0.0
20.0
-30.0
Y-location
645.0
607.0
572.0
572.0
Z-location
254.0
159.0
374.0
374.0
Adjustability
Y: +20.0
[mm]
Name
LCA front
LCA rear
UCA front
UCA rear
Tie Rod inner
X-location
-10.0
260.0
0.0
230.0
-40.0
Y-location
150.0
150.0
200.0
200.0
230.0
Z-location
119.0
119.0
300.0
300.0
305.0
Adjustability
Z: +20.0
Z: +20.0, +20.0*
Z: +20.0
Z: +20.0
Z: +20.0
[mm]
Table 7.2. Location of front suspension joints at the frame. * for Anti Dive adjustment
UCA outer
Tie Rod outer
UCA rear
Wheel Centre
LCA outer
Z
X
LCA front
LCA rear
UCA front
Tie Rod inner
The corresponding geometries of the front suspension can be viewed in Table 7.3.
Parameter
Anti Dive
Kingpin Inclination
Scrub Radius
Caster Angle
Trail
Camber Gain in Roll
Roll Centre Height
Value
0.0
9.2
4.6
5.3
19.1
0.32
13.8
Unit
[%]
[degrees]
[mm]
[degrees]
[mm]
[Camber/1Roll]
[mm]
By using the adjustment levels some parameters can be changed inside the intervals
listed in Table 7.4.
Parameter
Anti Dive
Camber Gain in Roll
Roll Centre Height
Ride Height
Camber Angle
Toe Angle
Low Value
0.0
0.16
2.8
25.0
-4.0
-2.0
High value
52.4
0.32
45.7
55.0
0.0
2.0
Unit
[%]
[Camber/1Roll]
[mm]
[mm]
[degrees]
[degrees]
The final design of the front suspension is a trade of between performance and
manufacturability. The hardest and most time-consuming part has been the design of the
front uprights. Due to lack of space available in the front rims the final design is a little
bit off from the design wanted. This mostly affects the kingpin inclination angle and the
scrub radius. It was desired to keep the kingpin inclination angle below 8.0 degrees but
the final design has a kingpin inclination angle of 9.2 degrees and this leads to a scrub
radius of 4.6mm.
is added. Adding anti dive to the front suspension also gives some other effects. Adding
around 50% anti dive will decrease the camber gain with around 0.05 degrees Camber
per degree Roll, lower the roll centre height with about 20mm, increase the lateral
movement of the roll centre with 80mm at one degree roll and will increase the roll steer
with 0.02 degrees at one degree roll.
7.3 Steering
The purpose of the steering geometry design was to minimize bump steer and have the
possibility to adjust toe angles and the Ackermann geometry. Adjustable outer tie rod
joints make the adjustability of Ackermann geometry possible. Another criterion that
the steering system had to fulfil was that the car had to manage a hairpin with an outer
diameter of 9m. Table 7.1 and Table 7.2 show the locations of the joints in the steering
system and the adjustability levels wanted. Table 7.5 shows the corresponding levels of
adjustability for Ackerman geometry and toe angles.
Parameter
Toe Angles
Percent
Ackermann
Low Value
-2.0
-9.4/-23.2
7.3.2 Ackerman
By adjusting the outer ends of the tie rods in Y-direction the amount of Ackerman can
be changed. By moving the outer end of the tie rod 20mm further out from the
centreline of the car the Ackerman is raised from around 0% to 65% at a steering angle
corresponding to a curve with a radius of 11.5m. By moving the steering rack forward
the Ackerman geometry can be reduced. A movement of 20mm forward will result in
10% less Ackerman for the 11.5m radius curve. Moving the inner end of the tie rods up
from the plane made by the upper A-arms in the front view will not affect the
Ackerman. The parameter study also shows that with a steering rack movement of
30mm the outside turn diameter is 9180mm with the Ackerman set to the low value and
8180mm with the Ackerman set to the high value. At 35mm steering rack movement the
corresponding values are 8040mm and 6850mm. The reasons for the smaller outer turn
diameter with Ackerman set to high value is not the change in percent Ackerman, the
reason is that the change made to the steering geometry also changes the steer camber
characteristics as shown in Figure 5.12 and Figure 5.13. With Ackerman set to low the
change in camber of the curve outer wheel is -1.1 degrees at 35mm steering rack
movement and -1.3 degrees with Ackerman set to the higher value. For the curve inner
wheels the values are +5.8 degrees and +7.7 degrees, both curve inner and curve outer
wheels are leaning into the curve as shown in Figure 7.5.
Figure 7.5. At big steering angles both curve inner and curve outer wheel leans into the curve.
X-location
1650.0
1590.0
1650.0
1650.0
Y-location
600.0
590.0
590.0
590.0
Z-location
250.0
140.0
370.0
140.0
Adjustability
[mm]
The locations of the joints on the frame can be read in Table 7.7. The toe link
configuration at the rear suspension is a so-called ungrounded toe link; it is connected to
the lower control arm instead of connected to the frame.
Name
LCA front
LCA rear
UCA front
UCA rear
X-location
1550.0
1850.0
1550.0
1850.0
Y-location
150.0
150.0
300.0
300.0
Z-location
109.0
109.0
250.0
250.0
Adjustability
Z: +20.0
Z: +20.0
Z: +20.0
Z: +20.0
[mm]
The corresponding geometries of the rear suspension can be viewed in Table 7.8, which
shows the parameters at the unchanged set up.
Parameter
Anti Squat
Anti Lift
Kingpin Inclination
Scrub Radius
Caster Angle
Cambergain positive Roll
Cambergain negative Roll
Roll Centre Height
Value
0.0
0.0
0.0
10.0
14.6
0.73
1.05
76.9
Unit
[%]
[%]
[degrees]
[mm]
[degrees]
[Camber/1Roll]
[Camber/1Roll]
[mm]
Camber Gain positive Roll refers to the curve outer wheel measured at 0.5 degrees
roll angle at contact patches and Camber Gain negative Roll refers to the curve inner
wheel.
By using the adjustment levels in Table 7.7 some parameters can be changed inside the
intervals listed in Table 7.9, which also shows the adjustment levels for static camber,
toe angles and ride height.
Parameter
Anti Lift
Anti Squat
Camber Gain positive Roll
Camber Gain negative Roll
Roll Centre Height
Ride Height
Camber Angle
Toe Angle
Low Value
0
0
0.58
0.84
53.4
25.0
-4.0
-2.0
High value
29
24
0.84
1.18
107.5
55.0
0.0
2.0
Unit
[%]
[%]
[Camber/1Roll]
[Camber/1Roll]
[mm]
[mm]
[degrees]
[degrees]
Note that anti squat and anti lift cant be changed independently of each other. The
parameters characteristics during operation of the car can be viewed in section 4.4.
8 Future Work
To be able to carry this work further the first thing to do is to take complete measures of
the Formula Student race car built. These measures can then be used to update the
computer models used. With models that correspond to the real car new simulations can
be made and further development is possible. This gives the opportunity to have two
development processes running simultaneously, one on the race car and another in the
computer. This also gives the opportunity to test how well the computer models agree
with the real race car.
One important parameter, not so much discussed in this thesis work, is the tires. As the
tires are the part of the race car having the biggest dynamic influence a lot of work can
be carried out investigating how different tires would affect the handling qualities of the
race car. Different sizes, brands and rubber compounds are available at the market.
Testing different tires against each other to find out which tires fits the race car best
could be of bigger importance than improving the design of the suspension. Especially
since the car is raced under such conditions that the wheel travel is very small and not
affecting the suspension geometry much.
If KTH decides to develop a new car for the 2005 Formula Student event the 2004 car is
a very useful tool in the engineering process. The car can be used in different tests and
evaluated to become base to start the new development process from. The computer
model of the car can be updated with drivetrain and brakes and used for evaluation of
the cars dynamic behaviour. By using the computer model unwanted behaviours on the
2004 car can be prevented from appear on the 2005 car as well.
9 Nomenclature
Fzi
l
ax
Fz i
lat
hCG
twi
Vertical axle load, i=1 for front axle and i=2 for rear axle
Wheelbase
Length from front axle to centre of gravity
Acceleration in X-direction
Length from ground up to centre of gravity
Lateral load transfer, i=1 for front axle and i=2 for rear axle
Lateral acceleration
Height to centre of gravity
Track width, i=1 for front axle and i=2 for rear axle
10 References
10.1 Literature
[1]
[2]
[3]
Reimpell, J. & Stoll, H. & Betzler, J.W. (2001), The Automotive Chassis:
Engineering Principles 2nd Edition, Butterworth-Heinemann, England
[4]
[5]
[6]
Race Tire Analysis and Plotting Toolbox Open Wheel Release version 7.2
(2001), The Goodyear Tire and Rubber Company
[7]
[9]