1-1

1. Suspension systems
1.1 Introduction to suspension design
Ride quality and handling are two of the most important issues related to vehicle
refinement. Together they produce some design conflicts that have to be resolved by
compromise. Also, the wide range of operating conditions experienced by a vehicle
affect both ride and handling and must be taken into account by the chassis designer.
This all adds up to some very challenging tasks for the chassis designer.
While the chassis designer has a checklist of functional requirements for a given
design, there are also a number of other constraints that have to be met. These include
cost, weight and packaging space limitations, together with requirements for
robustness and reliability, ease of manufacture, assembly and maintenance.
Suspension design like other forms of vehicle design are affected by development
times dictated by market forces. This means that for new vehicles, refined suspensions
need to be designed quickly with a minimum of rig and vehicle testing prior to launch.
Consequently, considerable emphasis is placed on computer-aided design. This
requires the use of sophisticated mathematical models and computer software that
enables a variety of what-if scenarios to be tested quickly and avoids the need for a
lot of prototype testing.
In order to understand the issues facing the suspension designer it is necessary to have
knowledge of:
- The requirements for steering, handling and stability
- The ride requirements related to the isolation of the vehicle body from road
irregularities and other sources of vibration and noise
- How tyre forces are generated as a result of braking, accelerating and cornering
- The needs for body attitude control
- Suspension loading and its influence on the size and strength of suspension
members.
This module addresses these issues. Further details can be found in a number of
textbooks, [1] to [6], with [1] being particularly appropriate for the course.
1.1.1 The role of a vehicle suspension
The principle requirements for a suspension are:
- To provide good ride and handling performance - this requires the suspension to
have vertical compliance to provide chassis isolation while ensuring that the
wheels follow the road profile with very little tyre load fluctuation.
- To ensure that steering control is maintained during manoeuvring, requiring the
wheels to be maintained in the proper positional attitude with respect to the road
surface.
1-2
- To ensure that the vehicle responds favourably to control forces produced by the
tyres as a result of longitudinal braking and accelerating forces, lateral cornering
forces and braking and accelerating torques - this requires the suspension geometry
to be designed to resist squat, dive and roll of the vehicle body.
- To provide isolation from high frequency vibration arising from tyre excitation -
this requires appropriate isolation in the suspension joints to prevent the
transmission of road noise to the vehicle body.
- To provide the structural strength to resist the loads imposed on the suspension.
It will be seen that these requirements are very difficult to achieve simultaneously
particularly when the added constraints of cost, packaging space, robustness and other
factors are also taken into account. This leads to some design compromises that often
result in less than perfect performance for some of the desired outcomes.
1.1.2 Definitions and terminology
There is a lot of terminology associated with suspension design that may appear novel
for a student meeting the subject for the first time. Most of this will be described as it
arises. The student can find a very useful summary of the most common vehicle
dynamics definitions in reference [2]. Some care is needed here however as there are
some differences between American and European terminology.
Figure 1.1 illustrates whole vehicle reference axes together with terms to describe
rotation about these axes. These axes have their origin at the centre of gravity of the
vehicle. Numerous axis systems are used in vehicle dynamics analysis.
Figure 1.1 Whole-vehicle reference axes
Roll
Pitch
Yaw
1-3
1.1.3 What is a vehicle suspension?
The isolation of a vehicle body from the road undulations that are fed into the tyre of a
vehicle at the road / wheel interface requires relative movement between the wheel
and the body. This motion is, in general, controlled by some form of linkage
mechanism. It is this mechanism that is termed a suspension.
1.1.4 Suspension classifications
In general suspensions can be broadly classified as dependent, independent or semi-
dependent types.
With dependent suspensions the motion of a wheel on one side of the vehicle is
dependent on the motion of its partner on the other side. That is when a wheel on one
side of the vehicle strikes a pothole, the effect of it is transmitted directly to its partner
on the other side. Generally this has a detrimental effect on the ride and handling of
the vehicle.
As a result of the trend to greater vehicle refinement, this type of suspension is no
longer common on passenger cars. However, they are still commonly used on
commercial and off-road vehicles. Their advantages are simple construction and
almost complete elimination of camber change with body roll (resulting in low tyre
wear). They are also commonly used at the rear of front-wheel drive light commercial
and off-road vehicles with rear driven axles (live axles) and on commercial and off-
highway vehicles. They are occasionally used in conjunction with non-driven axles
(dead axles) at the front of some commercial vehicles.
With independent suspensions the motion of wheel pairs is independent, so that a
disturbance at one wheel is not directly transmitted to its partner. This leads to better
ride and handling capabilities. This form of suspension usually has benefits in
packaging and gives greater design flexibility when compared to dependent systems.
Some of the most common forms of front and rear designs will be considered later.
Both McPherson struts; double wishbones and multi-link systems are employed for
front and rear wheel applications. Trailing arm, semi-trailing arm and swing axle
systems tend to be used predominantly for rear wheel applications.
There is also a group of suspensions that fall some way between dependent and
independent ones and are consequently called semi-dependent. With this form of
suspension, the rigid connection between pairs of wheels is replaced by a compliant
link. This usually takes the form of a beam that can twist providing both positional
control of the wheel carrier as well as compliance. Such systems tend to be simple in
construction while having scope for design flexibility when used in conjunction with
compliant supporting bushes.
1.1.5 Defining wheel position
Since one of the most important functions of a suspension system is to control the
position of the road wheels, it is important to understand the definitions relating to
wheel location. The location of the wheels relative to both the road and the vehicle
1-4
suspension are important and will in general be affected by suspension deflection and
tyre loading. In this sub-section we consider additionally the effect of wheel position
on handling behaviour.
Camber angle
This is the angle between the wheel plane and the vertical - taken to be positive when
the wheel leans outwards from the vehicle (Figure 1.2).
Figure 1.2 Wheel position (viewed in x-direction, i.e. from the front of the vehicle)
Camber is affected by vehicle loading and by cornering. A slight positive camber
(0.1) gives more even wear and low rolling resistance. However, on passenger cars
the setting is often negative (even when the vehicle is empty). Front axle values range
from 0 to -1 20. This improves lateral tyre grip on bends and improves handling
because of camber thrust effects (these are a consequence of tyre characteristics).
A disadvantage of an independent suspension is that the wheels incline relative to the
vehicle body on a bend (Figure 1.3). This tends to produce increased negative camber
on the inner wheels and increased positive camber on the outer wheels. The outer
wheels then try to roll towards the outside of the bend tending to produce an
understeer
1
condition. Note: the change of camber (relative to the road) is a
combination of changes due to suspension movement (rebound on the inside, bump on
the outside) and body roll.
Kingpin inclination (steered axles)
Sometimes called swivel pin inclination, KPI is the angle between the kingpin axis
and the vertical (Figure 1.2). It has the effect of causing the vehicle to rise when the
wheels are turned and produces a noticeable self-centring effect for swivel pin
inclinations greater than 15. When the wheels are turned, the magnitude of the self-
aligning moment is dependent on the kingpin angle, the kingpin offset and the caster
angle.
1
Understeer relates to a handling stability condition see sub-section 1.8.1
Kingpin ground offset, d
g
Kingpin
inclination,

Kingpin hub offset, d

h
Camber
angle,

i

o
1-5
Roll angle, |

o
i
Figure 1.3 Camber angles when cornering with an independent suspension
The effects of KPI are as follows:
- Tyre wear - induced camber changes can create wear particularly at high lock
angles
- Returnability
2
- this is improved with increased values of KPI due to work done in
lifting the vehicle from the straight ahead to the turned position.
- Torque steer
3
- this is related to the drive shaft angles.
- Lift-off / traction steer
4
on bends - increasing KPI may result in an increased
vehicle reaction to mid-bend traction changes.
- Steering effort - KPI causes lift of the vehicle as the steering angle is increased.
This results in an increasing amount of work done at the steering system to achieve
a given wheel angle. However an increase of KPI from zero to 15 is unlikely to
increase the steering effort by more than 10%
- Tyre clearance - changes in KPI affect wheel swept envelope and hence tyre
clearance.
Typical kingpin angles for passenger cars are between 11 and 15.5.
Kingpin offset
This is the distance between the centre of the tyre contact patch and the intersection of
the swivel pin axis at the ground plane (Figure 1.2) - taken to be positive when the
intersection point is at the inner side of the wheel. For a given suspension, KPO can
be changed by changing tyre width. Increasing KPO improves the returnability of the
steering (see below). The disadvantage is that longitudinal forces at the tyre contact
patch due to braking, or striking a bump or pothole is transmitted through the steering
mechanism to the steering wheel.
2
Returnability the tendency for the steering system to return to the straight-ahead position after an
induced steering change.
3
Torque-steer the tendency for a drive shaft to induce a change in the steering angle
4
Lift-off / traction steer the tendency for sudden accelerator change to produce a steer effect
1-6
The effects of kingpin offset are thus:
- Returnability - positive increases of KPO will increase returnability due to the
increased lift with steering angle
- Brake stability (outboard brakes) - with a diagonal split braking system, negative
KPO will assist in counteracting the steer effect of a failed system. A similar effect
occurs with split-
5
braking where one wheel locks or with unbalanced front
brakes. For braking on a bend, negative KPO produces toe-in on the outer wheels
creating a tendency to oversteer. (With inboard brakes the hub offset (see below) is
the critical dimension).
- Torque steer - with a conventional non-biased differential ensuring equal torque to
each shaft, if one wheel loses traction, large values of KPO will cause steering
fight.
- Steering effort - it is generally assumed that either positive or negative KPO will
reduce static steering effort by allowing some rolling of the tyre on the road surface
reducing tyre scrub. In practice there appears to be little effect even with large
values of KPO
In practice the KPO varies from small positive to small negative values.
Kingpin hub offset
This is defined as the perpendicular distance from the king pin axis to the intersection
of the hub axis and the tyre centreline as shown in Figure 1.2. Hub offset is defined as
positive when the tyre centre line lies outside the king pin axis at hub centre height.
Castor angle
- Castor angle is the inclination of the swivel pin axis projected into the fore-aft
plane through the wheel centre - positive in the direction shown (Figure 1.4). The
castor angle produces a self-aligning torque for non-driven wheels. It is dependent
on suspension deflection and for steered wheels it influences the camber angle as a
function of steering angle
Castor angle, v
Castor trail, l
t
Kingpin axis
Castor offset, l
off
Figure 1.4 Defining caster angle and mechanical trail
The effects of castor angle are as follows:
5
Split- different coefficients of friction between tyres and road on the right / left
sides of the vehicle
Front of
Vehicle
1-7
- Tyre wear - induced camber changes can create shoulder wear particularly at high
lock angles.
- Steer response - a positive castor angle will generate negative camber on a outside
wheel thus enhancing turn-in response.
Castor trail
Castor trail is the longitudinal distance from the point of intersection of the swivel pin
axis and the ground to the centre of the tyre contact patch - Figure 1.4. It influences
the magnitude of the self-righting moment. On some front wheel drive cars, there is an
increased self-righting moment during cornering due to the offset of the tractive force
and lateral force. This is undesirable in that the understeer effect is too great and the
steering is unduly influenced by rough road surfaces.
The effects of castor trail on vehicle behaviour are as follows:
- Straight line stability - improved with higher castor trail
- Returnability - stronger with higher castor trail
- Braking stability - castor angle will generally reduce with braking (due to hub
wind-up and vehicle pitch) resulting in reduced castor trail leading to a degradation
in braking stability, i.e. braking stability is enhanced with higher levels of castor
trail.
- Steering effort - castor trail has very little effect on steering effort at stationary or at
very low speeds, but otherwise steering effort increases with increasing castor trail.
Steering feel at high lateral accelerations (with impending loss of grip) is also
influenced by castor trail.
Castor offset
Castor offset is the longitudinal distance from the vertical centre line through the
wheel centre to the intersection of the longitudinal axis through the wheel centre and
the king pin axis. It is positive when in front of the wheel centre as shown in Figure
1.4. It is possible to use a negative offset to reduce mechanical trail.
Toe-in / Toe-out
This is the difference between the front and rear distances separating the centre plane
of a pair of wheels (quoted at static ride height and measured to the inner rims of the
wheels). Toe-in is when the wheel centre planes converge towards the front of the
vehicle as shown in Figure 1.5. Braking and rolling resistance forces tend to produce a
toe-out effect, while tractive forces (front wheel drive vehicles) tend to produce the
opposite effect. This leads to front wheels being set to toe-in for undriven and driven
front axles. In the later case this is to ensure driving stability in those cases when the
driver suddenly takes his foot off the accelerator. With independent front suspensions
body roll can produce changes of toe and hence roll steer
6
. This is discussed later.
6
Roll steer a tendency for the wheels to steer as result of suspension articulation arising from body
roll
1-8
Figure 1.5 Toe-in definition (plane view)
1.1.6 Tyre loads
Suspension loads result from the set of forces and moments that are generated at the
tyre / road interface during vehicle motion. These forces depend on the static loads
produced by masses of the vehicle, occupants and payload, together with the dynamic
forces arising from accelerating, braking, cornering, tyre rolling resistance and
aerodynamic effects. These combined forces affect the handling ability of the vehicle
and it is the chassis designers task to ensure that their effects are satisfactorily
controlled.
The tyre contact patch
Due to tyre compliance, the tyres deform to produce a contact area over which tyre
load is distributed. This is called the tyre contact patch. For static loading conditions
this is a symmetrical area as shown in Figure 1.6a. The centre of pressure in this case
lies vertically below the axle centre. During motion of the vehicle, the tyre contact
patch becomes distorted depending on whether the vehicle is accelerating, braking,
cornering etc. These actions produce a distributed set of normal and shear forces in the
tyre contact patch.
Vertical forces
Under static conditions, the load distribution in a vehicle determines the vertical load
supported by each wheel. These forces are altered by a process called load transfer.
During acceleration the vertical load increases on the rear wheels and decreases on the
front wheels. The reverse happens during braking. During cornering there are lateral
(centrifugal force) forces exerted on the vehicle through its mass centres. Since these
lie above the ground plane they increase the vertical load at the outer wheels and
decrease them at the inner wheel. Aerodynamic forces also affect the vertical tyre
loads. They become significant for passenger car speeds above 80 km/h and are
dependent on vehicle type and body design. Since the forces acting on a vehicle are
above the ground plane they produce moments that cause pitch and roll. These in turn
produce load transfer and hence affect tyre loading.
Front of vehicle
Toe is the difference
in these dimensions
1-9
a) Static loading b) Effect of cornering force
Figure 1.6 Tyre contact patch
Longitudinal tyre forces
Longitudinal tyre forces are produced by braking and accelerating, rolling resistance
and aerodynamic forces. They are affected by both wheel-slip and vertical loading.
Lateral force and slip angle
As a result of the cornering force described above there is also shear force produced in
the plane of the tyre contact patch. This produces distortion of the contact patch
resulting in the direction of wheel motion differing from the wheel heading by an
angle call the slip angle. Figure 1.6b shows this for the case of an unsteered wheel.
This clearly this has implications for directional control and handling stability of a
vehicle. Aerodynamic forces and camber angle also affect lateral tyre forces.
Pneumatic trail and self-aligning moment
As a result of distortion of the tyre contact patch the resultant lateral tyre force F
y
acts
at a distance t
p
(termed the pneumatic trail), behind the vertical centre line through the
wheel centre (Figure 1.6b). This produces a self-aligning moment, M
z
, of magnitude
t
p
F
y
about an axis normal to the road surface.
1.2 Vehicle suspensions
In this section we examine the factors that influence the selection of suspension types,
discuss the kinematic requirements for suspensions and examine a range of
suspension configurations.
Having selected a suspension mechanism for a given vehicle a number of issues need
to be considered. These include how the wheel movement changes relative to both the
vehicle body and the road surface over the range of suspension travel and whether the
Distorted
contact patch
Lateral
force
Wheel
heading
Direction
of travel
Slip
angle
Symmetrical
contact patch
t
p
1-10
suspension members are able to cope with the variety of loads imposed on them. Of
particular concern are changes in camber and toe angles together with changes in track
(this is related to tyre scrub and hence tyre wear). The loading in some suspension
mechanisms produces a need for additional support members. In general these do not
affect the kinematic behaviour of the suspension mechanism.
Initially, we will assume that suspension mechanisms undergo rigid body motion, with
no compliance at the joints between members. However, most modern suspensions do
incorporate some compliance at certain joints to facilitate subtle wheel / body
movements such as compliance steer to enhance handling performance. Also we will
assume initially that the motion of the wheel relative to the body is 2-dimensional, i.e.
occurs in a plane perpendicular to the longitudinal axis of the vehicle. In reality,
because of additional design requirements, e.g. pitch control of the vehicle body, the
wheel motion needs to be 3-dimensional.
The complexity of the issues facing the chassis designer can be appreciated when one
also considers how various operational factors affect chassis performance. These are
discussed in [1].
1.2.1 Factors influencing suspension selection
Choice of suspension is primarily determined by the engine-drive combination, which
is in turn determined by the following:
- The functional requirements of the vehicle (use, performance capabilities and cargo
space).
- The need to place a high proportion of the vehicle mass over the driven axles to aid
traction.
The most common combinations are outlined as follows. The advantages and
disadvantages of these configurations are more fully discussed in [1].
Front mounted engine and rear wheel drive
In this case there is not the same restriction on engine length as there is with the front
engine, front wheel drive combination. For this reason it is adopted for more powerful
passenger cars and estate cars. Under full load most of the vehicle mass is on the
driven axle. However, when lightly loaded, e.g. in the 2-up condition, it leads to poor
traction on wet and wintry roads. This problem can be overcome with the use of
traction control.
With this configuration, the implications for the choice of a rear suspension are that it
has to accommodate the differential and drive shafts while still providing an
acceptable boot space.
Front mounted engine and front-wheel drive
In this case the engine and transmission form one unit which sits in front of, over, or
just behind the front axle. This leads to a very compact arrangement compared to the
previous case and this is no doubt responsible for its adoption on small to medium
1-11
sized saloons (up to 2-litre capacity). The main advantages include a significant load
on the driven and steered wheels for normal vehicle loads providing good traction and
road holding on wet and icy roads. With a fully laden vehicle this advantage tends to
be lost.
The implications for front suspension design are that provision must be made for drive
shafts, and steering. In general the packaging space for the suspension is limited by the
size of the engine-transmission unit. This engine-drive configuration favours
McPherson strut and double wishbone types of suspensions. A variety of rear
suspensions are possible for this drive configuration. The type of vehicle and the
degree of refinement required usually influence the choice.
Four wheel drive
In this case it is possible to have all the wheels driven continuously or one of the pair
of wheels is always connected to the engine while the other pair are selected manually
or automatically. The aim is of course to improve traction for all road conditions
especially wet and wintry conditions. All wheel drive is particularly advantageous for
off-road conditions, and to improve climbing ability regardless of loading conditions.
From the point of view of suspension selection, the requirements for the front
suspension is similar to that on a vehicle with front mounted engine and front wheel
drive. The rear suspension requirement is similar to that for a rear suspension on a
vehicle with a front mounted engine and rear wheel drive.
Suspensions have been broadly categorised in section 1.1.4 as dependent, independent
or semi-dependent. Solid axles have been categorised as dependent systems, but other
connections across the vehicle, such as anti-roll bars, are not classed as dependent,
because although they influence suspension forces, they do not provide a geometric
constraint.
Other factors
The mass of the components associated with the wheel hub, brake components
(outboard brakes), and parts of the other masses connected to the wheel hub, e.g.
suspension links, suspension spring and damper, create what is know as the unsprung
mass. This produces an additional resonant frequency when the vehicle system is
subjected to road induced vibration (a 2 degree of freedom system). The effect is to
increase the dynamic tyre load and cause degradation in ride. The aim should therefore
be to keep the unsprung mass to a minimum (reduce the system to 1 degree of
freedom the vehicle body).
Another factor to be considered is roll centre (and hence roll axis) location. This
affects such things as the amount of body roll in cornering, the amount of lateral load
transfer and the loads in the suspension links. It will be discussed in detail later. There
is a roll centre in the vertical transverse plane passing through each axle. Suspension
type and geometry determine its location. Some suspension types allow much greater
flexibility in the choice of roll centre location.
1-12
1.2.2 Kinematic requirements for suspensions
The main objective in suspension design is to isolate the vehicle body from road
surface undulations. Ideally, this requires that the wheels of the vehicle perfectly
follow the road surface undulations with no vertical movement of the vehicle body. In
analysing this scenario, it is possible to assume the vehicle body is stationary while we
consider the motion of the wheels.
The motion requirements of the wheels relative to the vehicle body can be satisfied in
a number of different ways. If each wheel can move independently of its partner on
the other side of the vehicle (i.e. an independent suspension, as defined in section
1.1.4), this can be achieved kinematically as shown in Figure 1.7. In the first example,
the wheel moves vertically relative to the vehicle body via a sliding contact. In the
second example the required vertical motion is accompanied by rotation about a pivot
attached to the body, while in the third example the wheel and hub are attached to one
of the links in a 4-bar mechanism. This provides the required vertical wheel
movement accompanied by a small rotation.
All of these examples have some practical disadvantages. These will be discussed
later when we have a look at some applications. In an independent suspension the
mechanism coupling the wheel to the body of the vehicle is said to have a single
degree of freedom, i.e. the motion of the wheel relative the body can be described by a
single coordinate.
(a) (b) (c)
Figure 1.7 Examples of possible independent suspension mechanisms
If it is necessary to couple a pair of wheels together on a single axle (as in a dependent
suspension), the mechanism required must have two degrees of freedom to allow
vertical movement of each wheel while at the same time having a rigid connection
between the two wheels. An example of how this can be achieved is shown in Figure
1.8.
1-13
Figure 1.8 A possible dependent suspension mechanism
1.2.3 Examples of dependent types of suspensions
These may be categorised into three sub-groups:
- Drive axles sometimes these are called live axles they carry the driven
wheels and differential.
- Non-driven axles usually called dead axles - in which a solid beam simply
connects the two undriven wheels which may be steerable.
- De Dion axles - which connect the two driven wheels but do not carry the
differential and drive shafts.
Solid axles are used rarely in passenger vehicle design. They are however almost
universally used for trucks and commercial vehicles and for many off-highway
vehicles because of their simplicity (relative to independent systems). They virtually
eliminate wheel camber (except for the small amount arising from different tyre
deflections on the inner and outer wheels) and wheel alignment is easy to maintain,
providing good tyre wear properties.
Many methods of mounting solid axles have been and still are used in vehicle design.
Only the main types and associated principles are covered here. Further descriptions
of suspensions are covered in [1], [4] and [5].
Hotchkiss drive
The most familiar form of mounting a solid drive axle is on a pair of semi-elliptic leaf
springs; this is known as the Hotchkiss arrangement (Figure 1.9). It is a remarkably
simple suspension and produces satisfactory location of the wheels with the minimum
number of components. Its simplicity derives from the leaf spring properties, i.e.
compliant in the vertical direction but relatively stiff laterally and longitudinally.
Developments in leaf springs over the years have overcome problems with inter-leaf
friction. Single and taper leaf designs now tend to be used. However, for the low
spring rates required for good ride performance, single leaf springs are generally
inadequate to resist lateral loads and driving and braking torques on modern vehicles.
In such cases, additional strengthening links such as Panhard rods are used.
The detailed design of the leaf spring itself and the associated shackle mounting
geometry (Figure 1.10) which accommodates changes in effective spring length is not
1-14
straightforward. Design guidelines are published in the SAE Spring Design Manual
[7].
Four-link arrangements
The requirements for the axle to have two degrees of freedom and also withstand all
types of loading can be satisfied by a number of four-link suspension mechanisms.
In the mechanism shown in Figure 1.11, two upper and two lower links - usually
trailing (but not always), react the driving / braking torques while providing the
required vertical and roll degrees of freedom. Lateral control may be provided by
angling the upper links (as shown in Figure 1.11), triangulating the upper links, or by
providing an additional Panhard rod. Any springing medium can be used, but coil or
air springs are the most common. The advantages of this suspension design over the
Hotchkiss arrangement are greater design flexibility in arranging the roll centre, the
anti-dive/squat geometry and controlling roll steer.
Figure 1.9 Hotchkiss suspension system
Figure 1.10 Leaf spring kinematics
Bump stop
Leaf spring
Shackle
C
Leaf spring with shackle
Equivalent linkage to
represent leaf spring
C
1-15
De Dion system
In this suspension design the drive is separated from the solid axle (Figure 1.12).
Hence, the differential is chassis mounted and the simple axle is typically located by
one of many four-link variations. Its main advantage is that the unsprung mass is
reduced dramatically compared with a live axle. However, it does not have significant
cost or performance advantages over an independent system and hence is generally not
used these days.
1.2.4 Examples of independent front suspensions
Almost all passenger cars, and increasing numbers of light trucks, use independent
suspensions at the front. The gradual shift towards independent suspensions has
resulted from their benefits in packaging, (i.e. with front wheel drive vehicles they are
better able to fit into the limited space available around the engine- transmission
system). They also permit greater suspension design flexibility and overcome
problems
Figure 1.11 Four link suspension
Figure 1.12 De Dion suspension system
1-16
of steering shimmy
7
vibrations often associated with beam axle designs. Olley, one of
the early influential automotive engineers, is credited with proposing some of the first
designs for independent suspensions. His paper [9] makes interesting reading even
today. Over the years a bewildering number of designs have been proposed and the
descriptions here are restricted to the generic types.
McPherson strut suspensions
Great simplicity is the main benefit of this design with the wheel controlled by a lower
arm in combination with a sliding strut (Figure 1.13). The configuration shown in the
diagram is well able to withstand the longitudinal and lateral loads that will arise for
various operating conditions. Advantages of the design are good packaging and
simplicity, whilst its disadvantages are a high installation height (which may conflict
with the bonnet line) and wheel loads that produce a moment which must be reacted
by the strut. This produces friction problems in the strut. Angling the axes of the
spring and strut can reduce this problem.
Figure 1.13 McPherson strut suspension
Wishbone suspensions
Upper and lower wishbones - or arms as they are called in the United States - are
combined to form a classic four-bar linkage mechanism when viewed from the front.
(Figures 1.7c and 1.14) The wishbones are nearly always unequal in length; the upper
arm is invariably shorter than the lower one to meet the space limitations in front-
engined vehicles.
7
Steering shimmy a coupled bounce and steer oscillation found on some early
vehicles
1-17
1.2.5 Examples of independent rear suspensions
Differences in rear suspension design are dictated by whether the axle is driven or
undriven. In the case of rear-wheel drive, one must consider the packaging space
(affect by both differential and a need to minimise intrusion into the boot space) and
also torque-steer effects. This later point is particularly relevant for larger, more
powerful, saloon cars.
Trailing arm
A simple single trailing arm (Figure 1.15) can be used with a torsion bar spring (e.g.
as in the Renault 5) or with a rubber or hydro-elastic one (e.g. the early design of the
Mini). Again, the wheel camber is the same as the body roll angle.
Swing axle
This is also a simple way of achieving an independent rear suspension (Figure 1.16) -
made famous on the VW Beetle design. The length of the swing axle controls the
camber behaviour and this has to be relatively short to accommodate the differential.
Camber changes (and consequently scrub) are significant during normal wheel travel.
Figure 1.14 Wishbone suspension
Figure 1.15 Rear trailing arm suspension
1-18
Also this design is particularly prone to the problem of jacking
8
(vertical translation of
the vehicle) and is little used nowadays.
Semi-trailing arm systems
This design is a cross between a pure trailing arm and a swing axle (Figure 1.17). The
additional flexibility in the design allows a compromise in the control of camber and
jacking. However care must be taken with the geometry to control the amount of steer
which results from the inclination of the pivot axis. More recent variants of this type
of design try to exploit small amounts of rear steering to improve overall vehicle
handling qualities (e.g. Vauxhall Omega).
Multilink designs
Multi-link suspensions are now a common feature on passenger cars from mid-range
models upward. They are usually based on three, four, or five links per wheel station
and are designed to give more complete control over wheel positioning [10]. This
comes at a price due to the added number of components, and the extra effort required
in setting them up. The Mercedes Benz design (Figure 1.18) is typical of such designs.
Because of the number of links (5 in this case), the suspension mechanism is over-
8
Jacking in certain suspensions, lateral tyre forces produce progressive track
changes and body lift leading to instability
Figure 1.16 Swing axle suspension
Figure 1.17 Semi-trailing arm suspension
1-19
constrained kinematically. Suspension movement is only possible, because of
compliant suspension bushes at the ends of some of the links.
Figure 1.18 Mercedes Benz five link suspension
1.2.6 Semi-independent rear suspensions
This arrangement has become common at the rear of small front-wheel drive vehicles
due to its relative simplicity and low cost. It is a cross between a fully independent
trailing arm arrangement and a dependent beam axle arrangement. The axle is
replaced by a crossbeam that is allowed to twist by an amount controlled by the
geometry of the suspension and the structural properties of all the members.
An example is shown in Figure 1.19. Such designs are known by various names, e.g.
trailing twist-axles and semi-dependent suspensions. The cross member is effectively
rigid in bending but has torsional compliance and this performs a function similar to
that of an anti- roll bar.
Figure 1.19 Trailing twist axle
1-20
The support bushes can be angled and designed to have different stiffnesses along and
perpendicular to their principal axes (Figure 1.20). This can produce a shear centre
behind the wheel axis to give a compliance steer in the correct direction consistent
with understeer.
Figure 1.20 Trailing twist axle with lateral force understeer
The shape of the open section cross-member can be varied to alter the position of its
shear centre (a feature of open sectioned members in torsion). This permits some
flexibility in the choice of roll centre location. This will be discussed later.
1.3 Suspension components
Suspension components, particularly springs and dampers, have a profound effect on
ride and handling performance. In addition to the constraints imposed by suspension
performance requirements, component designers have a range of other constraints to
consider. These include cost, packaging, durability and maintenance. Because of the
hostile environment in which suspension components operate and the high fluctuating
loads (and hence stresses) involved, fatigue life is another of the designers prime
concerns.
1.3.1 Springs - types and characteristics
Suspension systems require a variety of compliances to ensure good ride, handling and
NVH performance. The need for compliance between the unsprung and sprung masses
to provide good vibration isolation has long been recognised. In essence, a suspension
spring fitted between the wheel and body of a vehicle allows the wheel to move up
and down with the road surface undulations without causing similar movements of the
body. For good isolation of the body (and hence good ride), the springs should be as
soft as possible consistent with providing uniform tyre loading to ensure satisfactory
handling performance.
The relatively soft springing required for ride requirements is normally inadequate for
resisting body roll in cornering, therefore it is usual for a suspension system to also
include additional roll stiffening in the form of anti-roll bars. Furthermore, there is the
Angled bush
Angled bush
Lateral tyre
force
Lateral tyre
force Shear centre
of bushes
Direction
of turn
Axle
rotation
1-21
possibility of the suspension hitting its stops at the limits of its travel as a result of
abnormal ground inputs (e.g. as a result of striking a pothole). It is then necessary to
ensure that the minimum of shock loading is transmitted to the sprung mass. This
requires the use of additional springs in the form of bump stops to decelerate the
suspension at its limits of travel.
Finally, there is also a requirement to prevent the transmission of high frequency
vibration (>20 Hz) from the road surface, via the suspension to connection points on
the chassis. This is achieved by using rubber bush connections between suspension
members.
The following compliant elements are thus required in suspension systems:
suspension springs, anti-roll bars, bump-stops and rubber bushes. In this section our
attention is concentrated on suspension springs and anti-roll bars. Further details of
suspension components can be found in [3] to [5], [11] and [12].
Suspension springs
The main types of suspension spring are:
- Steel springs (leaf springs, coil springs and torsion bars)
- Hydropneumatic springs
a) Leaf springs
Sometimes called semi-elliptic springs, these have been used since the earliest
developments in motor vehicle technology. They rely on beam bending principles to
provide their compliance and are a simple and robust form of suspension spring still
widely used in heavy commercial applications, e.g. lorries and vans. In some
suspensions (e.g. the Hotchkiss type) they are used to provide both vertical
compliance and lateral constraint for the wheel travel. Size and weight are among
their disadvantages.
Leaf springs can be of single or multi-leaf construction (Figure 1.21). In the latter case
(Figure 1.21 (b)) interleaf friction (which can affect their performance) can be reduced
with the use of interleaf plastic inserts. Rebound clips are used to bind the leaves
together for rebound motion. The swinging shackle accommodates the change in
length of the spring produced by bump loading. The main leaf of the spring is formed
at each end into an eye shape and attached to the sprung mass via rubber bushes.
Suspension travel is limited by a rubber bump stop attached to the central rebound
clip. Structurally, leaf springs are designed to produce constant stress along their
length when loaded.
Spring loading can be determined by considering the forces (see section 1.5) acting on
the spring and shackle as a result of wheel loading (Figure 1.22). The spring is a three
force member with F
A
, F
W
and F
C
at A, B and C respectively. The wheel load F
W
, is
vertical and the direction of F
C
is parallel to the shackle (a two-force member). The
direction of F
A
must pass through the intersection of the forces F
W
and F
C
(point P) for
the link to be in equilibrium. Knowing the magnitude of the wheel load enables the
other two forces to be determined. The number, length, width and thickness of the
1-22
leaves determine the stiffness (rate) of the spring. See [7] and [13] for stiffness and
stress formulae. Angling of the shackle link can be used to give a variable rate [7].
When the angle u < 90 (Figure 1.22), the spring rate will increase (i.e. have a rising
rate) with bump loading.
(a) Single leaf type (b) Multi-leaf type
Figure 1.21 Examples of leaf spring designs
(a) Wheel load (b) Force on the spring member
Figure 1.22 Leaf spring loading
b) Coil springs
This type of spring provides a light and compact form of compliance that is an
important feature in terms of weight and packaging constraints. It requires little
maintenance and provides the opportunity for co-axial mounting with a damper. Its
disadvantages are that because of low levels of structural damping, there is a
possibility of surging (resonance along the length of coils) and the spring as a whole
does not provide any lateral support for guiding the wheel motion.
Most suspension coil springs are of the open coil variety. This means that the coil
cross-sections are subjected to a combination of torsion, bending and shear loads.
Spring rate is related to the wire and coil diameters, the number of coils and the shear
modulus of the spring material. Cylindrical springs with a uniform pitch produce a
linear rate. Variable rate springs are produced either by varying the coil diameter
and/or pitch of the coils along its length. In the case of variable pitch springs, the coils
F
C
F
W
F
A
F
W
P
u
A
C
B
Bump stop
Shackle
Single tapered
leaf spring
Bump stop
Shackle
Multi-leaf
spring
1-23
are designed to bottom-out as the spring is loaded, thereby increasing stiffness with
load.
Coil spring design is well covered in the literature. In addition to Wahls classic text
[14], the reader is also recommended to consult [7] and chapters in, [14] and [15].
c) Torsion bars
These are a very simple types of springs and consequently very cheap to manufacture.
They are both wear and maintenance free. Despite their simplicity they cannot easily
be adopted for some of the more popular forms of suspension.
The principle of operation (Figure 1.23) is to convert the applied load F
W
into a torque
F
W
x R producing twist in the bar. A circular cross-section bar gives the lowest spring
weight for a given stiffness. In this case simple torsion of shafts theory can be used to
determine the stiffness of the spring and the stresses in it. As the lever-arm rotates
under load, the moment-arm changes, requiring twist angle corrections (for large
rotations) to be made in design calculations [14] and [7]. In general, stiffness is related
to diameter and length of the torsion bar and the torsion modulus of the material. A
bending moment F
W
x L is induced in the twist section of the member and supports
should be positioned to minimise this.
Figure 1.23 Principle of operation of a torsion bar spring
d) Hydropneumatic springs
In this case the spring is produced by a constant mass of gas (typically nitrogen) in a
variable volume enclosure. The principle of operation of a basic diaphragm
accumulator spring is shown in Figure 1.24. As the wheel deflects in bump, the piston
moves upwards transmitting the motion to the fluid and compressing the gas via the
flexible diaphragm. The gas pressure increases as its volume decreases to produce a
hardening spring characteristic.
The principle was exploited in the Moulton-Dunlop hydrogas suspension where
damping was incorporated in the hydropneumatic units. Front and rear units were
connected to give pitch control. See [11] for a detailed description of the system.
A further development of the principle has been the system developed by Citroen.
This incorporates a hydraulic pump to supply pressurised fluid to four hydropneumatic
struts (one at each wheel-station). Height correction of the vehicle body is achieved
L
Twist section
Fixing plane
R
F
W
1-24
with regulator valves that are adjusted by roll-bar movement or manual adjustment by
the driver. The system and its operation can be found in [11].
In general, hydropneumatic systems are complex (and expensive) and maintenance
can also be a problem in the long term. Their cost can, however, be offset by good
performance. Patents cover the two systems discussed in the preceding paragraphs, but
there is still scope for development of alternative hydropneumatic systems. Some of
these are incorporated into controllable suspensions.
Figure 1.24 Principles of a hydropneumatic suspension spring
Anti-roll bars (roll stabilisers)
These are used to reduce body roll and have an influence on a vehicles cornering
characteristics (in terms of understeer and oversteer). Figure 1.25a shows how a
typical roll bar is connected to a pair of wheels. The ends of the U-shaped bar are
connected to the wheel supports and the central length of the bar is attached to the
body of the vehicle. Attachment points need to be selected to ensure that bar is
subjected to torsional loading without bending.
If one of the wheels is lifted relative to the other, half the total anti-roll stiffness acts
downwards on the wheel and the reaction on the vehicle body tends to resist body roll.
If both wheels are lifted by the same amount, the bar is not twisted and there is no
transfer of load to the vehicle body. If the displacements of the wheels are mutually
opposed (one wheel up and the other down by the same amount), the full effect of the
anti-roll stiffness is produced.
(a) Anti-roll bar layout (b) Roll bar contribution to total roll moment
Figure 1.25 Anti-roll bar geometry and the effect on roll stiffness
Gas
Spherical container
Liquid Diaphragm
Piston
Wheel load
Roll
Moment
Bump stop
effect
Anti-roll bar
Roll bar
Roll angle
Chassis fixing
points
Total roll
moment
Suspension
1-25
Roll stiffness is equal to the rate of change of roll moment with roll angle. Figure
1.25(b) shows the contributions to the total roll moment made by suspension stiffness
and anti-roll bar stiffness.
1.3.2 Dampers - types and characteristics [8]
Frequently called shock absorbers, dampers are the main energy dissipaters in a
vehicle suspension. They are required to dampen vibration after a wheel strikes a pot-
hole and to provide a good compromise between low sprung mass acceleration
(related to ride) and adequate control of the unsprung mass to provide good road
holding.
Suspension dampers are telescopic devices containing hydraulic fluid. They are
connected between the sprung and unsprung masses and produce a damping force that
is related to the relative velocity across their ends. The features of the two most
common types of damper are shown in Figure 1.26. Figure 1.26(a) shows a dual tube
damper in which the inner tube is the working cylinder while the outer cylinder is used
as a fluid reservoir. The latter is necessary to store the surplus fluid that results from a
difference in volumes on either side of the piston. This is a result of the variable rod
volume in the inner tube. In the monotube damper (Figure 1.26(b)) the surplus fluid is
accommodated by a gas-pressurised free piston. An alternative form of monotube
damper (not shown in Figure 1.26) uses a gas/liquid mixture as the working fluid to
absorb the volume differences. Comparing the two types of damper shown in Figure
1.26, the dual tube design offers better protection against stones thrown up by the
wheels and is also a shorter unit making it easier to package. On the other hand, the
monotube strut dissipates heat more readily.
(a) Dual tube damper (b) Free-piston monotube damper
Figure 1.26 Damper types
In dealing with road surface undulations in the bump direction (the damper in
compression) relatively low levels of damping are required when compared with the
rebound motion (the damper extended). This is because the damping force produced
Rod seal
Gas
Piston rod
Piston and
valves
Floating
piston
Valves
Gas
1-26
in bump tends to aid the acceleration of the sprung mass, while in rebound an
increased level of damping is required to dissipate the energy stored in the suspension
spring. These requirements lead to damper characteristics that are asymmetrical when
plotted on force-velocity axes. The damping rate (or coefficient) is the slope of the
characteristic. Ratios of 3:1 for rebound to bump are quite common (Figure 1.27).
However, one manufacturer has discarded this reasoning and used linear dampers on
one of its vehicles with some success [16].
Figure 1.27 Non-linear damper characteristics
The characteristics called for in damper designs are achieved by a combination of
orifice flow and flows through spring-loaded one-way valves. These provide a lot of
scope for shaping and fine-tuning of damper characteristics
9
. Figure 1.28 illustrates
how the force-velocity characteristics can be shaped by a combined use of damper
valves. At low relative velocities damping is by orifice control until the fluid pressure
is sufficient to open the pre-loaded flow-control valves. Hence the shape of the
combined characteristic.
Figure 1.28 Shaping of damper characteristics
A driver operated adjustment mechanism can be used to obtain several damping
characteristics from one unit. Typical curves for a three position adjustable damper are
shown in Figure 1.29. A continuous electronically controlled adjustment forms the
basis of one type of controllable suspension designed to improve both ride and
handling.
9
For a more detailed account of damper operation and characteristics see [8]
Relative velocity
Valve control
Orifice control
Damping
force
Damping
force
Relative velocity
(bounce)
1-27
Figure 1.29 Different operating modes for adjustable dampers
1.4 The kinematics of suspensions
Kinematics is the study of the geometry of motion, without reference to the forces
involved. In the case of vehicle suspensions, which are simply examples of the many
linkages and mechanisms encountered in mechanical engineering, study of their
kinematic properties involve some important assumptions in relation to their practical
operation. For example, elastic deformations are excluded from the analysis (since
forces are excluded). This means that bush compliances cannot be included. Hence the
suspension must be treated as if all the link connections are perfect pivots, sliders or
ball joints.
A kinematic analysis of a proposed suspension is invariably the first analysis
performed by the designer in order to:
- Check that the mechanism actually works over the full bump/rebound range and
does not interfere with the body
- Check the way in which the wheel geometry is controlled over the suspension
working range
- Calculate the effective ratios involved between spring and damper travel relative
to wheel travel.
1.4.1 Kinematic analysis of suspension mechanisms
Suspensions are really three-dimensional mechanisms so that analyses of them should
be three-dimensional. However, the largest displacement components occur in a
transverse plane perpendicular to the longitudinal axis of the vehicle. It follows that a
lot of useful information can be obtained from a two-dimensional analysis.
There are three ways in which suspension kinematics can be analysed:
- Analytically using vector-based methods
- Using a computer package with a mechanisms analysis capability
- Graphically (only really relevant for a two-dimensional analysis).
It is the third approach that we are going to use in the following example. Readers
unfamiliar with graphical kinematic analysis should consult an undergraduate text
such as [17].
Damping
force
Relative velocity
Sport
Standard
Comfort
1-28
Example 1.1
A scaled drawing of a double wishbone suspension is shown diagrammatically in
Figure E 1.1
The damper is mounted between E and F and E can be assumed to be on the line AD.
The centre of the tyre contact patch is at W and WV is the centre-line of the wheel.
Use a velocity diagram to determine:
a) The scrub to bump derivative at the tyre-road interface (ratio of wheel horizontal
to vertical velocities)
b) The damper rate to bump rate (ratio of relative velocity across the damper to the
vertical velocity of the wheel)
c) The suspension ratio (the inverse of the damper rate to bump rate)
Figure E1.1
Solution
Begin by drawing the mechanism to scale as shown in Figure E1.2 and assume the
chassis fixing points to be stationary. Apply a unit angular velocity of e =1 rad/s to
link AD (the absolute magnitude is not important as the things to be determined are all
ratios). The magnitude of the velocity of D is v
D
= 360 x 1 = 360 mm/s and the vector
v
D
is perpendicular to AD as shown.
The velocity of C is given by: v
C
= v
D
+ v
CD
(a vector addition, where a symbol
underlined denotes a vector). v
C
is perpendicular to BC its magnitude is unknown at
this stage. Because CD can be considered as a rigid link, v
CD
is perpendicular to CD
its magnitude is unknown at this stage. The directions of these vectors are also
unknown (see dotted lines in Figure E1.2). Also draw a line from D to W it will be
assumed this imaginary link is rigid and form part of a link CDW.
A
E
D
B
C
F
V
W
Scale 1 : 6
Dimensions in mm
AE = ED 180
CD = 315
WD = 159
1-29
Construction of the velocity diagram (Figure E1.3) begins by first locating the pole O
v
Vectors drawn from here are absolute velocities.
1. Draw v
D
to some scale and locate D at its end.
2. Draw lines representing the directions of v
C
and v
CD
. Note that v
C
is an absolute
velocity and hence is drawn from O
v
3. Locate C at their intersection and add arrows to satisfy the vector equation above.
The magnitudes of v
C
and v
CD
can now be determined by scaling the diagram.
4. There should be a velocity image CDW in the velocity diagram of similar shape to
CDW in Figure E1.1. Determine W by proportion from FigureE1.1 and hence add
the vector v
WD
as shown in Figure E1.3
v
D
e
A
E
D
B
C
F
V
W
Figure E1.2
Velocity diagram
(to some scale)
Figure E1.3
v
WD
v
W
vert
O
v
v
D
D
v
C
C
v
CD
v
W
hor
v
E
v
E
r
v
E
t
W
E
1-30
5. The (absolute) vertical and horizontal velocity components of W, v
W
hor
and v
W
vert
can be added to the diagram as shown in Figure E1.3
6. The (absolute) velocity of E, v
E
, is half that of D and can now be added to the
diagram. There are components parallel and perpendicular to EF which are related
respectively to relative velocity across the damper and its angular velocity. Add
the component vectors v
E
r
and v
E
t
.
7. Relevant values can now be scaled from the diagram. These are: v
W
vert
= 336
mm/s; v
W
hor
= 114 mm/s and v
E
r
= 168 mm/s
Answers
a) Scrub to bump derivative:
v
W
hor
/ v
W
vert
= 114 / 366 = 0.393
b) Damper rate : bump rate
v
E
r
/ v
W
vert
= 168 / 366 = 0.459
c) Suspension ratio
v
W
vert
/ v
E
r
= 366 / 168 = 2.18
Bump to scrub ratio is related track change and tyre wear. Damper rate to bump rate
and suspension ratio are both related to modelling for ride analysis and the
determination of component parameters.
The facility to perform kinematic analysis of three-dimensional mechanisms is
available on many CAD systems. Alternatively, independent packages have been
written specifically to analyse three-dimensional suspension kinematics, for example,
the one embodied in VDAS, [18]. The industry standard is ADAMS [19].
Some typical output from the VDAS package for a three-dimensional analysis of
double wishbone suspension is shown in Figure 1.30. Note that the results have been
calculated over the full range of suspension travel and are for a particular steering
angle. The package also permits the evaluation of driveshaft data such as plunge at
CV joints and CV joint angles.
1.4.2 Roll centres and roll axis [20]
Roll centre and roll axis concepts are important aids for studying vehicle handling and
are also used in the calculation of lateral load transfer for cornering operations.
There are two definitions of roll centre. One is based on forces and the other on
kinematics. The first of these (the SAE definition) states that: the roll centre is a point
in the transverse plane through any pair of wheels at which a transverse force may be
applied to the sprung mass without causing it to roll. The second is a kinematic
1-31
definition that states that: the roll centre is the point about which the body can roll
without producing any lateral movement at either of the wheel contact areas.
Figure 1.30 Examples of results from a 3-D kinematic analysis
In general each roll centre lies on the vertical line produced by the intersection of the
longitudinal centre plane of the vehicle and the vertical transverse plane through a pair
of wheel centres. The roll centre heights in the front and rear wheel planes tend to be
different as shown in Figure 1.31. The line joining the centres is called the roll axis,
with the implication that a transverse force applied to the sprung mass at any point on
this axis will not cause body roll.
Figure 1.31 Roll centres and roll axis
1.4.3 Roll centre determination
For a given front or rear suspension, the roll centre can be determined from the
kinematic definition by using the Aronhold-Kennedy theorem of three centres. This
states: when three bodies move relative to one another they have three instantaneous
centres all of which lie on the same straight line.
To illustrate the determination of roll centre by this method consider the double
wishbone suspension shown in Figure 1.32. Consider the three bodies capable of
relative motion as being the sprung mass, the left-hand wheel and the ground. The
instantaneous centre of the wheel relative to the sprung mass I
wb,
lies at the
intersection of the upper and lower wishbones, while that of the wheel relative to the
ground lies at I
wg
. The instantaneous centre of the sprung mass relative to the ground
(the roll centre) I
bg,
must (according to the above theorem), lie in the centre plane of
the vehicle and on the line joining I
wb
and I
wg
, as shown in the diagram.
Rear RC
Front RC
Roll axis
1-32
Figure 1.32 Roll centre determination for a double wishbone suspension
I
wg
always lies in the position shown. Changing the inclination of the upper and lower
wishbones can vary I
wb
, thereby altering the location of I
bg
and ultimately the load
transfer between inner and outer wheels during cornering. The importance of roll
centre location is discussed later. Figures 1.33 to 1.36 illustrate the locations of roll
centres for a variety of independent suspensions. In the case of the McPherson strut
suspension (Figure 1.33) the upper line defining I
wb
is perpendicular to the strut axis.
Figure 1.33 Roll centre location for a McPherson strut
For the trailing arm suspension (Figure 1.35), the trailing arm pivots about a
transverse axis (forward of the wheel centre). In the front view (Figure 1.35(c)), the
wheel is constrained to move in a vertical plane (with no transverse movement) and
hence I
wb
lies at infinity along the pivot axis (to the right). The roll centre therefore
lies in the ground plane on the centre-line of the vehicle.
Figure 1.34 Roll centre location for a swing axle suspension
I
wb
I
bg
(RC)
I
wg
I
wb
I
bg
(RC)
I
wg
I
wb
I
bg
(RC)
I
wg
1-33
(b) Side view (c) Front view
Figure 1.35 Roll centre location for a trailing-arm suspension
For the semi-trailing arm suspension (Figure 1.36) the pivot axis is inclined and
intersects the vertical lateral plane through the wheel centre at I
wb
a distance L from
the centre plane of the wheel. The roll centre I
bg
lies on the line connecting I
wb
with
the instantaneous centre of the wheel relative to the ground I
wg
.
Figure 1.36 Roll centre location for semi-trailing arm suspension
Figure 1.37 shows the roll centre determination for a four-link rigid axle suspension.
In this case the wheels and axle can be assumed to move as a rigid body. The upper
and lower control arms produce instant centres at A and B respectively (projected to
I
w
I
w
I
bg
I
bg
I
w
(a) Plan view
I
w
(a) Plan view
Vehicle
c.l.
L
L
(c) Front (b) Side view
1-34
the vehicle centreline). Connecting these together produces a roll axis for the
suspension. The intersection of this axis with the transverse wheel plane defines the
roll centre.
Figure 1.37 Roll centre for a four link rigid axle suspension
Our final example illustrating roll centre determination is the Hotchkiss rear
suspension shown in Figure 1.38. The analysis in this case is somewhat different to
the previous examples. Lateral forces are transmitted to the sprung mass at A and B.
The roll centre height is at the intersection of the line joining these points and the
vertical transverse plane through the wheel centres. The roll centre is of course at this
height in the centre plane of the vehicle.
Figure 1.38 Roll centre location for a Hotchkiss suspension
1.4.4 Roll centre migration
When a suspension deflects, the roll centre locations move. For modest suspension
deflections this migration of the roll centre can be quite significant, (as shown in
Figure 1.39), arising in this case from the vertical deflection of the vehicle body.
Large roll centre migrations can have undesirable effects on vehicle handling. Hence
attempts are made at the design stage to limit them for the full range of suspension
(a) Plan view
(b) Side view
A
RC
B
B
A
Roll centre
height
RC
1-35
travel. Some suspension types are more amenable than others for controlling roll
centre migration.
Figure 1.39 Effect of suspension deflection on roll centre migration
1.4.5 Roll centre effects
A high roll centre produces low lateral load transfer (good anti-roll) but produces high
levels of half-track changes. The high roll centre is typically closer to the centre of
gravity and hence there is a reduced lever arm and lower roll moment. Lateral forces
generated under cornering therefore produce less roll. However, the suspension
geometry to give a high roll centre typically means that the wheel scrub is significant
over the suspension travel, hence the larger half-track changes.
A low roll centre produces small half-track changes but poor anti-roll properties.
The migration of the roll centre needs to be limited to minimise the steering balance
variation under different driving conditions.
Steering Response
The balance between steady-state understeer / oversteer is largely dictated by the front
to rear lateral load transfer ratio which is dependent on roll centre heights. This can be
modified by changes in roll stiffness usually with anti-roll bars. Migration of roll
centres can create unwanted variations in steering balance.
Straight Line Stability
Increasing front to rear lateral load transfer ratio produces an understeer tendency.
Unsymmetrical road undulations induce half-track changes producing lateral force
variations at the tyre contact patches with resultant steer effects. A positive half track
change produces a force towards the vehicle centreline and this effect, when combined
with bump steer, will affect vehicle stability.
Braking Stability
Roll centre migration due to induced pitch changes can cause plunging front and
rising rear roll centres resulting in an undesirable oversteer tendency
I
w
b
I
wb
I
bg
(RC)
I
bg
I
wg
1-36
Lift-Off / Traction Steer on Bends
Roll centre migration due to induced pitch changes can cause variations in steering
balance (see Steer Response above).
Ride
Half-track changes produce resisting forces having a vertical component. In some
cases this may be large enough to affect ride quality. On asymmetric undulations, the
lateral forces due to half-track changes may excite roll-rock. Vehicle roll is
fundamentally linked to roll centre height, but can be modified by the use of anti-roll
bars.
Jacking
Vehicle jacking is affected by roll centre height.
1-37
1.5 Suspension Force Analysis
A thorough analysis of all the forces in a suspension system is a complex problem
requiring the use of computer packages. FE packages can be used for static analyses
and stress distribution within members while multi-body system (MBS) packages are
required for the analysis of three-dimensional dynamic loading. Analysing static
forces is straightforward, but superimposed on these are dynamic conditions due to
vertical and longitudinal road irregularities, braking, accelerating, cornering etc. Many
of these dynamic loads are of a transient nature and are very difficult to model. For
this reason they tend to be modelled by dynamic load factors.
In this section we will discuss how dynamic loads can be quantified and analyse
graphically some typical loading situations. The relationship between the suspension
spring stiffness (spring rate) and the equivalent stiffness in the vertical plane of the
wheel (wheel rate) are also examined. This relationship is important in terms of
determining suspension forces and the natural frequencies of vibration of the vehicle.
A generalised schematic of wheel load against wheel vertical displacement is given
below. The figure identifies the critical points where the response changes, e.g. due to
engagement of a bump stop. Whilst the main suspension spring and its associated
spring rate may be linear (or intentionally designed to be non-linear) the wheel rate is
almost always non-linear owing to the suspension geometry which defines how the
spring rate translates to give the wheel rate.
A unladen condition
B fully loaded condition
C bump stop engagement
D - bump stop fully engaged
E bump stop engagement
F zero vertical wheel force
G bump stop fully engaged
1-38
1.5.1 Suspension loading
a) Longitudinal loads due to braking and accelerating
Vertical, longitudinal and lateral tyre loads have been discussed briefly in section
1.1.6. It is useful to examine in a little more detail the generation of longitudinal loads
arising from control actions such as braking and accelerating.
For these cases the loads are dependent on how braking and drive-torques are reacted.
They in turn depend upon whether the vehicle has inboard or outboard brakes and
whether the drive is to a live axle or an independent suspension.
Braking - outboard brakes
In this case the braking torque is reacted at the hub assembly. Consider the free-body
diagram of the wheel in Figure 1.40a. The corresponding force and moment on the
hub carrier are shown in Figure 1.40b. These provide the suspension loading. An
equivalent suspension loading is shown in Figure 1.40c.
Figure 1.40 Loads due to braking with outboard brakes
Motion
Moment exerted by
brake pads F
x
r
F
x
Tyre force
a) Wheel forces and moment
Force exerted
by
suspension
on hub, F
x
Moment exerted by
the wheel on the
brake pads, F
x
r
Force exerted by
hub on
suspension, F
x
b) Hub force and moment
F
x
c) Equivalent force exerted on suspension
r
1-39
Braking - inboard brakes
In this case the braking torque is reacted inboard at the vehicle body. The free-body
diagram of the wheel is the same as in Figure 1.40a. However, in this case there is no
moment exerted on the hub carrier. The suspension load is simply equal to the
longitudinal tyre force acting through the centre of the hub, Figure 1.41a. An
equivalent (resultant) force system is shown in Figure 1.41b.
Figure 1.41 Loads due to braking with inboard brakes
Drive to a live axle
In this case the drive torque is reacted outboard at the hubs. The analysis is similar to
the case of outboard brakes. The equivalent force is shown in Figure 1.42
Figure 1.42 Suspension load resulting from drive to a live axle
Drive to independent suspension
In this case the drive torque is reacted inboard. The analysis is similar to the case of
inboard brakes. The equivalent force system acting on the suspension is shown in
Figure 1.43.
Figure 1.43 Suspension load resulting from drive to an independent suspension of
turn (for speeds below critical).
Force exerted by the
hub on suspension, F
x
a) Force exerted on suspension
Compensating
moment, F
x
r
F
x Motion
F
x
Compensating
moment, F
x
r
b) Equivalent force system on the suspension
F
x
1-40
b) Modelling transient loads
In order to safeguard the robustness of a suspension it is particularly important that the
effects of worst-case loading scenarios are analysed. These tend to arise from the
transient (shock) loads due to wheels striking potholes and kerbs and panic braking.
Because every set of events is likely to be different, designers tend to cover worst-case
scenarios by using load factors based on the static wheel load. Different manufacturers
tend to adopt their own set of design factors. Alternatively, the manufacturers work to
a set of worst-case accelerations along the different vehicle axes; a typical example is
shown in Table 1.1. The force loadings can then be obtained using an appropriate
mass for the wheel and unsprung mass.
1.5.2 Wheel-rate required for constant natural frequency
Wheel-rate is the vertical stiffness of the suspension in a wheel plane. One of the
many problems facing the suspension designer is that as the body mass of a vehicle
changes from the unladen to the laden condition, the natural frequency of the mass can
change. This is undesirable and happens when the wheel rate k, is constant over the
range of suspension travel. The problem can be overcome by arranging a wheel-rate
that increases with suspension travel. The following analysis shows that it is possible
to determine how the wheel-rate needs to vary in order to provide a constant natural
frequency as the payload changes.
Load case Load factor
Longitudinal Transverse Vertical
Front / rear
pothole bump
3g at the wheel
affected
0 4g at the wheel
affected, 1g at
other wheels
Bump during
cornering
0 0 3.5g at wheel
affected, 1g at
other wheels
Lateral kerb
strike
0 4g at front
and rear
wheels on
side affected
1g at all wheels
Panic braking 2g at front
wheels, 0.4g at
rear wheels
0 2g at front
wheels, 0.8g at
rear wheels
Table 1.1 Typical worst case accelerations for dynamic suspension loads.
.
A vehicle body mass supported on its suspension can be modelled most simply as a
mass on a spring for which the natural frequency is given by:
n
k
m
e
=
|
\

|
.
|
1
2
where k = wheel rate and m = effective sprung mass.
(1-1)
1-41
Equation (1-1) may first of all be written in terms of an effective static deflection, o
s
,
which is defined by
s
mg
k
o =
Hence, substituting in Equation (1-1) gives
n
s
g
e
o
= ( )
1
2
Note the units. When g is in m/s
2
and o is in m, the units of e
n
are rad/s.
A typical load -deflection curve for a rising-rate spring is shown in Figure 1.44. The
effective rate is the local slope of this curve at a particular point. To maintain a
constant value of natural frequency requires a particular relationship between the rate
and the load i.e.
Load
Rate
constant =
W
dW dx
s
= =
o
constant
Equation (1-5) now needs to be re-arranged and integrated to obtain an expression for
W as a function of x. Re-arranging gives,
dW
W
dx
s
=
o
Hence by integration,
e
s
W
x
c log = +
o
(1-2)
(1-3)
(1-4)
(1-5)
Figure 1.44 Rising rate spring characteristics
(1-6)
(1-7)
1-42
The integration constant, c, is given by substituting values at the static load condition
as indicated in Figure 1.48.
W
Ws s x x
=
=
when
c
W
x
e s
s
s
= log
o
Hence equation (1-7) becomes
e
s
s
s
W
W
x
x
log
( ) |
\

|
.
|
=

o
1.5.3 The relationship between spring-rate and wheel-rate
The rising-rate stiffness of the suspension in the wheel-plane will dictate the required
characteristic for the suspension spring. The relationship between wheel-rate and
spring rate will depend on the kinematics of the suspension. In general, the
displacement relationships between wheel travel and deflection across the suspension
spring will be nonlinear. Some examples will illustrate how the relationship can be
obtained.
A simple example
This example is based on a swing axle suspension, a design that is not commonly used
nowadays because it is prone to the suspension-jacking problem mentioned earlier.
However, it serves as a useful example to introduce spring/wheel-rate calculations.
For the case shown in Figure 1.45, the wheel-rate may be calculated as follows.
Assume the suspension spring has stiffness k
s
.
(1-8)
(1-9)
(1-10)
1-43
Figure 1.45 Swing axle example
A small displacement, ow, at the wheel results in a displacement at the spring of,
o
o
s
c w
d
=
Hence the spring force is
)
d
w c
(
k F s s
o
=
Taking moments about the pivot at A (Figure 1.49b) gives,
w s F
d
F
c =
Then expressing the wheel force in terms of an effective wheel rate, k
w
, gives,
w
k F w w
o =
Substituting in equation (1-12) then gives
w s k k
c
d
= ( )
2
In this case the rates are related by the lever arm ratio squared. This is strictly valid
only for small displacements around the mean position shown. For larger
displacements the relationship is non-linear
(1-11)
(1-12)
(1-13)
(1-14)
(1-15)
Spring
c
d
o
s
o
w
A
c
d
F
A
F
S
F
W
(a)
(b)
1-44
A general example suspension with a coil spring
A more general relationship between spring rate and wheel rate can be derived using
Figure 1.46. The suspension ratio is defined as
R
S
F
=
The spring stiffness is
s k
dS
dx
d
dx
RF = = ( )
s k
R
dF
dX
dX
dx
F
dR
dX
dX
dx
= +
Also by virtual work,
FdX Sdx =
This leads to an alternative definition of the suspension ratio,
R
S
F
dX
dx
= =
Noting that the wheel rate is
w k
dF
dX
=
Equation (1-18) may be rewritten as
s w k k R
S
dR
dX
= +
2
( )
1.5.4 Analysis of forces in suspension members
Despite all the complexities associated with the practical situations, it is nevertheless
possible to carry out some simple design calculations. These can be carried out
(1-16)
(1-17)
(1-18)
Figure 1.46 Calculation of suspension ratio
(1-19)
(1-20)
(1-21)
(1-22)
X
1-45
graphically for a given load application and enables the loads in the suspension
members and at body mounting points to be determined. Inevitably these calculations
involve some simplifying assumptions. Some examples will illustrate the procedure.
a) Vertical loading
Double wishbone suspension example
A diagram of a double wishbone suspension in which the spring acts on the lower arm
is shown in Figure 1.47. The following assumptions are made.
- Loads due to the masses themselves are ignored on the basis that they will be
insignificant compared with loads arising from the static wheel load, F
w
.
- All joints are treated as simple pin joints so that effects of friction, compliance in
bushes etc. are ignored.
The problem can be treated in two dimensions.
The static wheel load, F
W
, would typically be calculated from the vehicle data. The
problem then is to calculate the forces transmitted to the body mounting points.
Clearly this is a statics problem since all the members are in equilibrium and the
simplest approach is a graphical one. The free body diagrams and associated force
vector diagrams are shown in Figure 1.47.
Note the following features in obtaining these results.
- The force in the upper arm DA must be along the link since it is pin jointed at both
ends.
- Consider the wheel plus kingpin assembly first and construct the FBD. The
directions of F
W
and F
D
are known so the direction of F
C
can be found (it must
pass through G for equilibrium)
- Next, the FBD for the lower arm can be drawn.
A full numerical solution to this problem requires that (i) the position diagram of the
linkage is drawn to scale to define all the geometry and (ii) the force vector diagrams
are either drawn to scale or the forces calculated using the geometry defined by the
force vector diagrams.
1-46
In order to analyse the suspension forces over the full range of suspension travel, the
procedure can be repeated at suitable increments between full bump and rebound
positions. To do this some data about either the spring rate or the wheel rate must be
known or assumed. For example, if the load-deflection curve for the spring is known,
then the load-deflection curve for the wheel can be plotted and the effective wheel rate
calculated (Figure 1.48). Note that because of geometrical effects, the relationship
between wheel deflection and spring deflection is not linear and so even a linear
spring will result in a non-linear wheel rate. An alternative calculation is to assume a
particular wheel rate and hence calculate the spring load-deflection curve (Figure
1.49). This defines the spring characteristics necessary to achieve a desired wheel-
rate, which may be linear or non-linear.
Force vector diagrams
Free Body Diagram
for wheel/kingpin assembly
Free Body Diagram
for lower arm
Figure 1.47 Force analysis of a double wishbone suspension
1-47
McPherson strut example
A diagram of a McPherson strut suspension is shown in Figure 1.50. The same
assumptions as in the previous section are used and in addition the strut is treated as a
piston which is free to slide within a cylinder. The free body and force vector
diagrams are constructed as follows.
- The force in the arm, CB, is along its longitudinal axis since it is pin-jointed at
both ends and there are no loads applied along its length.
- For the wheel-axle-strut assembly, DCE, the resultant force at D must pass
through point G for equilibrium
- The force at D is reacted by two components: the spring force in the direction of
AC and a side force on the cylinder. Note that this is only true if the spring is
concentric with the strut AC. The side force is undesirable since it produces
friction and wear on the piston in the strut.. Mounting the spring more nearly in
line with AG can reduce this side force and this is commonly done in McPherson
strut designs.
Slope = spring rate
Static deflection
Spring deflection
Spring Load
F
S
Slope = wheel rate
Static load
position
Wheel deflection
Wheel Load
F
W
Rebound Bump
Max
Min
Figure 1.48 Typical load vs. Deflection characteristics for spring and wheel
Slope = spring rate
Static load position
Spring length
Spring Force
F
S
Full bump
Full Rebound
max
min
Figure 1.49 Spring rate calculation
1-48
A
B C
D
E
G
F
D
F
c
F
w
F
spring
F
slide
F
D
F
D
F
c
F
w F
spring
F
slide
F
D
Free Body Diagram
for wheel/strut assembly
Force vector diagrams
Free Body Diagram
for piston slider
b) Lateral, Longitudinal and Mixed Loads
When a lateral tyre force is applied in addition to a vertical load, the analysis for the
example in Figure 1.47, is modified to that shown in Figure 1.51. The resultant wheel
force is first calculated from the vector combination of F
W
and F
Y
. This then defines a
new point for G and the force acting on the lower link is changed in both magnitude
and direction compared with the previous analysis.
For loads in the fore-aft direction, the situation is shown in Figure 1.52. A fore-aft
load may be applied at either points E or O (i.e. at right angles to the plane of the
diagram.) A force at the tyre-ground contact point, E, would arise from braking with
conventional outboard brakes. The braking torque is transmitted through the stub
axle/kingpin assembly. This may be treated as a single rigid member. If the fore-aft
force is due to drag, shock loading, braking with inboard brakes or traction, then its
effective point of application is at the wheel centre, O.
Figure 1.50 Analysis of a McPherson strut suspension
1-49
Figure 1.51 Combined lateral and vertical loads on double wishbone suspension
In general any force at a given point may be treated as the same force at another point
plus a couple around that point. This is a convenient way of analysing the effect of
these fore-aft forces. If the fore-aft force is moved to the kingpin axis then a moment
of either Fd
1
or Fd
2
acts around the kingpin axis. This moment is, in practice, reacted
by the steering arm and track rod so that the force in the track rod is either:
F
Y
F
W
F
C
F
D G
Resultant wheel force
Figure 1.52 Defining the loading points on a suspension
1-50
1
3
Fd
d
or
2
3
Fd
d
d
3
is the effective moment arm of the steering arm around the kingpin axis. In the side
view (Figure 1.53), this longitudinal force is reacted by the upper and lower arms. It
may then be possible to proceed to analyse all the forces on the upper and lower arms.
However, the analysis is now starting to become complicated. Remember also we are
not taking proper account of all the three dimensional effects. Despite this, our simple
analysis does provide a good indication of the primary forces in the suspension. The
subtle effects arising from a full three-dimensional treatment are of secondary
importance.
Figure 1.53 Longitudinal forces reacted by suspension arms
Suspension travel is limited by bump stops. Normally these are made of rubber and
produce a rapidly increasing spring-rate with deflection. This obviously prevents
damage from metal to metal contact and provides a rapidly increasing force to react
the extreme motion of the wheel. There are two philosophies involved in bump stop
design. The first, and perhaps more conventional approach is to make them relatively
short and stiff; this means that contact is relatively infrequent and restricted to extreme
events. The forces involved are high. The second is to make them relatively long and
compliant; this means that they come into contact much more often, but much lower
forces are involved and they may be viewed as a secondary spring. This effectively
helps to stiffen the suspension for large deflections. Whichever approach is adopted
contact with the bump stops changes the force system in the suspension members.
An example is shown in Figure 1.54. One method of tackling this problem is by
superposition, i.e. by separating the forces arising from the main suspension spring at
full deflection from those due to the bump stop spring. Both sets of suspension loads
are then added together using the principle of superposition.
The procedure is as follows;
(1-23)
(1-24)
Force on lower arm =
|
\

|
.
|
F
E D
CD
1
Force on upper arm =
|
\

|
.
|
F
E D
CD
1
Upper Arm
1-51
- Draw the suspension in the required position (in contact with the bumpstop).
- Ignoring the force due to the bump stop compression, determine the forces and
wheel load F
w
.
- Calculate the forces as described previously - this will give a nominal loads F
c
and F
s
.
- Calculate the excess wheel load carried by the bump-stop,
'
w w ex
F - F F =
- Ignoring the spring force, draw the FBD of the upper arm and determine the
suspension forces arising from the bump-stop load F
w
.
- Determine the total suspension forces by adding the results from steps (c) and (e).
A
B
C
D
F
ex
H
F
b
F
F
A
D
F
b
F
A
F
D
excess load
FBD for upper
arm
Force Diagram
Figure 1.54 Effect of bump stop on suspension force analysis
1-52
1.6 Suspension Geometry to Combat Squat and Dive
Squat (nose up) and dive (nose down) are changes in vehicle body attitude arising
respectively from acceleration and braking. It is possible to design a suspension that
will combat squat or dive, however, the geometry required for anti-squat will be
different to that for anti-dive. Again, this is one of those areas of suspension design
that requires compromises to be made. The design requirements to combat dive
depend on whether a vehicle has outboard or inboard brakes, while anti-squat design
depends on whether the vehicle has dependent or independent suspension. In the latter
case suspension stiffness also plays a part.
1.6.1 Anti-Dive Geometry
The free body diagram for a car during braking is shown in Figure 1.55. Note the
pseudo (DAlembert)-force, mf, in which f is the deceleration (m/s
2
). Braking forces,
B
f
and B
r
, are applied at the front and rear wheel pairs and the relationship between
them is defined by the fixed braking ratio,
k
B
B B
f
f r
=
+
Under these conditions, the vertical loads on the axles are different from their static
values, and can be calculated as follows.
Taking moments about the rear tyre contact patch,
f N
l mfh mgb = 0
f N
mgb
l
mfh
l
= +
(1-25)
(1-26)
Figure 1.55 Free Body Diagram of vehicle during braking
(1-27)
mg B
f
B
r
N
r N
f mf
a b
l
h
1-53
The first term is the static load and the second term is the load transfer due to braking.
Hence, the load at the rear is,
r N
mga
l
mfh
l
=
This load transfer term will cause an increased deflection at the front axle and a
reduced deflection at the rear axle. Hence the change of attitude of the vehicle body.
Outboard brakes
Consider now the diagram of the front suspension shown in Figure 1.56. The axis of
the link pivots are inclined such that the effective centre of rotation of the wheel in
side view is at O
f
.
Figure 1.56 Free Body Diagram for front suspension
The spring force may be expressed as static load plus a perturbation due to braking,
f f f S S S
=
'
+
o
Under static conditions (f = 0) the static spring load is
f S
mgb
l
'
=
For the braking condition, taking moments about O
f
gives,
f f f N
d
S
d
B
e = 0
d
mgb
l
mfh
l
d
S S B
e
f f f
( ) ( ) +
'
+ =
o
0
Removing the static load and setting oS
f
to zero to represent the case in which the
suspension load does not change gives
mfhd
l
B
e
f
= 0
Substituting for the braking force
(1-28)
(1-29)
(1-30)
(1-31)
(1-32)
(1-33)
o
B
f
d
N
f
e
S
f
Spring force
O
f
1-54
f B
mkf =
This results in
e
d
h
kl
= = tano
Hence, if the instantaneous centre of the suspension, O
f
, lies anywhere along the line
defined by equation (1-35) then the condition for no front suspension deflection is
satisfied. If the instantaneous centre is designed to lie below this line at an angle of
' o say, then the percentage anti-dive is defined as
(
tan
tan
)
'

o
o
100%
A similar analysis can be applied at the rear (Figure 1.57) to give
) k 1 ( l
h
d
e
tan

= = | (1.37)
Again, if the instantaneous centre lies along the line defined by equations (1-36) and
(1-37), then will be no change in spring force and hence no lift will occur. Therefore,
the condition for 100% anti-dive is that the front and rear instantaneous centres lie
along the two lines shown in Figure 1.58.
It is common practice to use only around 50% anti-dive. The reasons are:
- Zero pitch braking appears to be undesirable from a subjective viewpoint,
- Full anti-dive conflicts with anti-squat so that some compromise must be reached,.
(1-34)
(1-35)
(1-36)
Figure 1.57 Free Body Diagram for rear suspension
Figure 1.58 Condition for 100% anti-dive
|
B
r
d
N
r
e
S
r
Spring force
h
kl (1-k)l
| o
O
r
1-55
- Full anti-dive can cause large caster angle changes because all the brake torque is
reacted through the suspension links. Such caster changes may tend to make the
steering unacceptably heavy during braking
Inboard Brakes
All the foregoing analysis applies to the most common case in which the brakes are
mounted outboard, (i.e. at the wheel). Inboard brake systems differ in that the brake
torque is reacted through the driveshaft. (Section 1.5.1) The situation for say a front
wheel is shown in Figure 1.59
.
The additional torque applied by the driveshaft to the
wheel unit must be included in the freebody diagram. Taking moments about the
wheel centre,
f f M B
r mfkr = =
Now, taking moments about O
f
gives
f f f N
d
S
d
B
e mfkr + = 0
Comparing this with equation (1-31) it can be seen that an additional term is now
included. Using the same analysis as before leads to,
( )
tan
e r
d
h
kl

= = o
Note from the above expression that the angle o is now defined from the wheel centre
rather than the ground level. The analogous situation occurs at the rear to give
| =

tan
) k 1 ( l
h
d
) r e (
1.6.2 Anti-squat geometry
The analysis for anti-squat or anti-pitch design during acceleration follows a very
similar process to that adopted in the previous section. In Figure 1.60, the pseudo
force is now reversed in direction where f is now defined as acceleration. Tractive
forces may be applied at either axle, or both in the special case of four-wheel drive.
(1-38)
Figure 1.59 FBD for front suspension with inboard brakes
(1-39)
(1-40)
(1-41)
e
M
f
N
f
o
B
f
d
S
f
Spring force
O
f
1-56
There are two general cases to consider; solid axle drive or independent suspension /
driveshaft designs.
Solid rear drive axle
This configuration is shown in Figure 1.61 with the axle assumed to be mounted on
twin trailing arms, with an instantaneous centre of rotation at O
r
. This point is
sometimes referred to as the virtual reaction point (i.e. the point at which the forces in
the suspension arms can be resolved and assumed to act on the vehicle body). Note
that this treatment of the suspension is exactly the same as if a trailing arm of length,
d, connected rigidly to the rear axle acting about a pivot at O
r
. Note that the drive
torque is reacted through the axle casing and the suspension members (thus making
the analysis analogous to the case of outboard brakes). The analysis assumes that the
axle mass is small relative to the vehicle mass. The vertical wheel load is,
r
N
mga
l
mfh
l
= +
mfh
l
is the load transfer onto the rear axle.
The tractive force is,
r
T mf =
Taking moments about O
r
gives
Figure 1.60 FBD for an accelerating car (all-wheel drive)
(1-42)
e
O
Sr
T
r
N
r d
Figure 1.61 Rear axle mounted on twin trailing arms
(1-43)
mg
T
f
T
r
N
r N
f
mf
a b
l
h
O
r
1-57
r r r
T e S d N d + = 0
Substituting for all the forces and eliminating the static spring force gives
r r r S
mf
h
l
e
d
k o o
= = ( )
where o
r
= rear suspension deflection, k
r
= rear spring rate
The front suspension must also deflect under these conditions due to the load transfer
term, mfh l , reacted there. Note that it is not possible to design the front suspension
to directly accommodate this. It was possible to do this for the braking case, because
part of the braking force was reacted there. Hence, the spring force change at the front
is,
f f f S k
mfh
l
o o
= =

The pitch angle, u, of the vehicle is simply,
u
o o
=
( )
r f
l
u =
+
mf
k
h
l
e
d
mfh
l
k
l
r f
( )
) (
d
e h h
l
mf
k lk lk r r f
+ = u
Hence the condition for zero pitch may be expressed as,
e
d
k
k
h
l
r
f
=
+ ( ) 1
Independent suspension rear wheel drive
The freebody diagram is now altered to include the moment provided by the driveshaft
(Figure 1.62). Note that this is similar to the case of the inboard mounted brakes.
Thus equation (1-48) becomes,
r r r r
T e S d N d M + = 0
where M
r
= T
r
r
(1-44)
(1-45)
(1-46)
(1-47)
(1-48)
(1-49)
(1-50)
(1-51)
1-58
Carrying this small difference through the analysis results in the condition for zero
pitch,
( )
( )
e r
d
k
k
h
l
r
f

=
+ 1
Example E2
A multi-purpose vehicle has a live rear axle with outboard brakes. The pivot centre of the rear
suspension is located at O
r
, which is positioned relative to the underside of the wheels by the
dimensions d and e, as shown in Figure E2.1
Relevant vehicle data is as follows:
Vehicle mass 2.1 tonnes
Wheelbase 2.8 m
Vehicle c.g. height above road 0.85 m
Front spring-rate (both sides) 52.5 kN/m
Rear spring-rate (both sides) 35.2 kN/m
Braking ratio (front: rear) 70: 30
a) Determine the ratio e/d for zero pitch during acceleration.
b) What then is the lift at the rear suspension when the vehicle brakes with a constant
deceleration of 0.3g.
Figure E2.1
e
O
Sr
T
r
N
r d
M
r
Torque applied
by driveshaft
Figure 1.62 Independent rear suspension
(1-52)
O
r
d
e
O
r
1-59
Solution
a) The free body diagram of wheel and axle during acceleration is shown in Figure E2.2:
e
O
S
r
T
r
N
r
d
Figure E2.2
It has been shown that for zero pitch during acceleration:
e
d
k
k
h
l
r
f
=
+ ( ) 1
..(E-1)
1-60
Using the vehicle data: 507 . 0
8 . 2
85 . 0
5 . 52
2 . 35
1
d
e
=
|
.
|

\
|
+
=
b) For the lift at the rear suspension due to 0.3g deceleration, note that the vehicle has
outboard brakes
In this case, a braking force replaces the traction force in the above diagram and
acts in the opposite direction.
The moments equation about O
r
is: B
r
e + S
r
d N
r
d = 0.(E-2)
l
h f m
l
a g m
N
r
= ,
r r r
S S S o + ' = , ( ) f m k 1 B
r
=
Substituting into E-2 and cancelling out terms gives:
( ) ( )
(

= = o
l
h
d
e
k 1 f m
l
h f m
d
e
k 1 f m S
r
Change of spring force
936
8 . 2
85 . 0
507 . 0 x 3 . 0 81 . 9 x 3 . 0 x 2100 S
r
= |
.
|

\
|
= o N
The negative sign indicates a force reduction
Lift at rear suspension:
mm 6 . 26
mm / N 2 . 35
N 936
k
S
r
r
r
= =
o
= o
The negative sign indicates lift at the rear axle.
.
1.7 Vehicle ride analysis
Ride comfort is one of the most important characteristics defining the quality of a
vehicle. It is principally related to vehicle body vibration, the dominant source of
which is due to road surface irregularities. In order to design vehicles that have good
ride properties it is essential to be able to model ride performance in the early stages
of vehicle development. This requires an understanding of road surface characteristics,
human response to vibration and vehicle modelling principles. These are introduced in
this section.
1.7.1 Road surface roughness and vehicle excitation
When a vehicle traverses over a road profile, the irregularities in the profile are
converted into time-varying vertical displacement excitation inputs at each tyre
1-61
contact patch. In general, road surfaces have random profiles, which means that the
excitations that they generate are also of a random nature.
The spatial random profile of a road surface can be frequency analysed and shown to
comprise a set of harmonic components having different amplitudes and wavelengths
as shown in Figure 1.63. In general the components having the longest wavelengths
have the largest amplitudes. While only three components are shown in Figure 1.63,
there are in reality an infinite number. For a given frequency component it is possible
to define spatial
10
frequency n as the number of cycles per metre. It can then be shown
[21] that the frequency characteristics of the profile are described as a function of n,
by the spatial power spectral density S(n). The units of S are m
3
/ cycle.
Figure 1.63 Harmonic components of a random road surface profile
From large amounts of measured road data it has been established [22], that S and n
are related approximately by:
( ) S n n =

k
2 5 .
k is the road roughness coefficient.
When a vehicle travels at a velocity V m/s over a random road profile described by
S(n), the resulting random excitation to the vehicle is described by a temporal
11
spectral density S(f), where f = n V is the excitation frequency in Hz.
It may be shown [21] that
S f
S n
V
( )
( )
=
10
Spatial position related
11
Temporal time related
(1-53)
(1-54)
1-62
It follows that S f V f ( )
. .
=

k
1 5 2 5
The units for S(f) are m
2
/ Hz. Typical values of k for a motorway, a principal road
and a minor road are 0.25.(10)
-6
, 4.(10)
-6
and 15.(10)
-6
m
2
/(cycle/m)
1.5
respectively for
0.01<n<10 cycle/m [22]. The variation of S(f) for a vehicle traversing a poor minor
road at 20 m/s is shown in Figure 1.64. It is important to note the distribution of
frequency components in this spectrum in relation to the natural frequencies of a
suspended vehicle.
1 10 100
1 10
7
1 10
6
1 10
5
1 10
4
0.001
0.01
Frequency, Hz
Figure 1.64 Spectral density of a road input as a function of vehicle speed (poor
minor road)
1.7.2 Human perception of ride
Ride is dominated by the level of road induced vibration in a vehicle body. In
assessing ride performance of a vehicle it is necessary to take into account the
sensitivity of the human body to vibration.
The human body is sensitive to vibration and has a number of natural frequencies. It
follows that when the human body is subjected to excitation at any of these
frequencies resonance and discomfort will result. From the last section it obvious that
road induced vibration will be broadband and hence some of the components of the
excitation will tend to produce human body resonance. From an understanding of
human body resonances ([23], [24]) it is possible to weight the vehicle body response
in order to obtain a single figure assessment of ride performance.
Human response to whole body vibration
From a vibration point of view the human body can be considered to be a complex
assemblage of mass, elastic and damping elements that result in a number of natural
frequencies in the range from 1 to 900 Hz. From a ride point of view we are concerned
with whole-body vibration of a seated person in the range from 0.5 to 15 Hz. There
are two whole-body natural frequencies in this range. They are associated with the
head-neck (1-2 Hz) and the thorax-abdomen (4-8 Hz). Head-neck resonance can be
excited by vehicle pitching and rolling motions, while resonance of the thorax-
abdomen can be produced by vehicle bounce motion.
(1-55)
Spectral
Density
m
2
/ Hz
1-63
In general the tolerance to whole-body vibration decreases with time, so a high level
of vibration for a short time can produce the same level of discomfort as a low level
for a long time. Equal discomfort curves have been produced in standards, e.g.
International Standard ISO 2631 [25] and British Standard BS. 6841:1987 [26]. Some
examples are given in Figure 1.65 for vertical and longitudinal / lateral excitation. It
should be noticed that the human body is acceleration-sensitive and the minimum
values of the graphs correspond to the resonances discussed above.
1 10 100
0.1
1
10
Frequency, Hz
R
.
m
.
s
.
a
c
c
e
l
e
r
a
t
i
o
n
a
z
,
m
/
s
^
2
1 10 100
0.01
0.1
1
10
Frequency, Hz
R
.
m
.
s
.
a
c
c
e
l
e
r
a
t
i
o
n
s
a
x
a
n
d
a
y
,
m
/
s
^
2
(a) Vertical vibration (b) Longitudinal and lateral directions
Figure 1.65 Equal discomfort curves
The shapes of the graphs are used to establish weighting curves that can be applied to
vehicle body accelerations determined from simulations of vehicle response to road
induced vibration. Thus enabling a single figure estimate of human response to be
obtained. A typical weighting curve for vertical vibration is shown in Figure 1.66.
Figure 1.66 Weighting curve for vertical whole-body acceleration
1.7.3 Vehicle models
Ride performance is assessed at the design stage by simulation of vehicle response to
road excitation. This requires the development of a vehicle model and analysis of its
response.
2 8 4
8h
8h
4h
4h
Weighting,
dB
-12
--8
--4
-0 -0
+4
0.5 1 2 4 16
1-64
Models of varying complexity are used. For a passenger car, the most comprehensive
of these has seven degrees of freedom (DOFs) - Figure 1.67. These comprise three
DOFs for the vehicle body (pitch, bounce and roll) and a further vertical degree of
freedom at each of the four unsprung masses. This model allows the pitch, bounce and
roll performance of the vehicle to be studied.
Figure 1.67 Full vehicle model
The suspension stiffness and damping rates in the model are derived from the
individual spring and damping units using the approach discussed in Section 1.3.
Various tyre models have been proposed [27]. The simplest of these uses a point-
contact model to represent the elasticity and damping in the tyre with a simple spring
and viscous damper. Since tyre damping is several orders of magnitude less than
suspension damping it can be neglected in basic vehicle models.
Much useful information can be derived from simpler vehicle models. Because for a
normal road surface profile the components having the longer wavelength components
are in phase (coherent) across the left and right tracks of a vehicle, there is no
tendency to excite body roll. This then justifies the use of a half vehicle model as
shown in Figure 1.68. This has four DOFs; a body mass translation and rotation, plus
a translation for each of the unsprung masses.
Figure 1.68 Half-vehicle model
The half vehicle model can be simplified still further to a quarter vehicle model, if the
body mass satisfies certain conditions. Let the sprung mass in Figure 1.68 be denoted
Sprung
mass
Unsprung
masses
a b
L
G
1-65
by M and its moment of inertia about its centre of gravity G by I. G is located by
distances a and b. The wheelbase is L. For the sprung mass in Figure 1.69 to be
dynamically equivalent to that in Figure 1.68 the following conditions have to be
satisfied.
Equal mass: M M M M
G r f
= + + (1.56)
Mass centre in the same position: b M a M
r f
= (1.57)
The same moment of inertia: I b M a M
2
r
2
f
= + (1.58)
Figure 1.69 Half-vehicle model with dynamically equivalent sprung mass
From equation 1.57 and 1.58
I b
b
a
M a M
2
f
2
f
= +
Re-arranging ( ) I b a a M
f
= +
But L b a = +
Hence
L a
I
M
f
= (1.59)
It follows that
L b
I
M
r
= (1.60)
Substituting into equation 1.56:
( )
b a
I
M
L b a
b a
I M
L b
I
L a
I
M M
G
=
+
= =
Denoting I by M r
G
2
, where r
G
is the radius of gyration of the sprung mass about
lateral axis through G, it follows that:
Unsprung
masses
a b
L
M
f
M
G
M
r
G
1-66
|
|
.
|

\
|
=
b a
r
1 M M
2
G
G
(1.61)
If the inertia coupling ratio 1
b a
r
2
G
= , M
G
= 0
12
. Then the dynamically equivalent
system comprises a sprung mass M
f
at the front suspension and a sprung mass M
r
at
the rear suspension and the motions at the front and rear suspensions are uncoupled.
For this case a quarter vehicle model as shown in Figure 1.70 can represent either the
front or rear suspension.
Figure 1.70 Quarter vehicle model
1.7.4 Vibration analysis of the quarter vehicle model
13
The model shown in Figure 1.70 consists of linear elements, i.e. the springs and
dampers have linear characteristics. These are valid approximations for small
oscillations about the static equilibrium position of the system. Let x
0
, z
1
and z
2
be
time varying displacements from this equilibrium position.
a) Vibration characteristics of a quarter vehicle model (see Appendix A1)
For this case we assume C
s
= 0 and examine the free vibration of the model (x
0
= 0).
The equations of motion are
( )
)
`

=
)
`

+
+
)
`

0
0
z
z
K K
K K K
z
z
M 0
0 M
2
1
s s
s s t
2
1
s
u


(1.62)
Assume solutions z
1
= A
1
sin et and z
2
= A
2
sin et. Substitute these expressions and their
second differentials into (1.62) and cancel throughout by sin et:
12
The inertia coupling ratio ranges typically from 0.8 for sports cars to 1.2 for some
front wheel drive cars.
13
For further details of vibration analysis see [28]
Sprung
mass, M
s
Unsprung
mass, M
u
C
s
K
s
K
t
z
1
z
1
x
0
1-67
( )
( )
)
`

=
)
`

(
(

e
e +
0
0
A
A
M K K
K M K K
2
1
2
s s s
s
2
u s t
(1.63)
The non-trivial solution of (1.63) is
( )
( )
0
M K K
K M K K
2
s s s
s
2
u s t
=
e
e +
(1.64)
This is called the frequency or characteristic determinant, i.e. describes the vibration
characteristics of the system.
Consider a typical set of data for the rear of a mid-range saloon car:
Sprung mass M
s
= 317.5 kg
Unsprung mass M
u
= 45.4 kg
Suspension stiffness (wheel-rate) K
s
= 22 kN/m
Tyre stiffness K
t
= 192 kN/m
Substitute the into (1.64), expand and then simplifying the result gives:
0 10 x 293 4786
3 2 4
= + e e
The solution to this quadratic equation in e
2
is:
( )
2
10 x 293 x 4 4786 4786
,
3 2
2
2
2
1

= e e

From which the natural frequencies are e
1
= 7.87 rad/s (f
1
= 1.25 Hz) and e
2
= 68.73 rad/s (f
2
= 10.9 Hz). These results are very typical of what one would expect for a car suspension.
Corresponding to each natural frequency there is a mode of vibration. These describe the
relative amplitudes of free vibration of the masses at each of the natural frequencies. They
are determined by substituting first e
1
2
and then e
2
2
into either of the equations in (1.63).
First mode of vibration
Substituting e
1
2
into the second equation and add a second suffix 1 to A
1
and A
2
to denote
that they are the amplitudes at the first frequency:
( ) 0 A M K A K
21
2
1 s s 11 s
= e + (1.65)
For the numerical data: 0122 . 0
A
A
1
2
1
=
|
|
.
|

\
|
This indicates that since the result is positive, both masses move in the same direction at any
instant and the amplitude at the unsprung mass is 0.0122 that of the sprung mass. This mode
of vibration is called the body mode and can be represented graphically as shown in Figure
1.71a. Note amplitudes are plotted horizontally for clarity.
1-68
Second mode of vibration
Substituting e
2
2
into the second equation and add a second suffix 2 to A
1
and A
2
to denote
that they are the amplitudes at the second frequency:
( ) 0 A M K A K
22
2
2 s s 12 s
= e + (1.66)
For the numerical data: 7 . 7
A
A
2
2
1
=
|
|
.
|

\
|
This indicates that since the result is negative, the masses are moving in opposite directions
at any instant and the amplitude at the unsprung mass is 7.7 times that of the sprung mass.
This mode of vibration is called the wheel-hop mode and can be represented graphically as
shown in Figure 1.71b.
(a) Body mode (b) Wheel-hop mode
Figure 1.71 Vibration modes for a quarter-vehicle suspension.
From the above analysis it is apparent that for the body mode the amplitude of the unsprung
mass is very small compared to that of the sprung mass, so in this case the system could be
approximated as shown in Figure 1.72a. Also, for the wheel-hop mode the amplitude of the
sprung mass is very small in comparison to that of the unsprung mass. For this case the
system could be approximated as shown in Figure 1.72b. The corresponding approximate
natural frequencies and damping ratios are as shown in the diagrams.
(a) Body mode
(b) Wheel-hop mode
Figure 1.72 Approximate models for body mode and wheel-hop mode
A
22
7.7
-1
1
0.0122
A
21
A
11
f
1
= 1.25 Hz
K
s
M
u
K
t
M
s
K
s
C
s
C
s
s
s
1
M
K
2
1
f
t
=
s s
s
1
K M 2
C
= ,
u
t s
2
M
K K
2
1
f
+
t
=
( )
t s u
s
2
K K M 2
C
+
= ,
A
12
f
2
= 10.9 Hz
A
22
1-69
b) Frequency response of the quarter vehicle model (see Appendix A1)
In this case we are concerned with the complete quarter vehicle model shown in Figure 1.70
and begin with the equations of motion for the system:
( )
)
`

=
)
`

+
+
)
`

+
)
`

0
x K
z
z
K K
K K K
z
z
C C
C C
z
z
M 0
0 M
0 t
2
1
s s
s s t
2
1
s s
s s
2
1
s
u

(1.67)
Assume
i) A harmonic input at the base of the tyre spring, ( ) | |
t i
e X t X x
e
e
0 0 0
Im sin = = .
ii) Steady state solutions for the response of the masses as | |
t i
e Z z
e
1 1
Im = and
| |
t i
e Z z
e
2 2
Im = . The quantities Z
1
and Z
2
are in general complex quantities where a
complex exponent indicates a phase change with respect to the input waveform.
For the purpose of solving the equations of motion, the exponential form of the input and
output responses is used. The square root of the sum of the squares of the resulting complex
expression give the amplitude response and the inverse tangent of the imaginary part divided
by the real part give the phase change with respect to the input waveform.
When these expressions (and their first and second order differentials) are substituted into
(1.67) it is possible to cancel throughout by e
iet
. After some manipulation the result is:
( )
( )
)
`

=
)
`

(
(

e + e e
e e + + e
0
X K
Z
Z
M K i C K i C
K i C M K K i C
0 t
2
1
2
s s s s s
s s
2
u s t s
(1.68)
Note that Z
1
and Z
2
are complex in nature due to the inclusion of damping, i.e. the vibration
of the masses lags behind the excitation. Solving this pair of non-homogeneous equations in
Z
1
and Z
2
produces the following expressions:
( ) | |
( )
( )
2
s s s s s
s s
2
u s t s
2
s s s t
0
1
M K i C K i C
K i C M K K i C
M K i C K
X
Z
e + e e
e e + + e
e + e
= (1.69)
and
( )
( )
( )
2
s s s s s
s s
2
u s t s
s s t
0
2
M K i C K i C
K i C M K K i C
K i C K
X
Z
e + e e
e e + + e
+ e
= (1.70)
Replacing e with 2 t f (where f is the excitation frequency in Hz), in these expressions leads
to frequency response functions H
1
(f) and H
2
(f) as follows:
( ) ) f (
X
Z
f H
0
1
1
= (1.71) and ( ) ) f (
X
Z
f H
0
2
2
= ..(1.72)
H
1
(f) and H
2
(f) are complex of the form H
1r
(f) + i H
1i
(f) and H
2r
(f) + i H
2i
(f)
respectively. It follows that the amplitude responses are given respectively by:
1-70
( ) ( ) ( ) ( ) ( )
2
i 1
2
r 1 1
f H f H f H + = (1.73)
and ( ) ( ) ( ) ( ) ( )
2
i 2
2
r 2 2
f H f H f H + = (1.74)
where r denotes the real part and i the imaginary part of each function. H
1
(f) and
H
2
(f) can be thought of as transfer functions between the respective parts of the
system.
1.7.5 Response of a linear system to random excitation [28]
Consider a general linear structural system as shown in Figure 1.73. If A is the
location of a random excitation having a psd S
A
(f), and B is the location of the output
from the system, having a psd S
B
(f), it can be shown [22] that:
( ) ( ) ( ) f S f H f S
A
2
B
= , (1.75)
where H(f) is the frequency response function relating A to B. Furthermore it can be
shown that the r.m.s. value of a stationary random signal is equal to the area under the
psd graph. This requires lower and upper frequency limits to be specified. For
suspension performance analysis these can be taken to be between 0.5 and 15 Hz.
Figure 1.73 Input and output spectral densities
1.7.6. Parameters used in assessing suspension performance
There are three key parameters as follows.
a) Discomfort Parameter ACC
This is defined as the r.m.s. sprung-mass acceleration weighted for human response.
The FRF for the weighted sprung-mass acceleration is given by:
( ) ( ) ( )
acc 2 hrw ACC
f H f H f H = (1.76)
H
hrw
(f) = FRF for the human response weighting for vertical vibration
( ) ( ) ( ) f H f 2 f H
2
2
acc 2
t = = FRF for the sprung mass acceleration
ACC is then given by:
( ) ( )
}
=
15
5 . 0
2
ACC
df f S f H ACC (1.77)
S(f) = power spectral density associated with the displacement input x
0
.
i
o
Input PSD, S
i
(f)
Output PSD, S
o
(f)
System
1-71
b) Dynamic Tyre Load Parameter DTL
This is defined by the r.m.s. tyre-load fluctuation. Since the tyre load variation from
static is given by K
t
(x
0
z
1
), it follows that the FRF for tyre load fluctuation is given
by:
( )
( )
( ) ( ) f H 1 K
X
Z X
K f H
1 t
0
1 0
t DTL
=

= (1.78)
DTL is then given by:
( ) ( )
}
=
15
5 . 0
2
DTL
df f S f H DTL (1.79)
c) Suspension Working Space Parameter SWS
This is defined as the r.m.s. relative displacement between sprung and unsprung
masses. The relative displacement is random and Gaussian. If the bump-stop to bump-
stop displacement is some multiple of SWS, there will be some bump-stop contact. If
for example the unsprung mass is equidistant between the bump-stops in the static
position and the maximum suspension travel is n x SWS, the percentage of time that
bump-stop contact occurs is given in Table 1.2.
n 1 2 3
Contact time, % 31.7 4.6 0.03
Table 1.2 Bump-stop contact related to suspension travel and SWS
One of the design aims is to minimise bump-stop contact, therefore a value of n = 3
would be a realistic choice.
The FRF for relative movement between unsprung and sprung masses is:
( ) ( ) ( ) f H f H f H
SWS 2 1
= (1.80)
SWS is then given by:
( ) ( )
}
=
15
5 . 0
2
SWS
df f S f H SWS (1.81)
1.7.7 Quarter vehicle model results and design guidelines
The equations given above can be used to investigate the effect of suspension stiffness
and damping for a given road surface and vehicle speed. Some typical results are
shown in Figure 1.74.
The objective from a design point of view is to have low body acceleration and low
dynamic tyre load. For a given SWS of 30 mm it is seen (upper diagram) that low
damping requires higher suspension stiffness than would be required by high
damping. In terms of ACC (middle diagram), high damping produces higher body
acceleration than low damping. From the bottom diagram, DTL is relatively
insensitive to change in stiffness, while over the design range DTL favours higher
1-72
damping. In general, damping should be between 0.3 and 0.5 of critical - at the lower
end for optimum comfort and the higher end for optimum handling.
Figure 1.74 Typical quarter vehicle model results
The conflicting issues regarding stiffness and damping selection can be clarified if a
full set of results (for a number of different damping values), are plotted on a conflict
diagram for a selected SWS (Figure 1.75). Note f
1
and ,
1
are given by the equations in
Figure 1.72a.
Figure 1.79 Conflict diagram
Figure 1.75 Conflict diagram
DTL, kN
ACC, m/s
2
Point f
1
,
1
Conventional
design range
Stiffness
increasing
1-73
The optimum points on the conflict diagram (to minimise ACC and DTL) are 2 and 3.
However, from a practical point of view, this will tend to make the suspension
stiffness too low. The result will be large amounts of body roll during cornering. This
can be counteracted with very stiff anti-roll bars. Unfortunately these can degrade ride
on roads having significantly different characteristics across the road lane width. A
better selection of parameters would be based on point 1.
1.8 The Influence of Chassis Design on Vehicle Handling
1.8.1 Steady State Cornering and Stability
Appendix A2 outlines the dynamics of a simple handling model. Note the
assumptions and the notation. When forward speed U and steering input o
f
are
constant, v and r are constant for steady state conditions representing the lateral
velocity and yaw velocity (an angular term) respectively (i.e. 0 = = r v ). Under these
conditions equation A2.11 reduces to:
Denoting r as r
ss
(steady-state yaw rate) the motion is described in Figure 1.76 and the
simultaneous equations 1.82 can be solved to give:
where L = a + b = wheelbase
Figure 1.76 A steady-state turn
Steady-state yaw rate can be related to path curvature:
82 . 1 ... ..........
C a
C
r
v
U
C b C a
U
C b C a
U
C b C a
U m
U
C C
f
f
f
r
2
f
2
r f
r f r f
o
)
`

=
)
`

(
(
(
(

+
+
( )
83 . 1 .......... ..........
C a C b U m C C L
C C L U r
f r
2
r f
2
r f
f
ss
+
=
o
Tyre force
Tyre force
Front slip angle
Rear slip angle
Yaw rate r
ss
v
ss
Centre of turn
Radius of turn
Constant forward
speed U
o
f
1-74
84 . 1 .......... .......... U
R
1
U r
ss ss
= |
.
|

\
|
=
Substitute into 1.83 and re-arrange to give:
Where:
In this relationship b C
r
a C
f
is called the Stability Margin. Its magnitude (positive,
zero or negative) determines the behaviour of the vehicle. Possible effects as a
function of speed U are shown in Figure 1.77.
Figure 1.77 Effect of speed on
ss
/ o
f
for different K
Consider the three possibilities.
K = 0 (b Cr = a Cf) Neutral Steer condition
L
1
f
ss
=
o

. This represents a pure rolling (Ackermann) condition. According to Figure

1.78,
R
L
f
~ o
K > 0, (b C
r
> A C
f
), Understeer condition
The vehicle is always stable and tends to run wide. o
f
needs to be increased to
maintain
ss
.
K < 0, b (C
r
< a C
f
), Oversteer condition
85 . 1 ...... .......... ..........
U K L
1
2
f
ss
+
=
o

( )
86 . 1 .......... ..........
C C L
C a C b m
K
r f
f r

=
K<0, Oversteer
K<0, Oversteer
Forward
speedU
K>0, Understeer
K=0, Neutral steer
Critical speed
U
crit

ss
/ o
f
1-75
There is a critical speed at which ss tends to . This occurs at a speed U
crit
given by:
Figure 1.78 Ackermann vehicle in a steady turn
U < Ucrit

ss
increases with speed. At a given speed (< Ucrit) it is necessary to decrease
o
f
to maintain
ss
.
U > U
crit
Need to apply negative steering lock to maintain
ss
requires special skill.
1.8.2 The Influence of Chassis Design on Handling
Despite the many limitations in the analysis of the basic handling model, a number of
important deductions can be made as to how chassis design can influence handling
stability. In effect we envisage ways of modifying cornering stiffness producing
revised values C
f
and C
r
. We limit our attention to the following:
a) Load transfer
b) Camber angle
c) Compliance steer.
Load Transfer
When a vehicle corners load is transferred from the inside to the outside wheels. In the
case of an unsprung vehicle (Figure 1.79) with a static axle load F
z0
= mg / 2, the
lateral load transfer is:
o
f
Centre
of turn
a
b
L
R
o
f
tan o
f
L
R
=
87 . 1 .... .......... ..........
K
L
U
crit

=
1-76
88 . 1 ........ ..........
t 2
h a m
F
y
z
= A
Figure 1.79 Load transfer unsprung vehicle
Figure 1.80 shows a typical tyre characteristic for F
y
as a function of F
z
. The graph
shows the vertical tyre load for zero acceleration F
z0
and the corresponding lateral tyre
force capability F
y0
. For a lateral load transfer AF
z
the lateral tyre force capability at
inner and outer wheels are F
y,inner
and F
y,outer
,.
Figuer 1.80 Tyre characteristics and lateral load transfer
The following conclusions can be drawn regarding the total lateral tyre force
capability F
y,total
(= F
y,inner
+ F
y,outer
) as a function of AF
z
:
1. F
y,total
is less than 2F
y0
.
2. F
y,total
decreases as AF
z
increases, i.e. F
y,total
decreases with increasing acceleration.
3. F
y,total
is related to the effective cornering stiffness C (C decreases in proportion
to F
ytotal
), so by inducing more load transfer AF
z
, C can be reduced.
4. Since bC
r
needs to be greater than aC
f
for an understeer vehicle, understeer can
be promoted by making C
f
smaller. This can be achieved by inducing more load
transfer at the front axle.
On a sprung vehicle, load transfer is reacted by the roll stiffness at front and rear axles
- the higher the roll stiffness at a given axle the more the load transfer. The balance of
t t
F
zo
-AF
z
h
ma
y
mg
F
zo
+AF
z
Vertical tyre
load, F
z
Lateral tyre force, F
y
Outer wheel
Inner wheel
Static wheel load
Outer wheel Inner wheel
F
z0
AF
z
AF
z
F
y,outer
F
y,inner
F
y0
1-77
load transfer at front and rear axles can be changed by altering the roll stiffness at
these axles. It follows that by increasing the roll stiffness at the front axle, the
cornering stiffness can be decreased there and vehicle understeer can be made more
pronounced. In practice this can be achieved by adding or increasing anti-roll bar
stiffness. It is then possible for the vehicle to be stable up to its cornering limit.
Camber Angle
Wheel camber produces a lateral tyre fiorce called camber thrust. Its magnitude is a
function of camber angle and vertical load. Denoting camber angle (see Fig 1.2),
camber stiffness can be defined as:
When a vehicle is stationary is small, but with suspension movement can change,
depending on the type of suspension. For small camber angles camber thrust is given
by:
90 . 1 .... .......... C F
camber , y
=

F
y,camber
is a second order force compared to F
y
due to a slip angle and together they
produces an effective cornering stiffness C. Camber thrust can be used to make minor
adjustments to handling balance. Figure 1.81 shows how a net reduction in cornering
force can be achieved due to camber thrust. Front positive camber used with negative
rear camber can be used to promote understeer.
Figure 1.81 Reduction of cornering force due to camber thrust
Compliance steer
Cornering force can induce steer due to compliance in the suspension system. This can
be deliberately used to fine-tune and improve the handling of a vehicle. It is a
feature exploited currently by many vehicle manufacturers.
89 . 1 ....... ..........
F
C
0
y
=

c
c
=
m a
y
Lateral tyre force
due to camber
Lateral tyre force
due to slip angles
Parallel links result in
equal body roll and
camber angles
1-78
Assume that the cornering forces produce a small compliance steer o
r,comp
at the rear
axle. If this is in the same direction as the steering input at the front axle o
f
, it moves
the centre of curvature away from the vehicle, thereby increasing the understeer
tendancy Figure 1.82. It can be viewed as an increase in the rear cornering stiffness
C
r
, thereby increasing the margin of stability.
Figure 1.82 The effect of rear compliance steer
o
f
o
f
o
r
o
r
To centre
of turn
1-79
References 1
1. Reimpell, J; Stoll, H and Betzler, JM. The Automotive Chassis: Engineering
Principles, 2
nd
Ed; Butterworth-Heinemann 2001.
2. Gillespie, T.D. Fundamentals of Vehicle Dynamics. SAE. 1992
3. Dixon, J.C. Tyres, Suspension and Handling, (2nd Ed). Edward Arnold.
(distributed by SAE), 1996.
4. Milliken, W.F. and Milliken, D.L. Chassis Design: Principles and Analysis,
Professional Engineering Publishing, 2002.
5. Bastow, D and Howard, G. Car Suspension and Handling, (3rd Ed). Pentech
Press, London. (distributed by SAE, Warrendale, USA.) 1993.
6. Happian-Smith, J (Ed). An Introduction to Modern Vehicle Design, Butterworth-
Heinemann, 2002.
7. Anon. Spring Design Manual. 2
nd
Ed; SAE. 1999
.
8. Dixon, JC. The Shock Absorber Handbook, SAE. 1999
9. Olley, M. Independent wheel suspensions - its whys and wherefores. SAE
Journal, Vol 34, pp73-81, 1934.
10. Enke, K. Improvement of the ride and handling compromise by progress in the
elasto-kinematic system of wheel suspension. Conf on Road Vehicle Handling,
Nuneaton. Proc I MechE pp91-96, 1983.
11. Heisler, H. Advanced Vehicle Technology, Edward Arnold, 1989
12. Hillier, V.A.W. Fundamental of Motor Vehicle Technology. 4th Ed. Thornes.
1991
13. Wahl, A.M. Mechanical Springs, 2nd Ed. McGraw-Hill 1963
14. Shigley, J.E. and Mitchell, L.D. Mechanical Engineering Design. McGraw-Hill.
1983.
15. Mott, R.L. Machine Elements in Design. Charles E Merrill. 1985.
16. Bosworth, R. Rovers System Approach to Achieving First Class Ride Comfort
for the New Rover 400. Automotive Refinement (Selected papers from
Autotech95), pp113-122, IMechE, 1996.
17. Grosjean, J. Principles of Dynamics. Stanley Thornes. 1986.
1-80
18. Horton, D. VDAS - Vehicle Dynamics Analysis Software, Users Manual. Version
3.4. University of Leeds, Feb 1992.
19. Ryan, R. Multibody Systems Analysis Software, Multibody Systems Handbook
(Ed W Schielen), Springer Verlag, 1990
20. Dixon, J.C. The roll centre concept in vehicle dynamics. Proc IMechE, Vol 201,
No D1, pp69-78, 1987.
21. Wong, J.Y. Theory of Ground Vehicles. (2nd Ed) SAE. 1993
22. Dodds, C.J. and Robson, J.D. The description of road surface roughness J
Sound and Vib. 31 (2), pp175-183, 1973.
23. Butkunas, A.A. Power spectral density and ride evaluation. Sound and
Vibration, 1 (12), pp25-30, 1967.
24. Reason, J. Man in Motion: the Psychology of Travel. Weidenfeld and Nicholson.
London. 1974.
25. ISO 2631-1 Guide for the evaluation of human exposure to whole-body
vibration. International Standards Organisation. International Standards
Organisation. 1978.
26. B.S. 6841:1987 Guide to the measurement and evaluation of human exposure to
whole-body mechanical vibration and repeated shock. British Standards
Institution, 1987.
27. Bohm, F and Wilhemeit, H-P (Eds). Tyre Models for Vehicle Dynamic Analysis.
Proc 2nd Int Colloquium on Tyre models for vehicle dynamic analysis, Berlin
1997. Supplement to Vehicle System Dynamics, vol 27. Swets and Zeitlinger,
1997.
28. Thompson, W.T. Theory of Vibration with Applications, 3
rd
Ed. Unwin Hyman.
1989.
29. Meirovitz, L. Elements of Vibration Analysis, 2nd Ed; McGraw-Hill, 1986.
30. Rao, S.S; Mechanical Vibrations, 3rd Ed, Addison-Wesley, 1995.
31. Dimaragonas, A.D and Haddad, S; Vibration for Engineers, Prentice-Hall
International, 1992.
32. Schwarzenbach, J and Gill, K.F. System Modelling and Control. Arnold, 1984
1-81
Appendix A1: Summary of vibration fundamentals [29,30,31]
The general approach to vibration analysis is to:
- Develop a mathematical model of the system and formulate the equations of
motion
- Analyse the free vibration characteristics (natural frequencies and modes)
- Analyse the forced vibration response to prescribed disturbances
- Investigate methods for controlling undesirable vibration levels if they arise.
Mathematical models
These provide the basis of all vibration studies at the design stage. The aim is to
represent the dynamics of a system by one or more differential equations. The most
common approach is to represent the distribution of mass, elasticity and damping in a
system by a set of discrete elements and assign a set of coordinates to the masses.
- Elasticity and damping elements are assumed massless there is a need to know
the constituent equation describing the character of these elements
- The number of degrees-of-freedom (DOFs) of the system is determined by the
number of coordinates, e.g. the model of a simply-support beam in Figure A1.1
- Each coordinate z
1
, z
2
.. etc is a function of time t
- The number of elements chosen dictates accuracy - aim to have just sufficient
elements to ensure that an adequate number of natural vibration modes and
frequencies can be determined while avoiding unnecessary computing effort
- An n-DOF system will be described by n-second order differential equations and
have n-natural frequencies and modes.
- The simplest model has one DOF
Formulating equations of motion
- Equation(s) of motion can be determined by drawing a free-body diagram (FBD)
of each mass including all relevant forces and applying Newtons 2
nd
law
- Each equation of motion is a second order differential equation
- For a multi-DOF system the equations can be assembled in matrix form
- Figure A1.2 shows the FBDs of the quarter vehicle model discussed in Section
1.7.4
- z
1
(t) and z
2
(t) are the generalised coordinates (measured from the static
equilibrium position). x
0
(t) represents the dynamic displacement input from the
terrain
z
n
(t)
z
3
(t) z
2
(t) z
1
(t)
Figure A1.1 A lumped-parameter model of a simply-
supported beam
1-82
- Applying Newtons second law to the unsprung mass gives:
( ) ( ) ( )
1 u 2 1 s 2 1 s 1 0 t
z M z z C z z K z x K =
( ) ( )
0 t 1 t 1 t 2 1 s 2 1 s 1 u
x K z K z K z z K z z C z M = = + + +
- Applying Newtons second law to the sprung mass gives:
( ) ( )
2 s 2 1 s 2 1 s
z M z z C z z K = +
( ) ( ) 0 z z K z z C z M
2 1 s 2 1 s 2 s
=
- These equations can be re-expressed in matrix form as:
K
s
(z
1
z
2
)
C
s
( )
2 1
z z
K
t
(x
0
z
1
)
M
u
1 1 1
, , z z z
K
s
(z
1
z
2
) C
s
( )
2 1
z z
M
s
2 2 2
, , z z z
z
2
(t)
z
1
(t)
x
0
(t)
(b) Free body diagrams
Figure A1.2 The quarter vehicle model and free-body diagram
diagrams
(a) Model
M
s
K
s
C
s
M
u
K
t
1-83
( )
)
`

=
)
`

+
+
)
`

+
)
`

0
x K
z
z
K K
K K K
z
z
C C
C C
z
z
M 0
0 M
0 t
2
1
s s
s s t
2
1
s s
s s
2
1
s
u

- The equations can be written concisely as:
| |{ } | |{ } | |{ } ( ) { } t F z K z C z M = + +
[M], [C] and [K} are the mass / inertia, damping and stiffness matrices
{ } { } and z , z { } z are displacement, velocity and acceleration vectors
{F(t)} is the excitation vector
Single degree of freedom systems
- Knowledge of SDOF behaviour provides an understanding of more complex
systems.
- For the SDOF system in Figure A1.3 the equation of motion is:
( ) t F z k z c z m = + +
- The system characteristics are obtained from the free-vibration behaviour (F(t) =
0)
- With c = 0, there is simple harmonic motion
( ) | e = t cos Z z
n
,
Z (the amplitude) and | are arbitrary constants (determined from the conditions at t
= 0) and e
n
is the undamped natural frequency given by:
m
k
n
= e
Figure A1.3 Notation for SDOF system
- When 0 c = , the level of damping can be described in terms the damping ratio ,
defined as:
c
c
c
= , , mk 2 c
c
=
- When disturbed, the mass either returns to its equilibrium position with or without
oscillation
- Two possible characteristics:
, < 1, the system is under damped - decaying oscillation
m
k
F(t)
z(t)
c
1-84
1 > , , the system is over damped - the mass returns to its equilibrium without
oscillation
- When , = 1 the system is critically damped
- For an under damped system: ( ) | e =
e ,
t cos e Z z
d
t
n
, where Z and | are
arbitrary constants and e
d
is the damped natural frequency given by:
2
n d
1 , e = e
- If F(t) = 0 the solution to the equation of motion has two components: the
Complementary Function (CF) and the Particular Integral (PI).
- The CF is identical to the free vibration solution and quickly dies away with
realistic levels of damping
14
to leave z = PI.
- If ( ) F t F t =
0
sine , the steady-state response of the mass (after the CF has become
zero) is given by: ( ) ( ) | | e o e e = t sin A z
- A(e) is the steady-state amplitude and o(e) is the phase lag - both are dependent
on e.
- It may be shown that
( ) ( )
( ) ( )
( ) ( )
2
2
2
0
2
0 0
c m k
F
i c m k
1
F H F A
e + e
=
e + e
= e = e
- H(e) is called the frequency response function (complex) - relates the input
(excitation) to the output (response) in the frequency domain
- The amplitude response can be presented in dimensionless form in terms of the
dynamic magnifier
0
F A k D = and frequency ratio r
n
= e e
- It may be shown [30] that
(
(

|
|
.
|

\
|
e
e
, +
|
|
.
|

\
|
e
e

=
2
n
2
n
2 1
1
D
- The variation of D for two different values of , are shown in Figure A1.4:
0 1 2 3 4
0
1
2
3
4
5
6
Frequency ratio, r
D
y
n
a
m
i
c
m
a
g
n
i
f
i
e
r
,
D
Figure A1.4 Amplitude response in terms of D
- Conclude from Figure A1.4 that:
a) Maximum amplitude occurs at resonance - when e e ~
n
14
In practice , ranges from approximately 0.02 for low damping elastomers to 0.5 for
vehicle suspensions.
, = 0.1
, = 0.3
1-85
a) The amplitude is strongly dependent on the damping in the system when
e e ~
n
b) When e and e
n
are very dissimilar, damping has very little effect on response
amplitude an important point when considering the use of damping to
control vibration levels.
Multi-degree of freedom systems (MDOF-systems)
- Equations of motion in matrix form: | |{ } | |{ } | |{ } ( ) { } t F z K z C z M = + +
- Characteristics (F(t)=0), given by the solution of:
| |{ } | |{ } | |{ } { } 0 z K z C z M = + +
- For negligible damping:([C] = [0]) and equations of motion are:
| |{ } | |{ } { } 0 z K z M = +
- Assuming solutions of the form { } { }
st
e A z = leads to a set of homogeneous
equations: | | | | ( ){ } { } 0 A K s M
2
= +
- The non-trivial solution can be written in the form of a characteristic equation (or
frequency determinant): | | | | 0 K s M
2
= + or eigenvalue equation:
| |{ } | |{ } u K u M =
- This leads to a set of real roots, typically
2
i i
2
i
s e = = , where
i
is the i-th
eigenvalue and e
i
the i-th natural frequency.
- For each eigenvalue
i
, there is an eigenvector {u}
i
,which relates the relative
amplitudes at each of the degrees of freedom.
- Vibration in the i-th mode then can be described by: { } { } ( )
i i i i
t sin A u z
i
o + e =
- The general free vibration of the systems is a combination of all modes:
{ } { } | | ( ) { } t q u z z
n
1 i
i
= =

=
, where: | | { } { } { } | |
n 2 1
u ...... u u u = is called the modal
matrix and { }
( )
( )
q t
A t
A t
n n n
( )
sin
sin
=
+
+

1 1 1
e o
e o
is a vector of modal (principal)
coordinates.
- [u] represents a linear transformation between generalised coordinates {z} and
modal coordinates {q}
- For a lightly damped system ( 0 C = ), the frequency determinant is:
| | | | | | 0 K s C s M
2
= + +
- This equation has a set of complex conjugate roots having negative real parts -
provides information about the frequency and damping associated with each mode
of vibration
- For harmonic excitation of damped MDOF system
a) The system vibrates (in its steady state) at the same frequency as the excitation
b) The dynamic displacement at each DOF lags behind the excitation.
c) The displacement amplitudes at each of the degrees of freedom are dependent
on the frequency of excitation
1-86
- In MDOF systems the excitation can be applied simultaneously at any of the
DOFs.
- For a linear system the response at any of the DOFs is the sum of the responses
due to each excitation force.
- When the excitation is applied simultaneously at all DOFs, F(t) can be written as:
{F(t)} = {F}f(t) = {F}sin(et), then| | { } | | { } | | { } { } ) t ( f F z K z C z M = + +
- Take Laplace transforms of both sides with zero initial conditions, replace s with
ie [32] and pre-multiply both sides by | | | | ( )
1
2
K C i M

+ e + e gives:
( ) { } ( ) | |{ } F H H
x
e = e
- {H
z
(e)} is a vector of frequency responses at the DOFs
- ( ) | |
(
(
(
(

= e
nn 1 n
n 2 22 21
n 1 12 11
H H
H H H
H H H
H

is a matrix of frequency response functions such

that H
ij
= frequency response at i due to unit amplitude excitation at j.
- The frequency response at DOF i is:
( )
j
n
1 j
ij i in 2 2 i 1 1 i
i
z
F H F H F H F H H

=
= + + = e
- The response at a particular DOF is made up of contributions from excitations at
the various DOFs in the systems.
- FRFs and hence responses are complex (when damping is included) amplitude
and phase at a given DOF is given by the modulus and argument of the complex
response
Appendix A2: The Basic Handling Model and its Dynamics
Steady state analysis of a basic vehicle-handling model leads to some important
conclusions. First of all consider the following modelling assumptions:
1. The vehicle moves on a flat horizontal plane
2. The forward velocity is constant
3. There is no suspension compliance (rigid suspension, no body roll)
4. No compliance in the steering mechanism
5. Input to the model is a small steering angle o
f
at the front axle
6. No wheel camber
7. No lateral load transfer
8. No aerodynamic forces
The model is shown in Figure A2.1. The origin of the moving axes is at A, the vehicle
c.g; and the following notation is used:
u = forward velocity (= U, the constant forward speed)
v = lateral velocity
r = yaw velocity
m = vehicle mass
1-87
I = moment of inertia about the z-axis
a,b = distance from A to front and rear axles
Figure A2.1 Basic handling model
Lateral forces at the front and rear axles can be lumped together to give a bicycle
model Figure A2.2. Longitudinal tyre forces are negligible for constant forward
velocity, so two equations of motion are sufficient to describe the dynamics of the
vehicle. In the lateral direction:
yr yf
F F
R
U
v m + =
|
|
.
|

\
|
+
2

where R is the radius of the curve. Noting that Rr U = (linear velocity, U, equals the
product of the radius, R, and the angular velocity, r):
( ) 1 . 2 A ........ ..........
yr yf
F F Ur v m + = +
2 . 2 A ...... .......... .......... F b F a r I
yr yf
=
Figure A2.2 Bicycle model
Ground-fixed
axis system G
X
Y
o
f
x
y
u
v
r
Vehicle-fixed
axis system A
Steering angle
x
y
U
v
r
X
Y
G
A
F
yf
F
xf
F
yr
F
xr
a
b
1-88
Equations for the tyre forces can be derived from the definition of slip angle and tyre
characteristics. Slip angle o is given by:
3 . 2 A .. ..........
u
v
tan
w
w
= o
Relevant tyre characteristics are shown in Figure A2.3 where lateral tyre force F
y
is
plotted as a function of o, for a range of vertical loads F
z
, Cornering stiffness C is the
slope of the appropriate curve at the origin.
Figure A2.3 Lateral tyre force as a function of slip angle
For small slip angles (o < 6)
4 . 2 A ........ .......... .......... C F
y
o =
The negative sign is required to make C positive consistent with the sign convention
used in the analysis. C is the cornering stiffness for the whole axle and depends on the
vertical load. F
y
is assumed to act in the y direction (based on the assumption that o is
small).
It can be shown that the slip angles at the front and rear axles o
f
and o
r
are:
5 . 2 A ... .......... ..........
U
r a v
f f
o
+
= o
6 . 2 A . .......... .......... ..........
U
r b v
r

= o
Increasing
vertical tyre load
F
z
Lateral tyre
force F
y
, N
0 1 2 3 4 5 6
0
1000
2000
3000
4000
Cornering
stiffness
C, is slope
at origin
Slip angle, o, deg
1-89
Lateral forces F
yf
and F
yr
are then:
7 . 2 A .... ..........
U
r a v
C F
f f yf
|
.
|

\
|
o
+
=
8 . 2 A .... .......... ..........
U
r b v
C F
r yr
|
.
|

\
|
=
C
f
and C
r
are cornering stiffness at front and rear axles. F
yf
and F
yr
can be substituted
into equations A2.1 and A2.2. After some manipulation the following equations
result.
These two first order differential equations can be written in matrix form:
With constant steering angle
f
and forward speed U, and considering the steady state
response under these conditions, v, r are constant hence, 0 r v = = , hence:
f
f
f
r f r f
r f r f
aC
C
r
v
U
C b C a
U
bC aC
U
b C a C
mU
U
C C
o
(
(

=
(

(
(
(
(

+
+
2 2
Solving for r, which is denoted as r
ss
to represent the steady state yaw rate, and
representing the terms of the 2x2 matrix on the left hand side as a
11
, a
12
etc:
(
(

=
(

f f
f f
aC
C
r
v
a a
a a
o
o
22 21
12 11
(
(

=
(

f f
f f
aC
C
a a
a a
a a a a r
v
o
o
11 21
12 22
12 21 22 11
1
10 . 2 A ........ C r
U
C b C a
U m v
U
C C
v m
f f
r f r f
o +
(

+ |
.
|

\
| +
=
9 . 2 A ....... C a r
U
C b C a
v
U
C b C a
r I
f f
r
2
f
2
r f
o +
|
|
.
|

\
|
+
|
.
|

\
|
=
11 . 2 A ...
C a
C
r
v
U
C b C a
U
C b C a
U
C b C a
U m
U
C C
r
v
I 0
0 m
f
f
f
r
2
f
2
r f
r f r f
o
)
`

=
)
`

(
(
(
(

+
+
+
)
`

1-90
Hence r maybe found as:
| |
f f f f
aC a C a
a a a a
r o o
11 21
12 21 22 11
1
+

= ,
Which simplifies to:
( )
f r r f
r f
aC bC mU L C C
L C C U
r
+
=
2 2
Considering that r is the yaw rate (an angular velocity) and in this steady state
condition, r = r
ss
,
ss ss
U
R
U
r = =
where R is the radius of the corner, and
ss
is the curvature.
Hence:
( )
f r r f
r f
f
ss
aC bC mU C C L
L C C
+
=
2 2
o

2
1
KU L +
=
with
( )
L C C
aC bC m
K
r f
f r

margin stability