INSTITUT FÜR

ika KRAFTFAHRWESEN
AACHEN
Univ.-Prof. Dr.-Ing. Henning Wallentowitz

Henning Wallentowitz

Automotive Engineering II
Vertical Vehicle Dynamics
Suspension Systems
Driving Behaviour
Editor
Prof. Dr.-Ing. Henning Wallentowitz
Institut Für Kraftfahrwesen Aachen (ika) RWTH Aachen
Steinbachstraße 7, D-52074 Aachen - Germany
Telephone (0241) 80-25 600 Fax (0241) 80 22-147
e-mail [email protected]
internet htto://www.ika.rwth-aachen.de

Editorial Staff
Dipl.-Ing. Florian Fuhr Dipl.-Ing. Ingo Albers
Telephone (0241) 80-25 646, 80-25 612

4th Edition, Aachen, February 2004

Printed by
Vervielfältigungsstelle der Hochschule

Reproduction, photocopying and

electronic processing or translation is prohibited

c ika 5zb0498.cdr-pdf
Contents 1

Contents

1 Vertical Dynamics (Suspension) ......................................................................................3

1.1 Suspension - Demands and Possibilities of Implementation......................................3

1.2 The Road as the Source of Excitation........................................................................5

1.2.1 Spectral Density of the Road Unevenness .............................................................8
1.2.2 Data Acquisition for Road Unevenness ................................................................11

1.3 Components of the total Suspension System ..........................................................15

1.3.1 Tires......................................................................................................................15
1.3.2 Body Springs ........................................................................................................21
1.3.2.1 Leaf Springs ..................................................................................................21
1.3.2.2 Torsion-bar Springs .......................................................................................27
1.3.2.3 Coil Springs ...................................................................................................29
1.3.2.4 Gas Springs...................................................................................................33
1.3.3 Vibration Dampers ................................................................................................44
1.3.4 Seats.....................................................................................................................56
1.3.5 Evaluation of Oscillations by Humans ..................................................................57

1.4 Single Wheel Suspension Model..............................................................................63

1.4.1 Single Mass Model ...............................................................................................65
1.4.2 Dual-Mass Equivalent System..............................................................................67
1.4.2.1 Parametric Study - Automobile Suspension ..................................................72
1.4.2.2 Parametric Study of a Truck Suspension ......................................................85
1.4.3 Enhancement of the Model by Seat Suspension..................................................87

1.5 Single-Track Suspension Model...............................................................................89

1.5.1 Double-axle Vehicle with bending resistant Structure ..........................................89
1.5.1.1 Excitation by real unevenness Routing .........................................................91
1.5.1.2 Mass and Spring Coupling ............................................................................93
1.5.2 Two axle Vehicle with additional Degrees of Freedom.........................................97

1.6 Two-Track Suspension Model................................................................................101

1.6.1 Roll Springing .....................................................................................................101
2 Contents

1.6.1.1 Stabilizer and Compensating Spring ...........................................................103

1.6.1.2 Vehicle Conception and Suspension Characteristics ..................................104
1.6.2 Distortion of the structure (twisting) ....................................................................106
1.6.3 Rigid axle Tramp.................................................................................................108

1.7 Suspension Investigation Methods.........................................................................110

Suspension - Demands and Possibilities of Implementation 3

1 Vertical Dynamics (Suspension)

1.1 Suspension - Demands and Possibilities of Implementation

The roads commonly used by motor vehicles are uneven. This unevenness causes vertical
movements of the vehicle and the passengers during the driving process.

The vehicle is connected to the road by the tire. Small unevenness in comparison to the tire
contact patch size can be compensated by the tire elasticity, whereas larger unevenness
entails a vertical acceleration or deflection of the wheels. In order not to transfer these
accelerations into the vehicle structure, a length compensating element has to be placed
between the wheel and the vehicle structure.

Steel springs are the technologically most simple elements with variable length. Due to this
fact it is also the most common length compensating element, whose force is a function of
the length variation. It is usually used in the suspensions of motor vehicles. Different parts
connected with springs generate oscillating systems. So there has to be added an energy
absorbing element, the damper.

The suspension’s job in the motor vehicle is to reduce these vertical movements. The
essential criteria specifying the quality of a suspension can be listed as follows:

• suspension comfort for the passengers (Effective acceleration affecting the

passengers)

• forces affecting the load (Effective value of structure acceleration)

• wheel load variation (Effective value of the dynamic wheel load), which influences the
grip between tires and road (driving safety) and the load application upon the road
surface.

The further demands on the suspension in a motor vehicle are various and partially
contradicting (Fig. 1.1-1)
4 Suspension - Demands and Possibilities of Implementation

vehicle level and vibrational

little body accelerations behaviour independent from load
..
z, z B
little pitch- and m
rolling motions y load
υ

F wheel 3 Sr
ϕ
F wheel 4
x
spring travel limited
F wheel 2 ∆ zw by required space
wheel load distribution according
little variation in F wheel 1 to good driving behaviour
wheel load

Fig 1.1-1: Demands on a vehicle suspension

Before dealing with the technical details of the spring and absorber elements, first the road
and the mathematical description of its unevenness are presented.
The Road as the Source of Excitation 5

1.2 The Road as the Source of Excitation

The unevenness of the road represents the most intensive source of excitation for the
vibratory system of the motor vehicle in the frequency range up to approximately 30 Hz. The
road`s unevenness causes vertical movements of the vehicle structure, and as a
consequence, the road is affected by tire load variation itself.

Generally road unevenness appears as an excitation with different amplitudes and

wavelengths at irregular periods of time. This is called a stochastic excitaion of the vehicle. In
order to be able to examine the effect of the road unevenness on the vibratory motor vehicle
system (see chapter 1.4), this unevenness first has to be described mathematically.

Concidering the most simple case of a harmonic (sinusoidal) radiation, where the road
unevenness excites an amplitude ‘h’ at equal distances L, you will get a characteristic
unevenness as a result (Fig. 1.2-1).

Fig. 1.2-1: Sinusoidal pattern of unevenness

The unevenness height can be described as follows:

h(x ) = ĥ ⋅ sin(Ω ⋅ x + ε ) (1.2-1)

2Π
including: Ω = as track-dependent circular frequency and ε as phase shift.
L

When driving on this roadway with constant velocity v this distance-dependent unevenness
can be changed into a time-dependent relation:
6 The Road as the Source of Excitation

h(t ) = ĥ ⋅ sin(ω ⋅ t + ε ) (1.2-2)

with: ω as time-dependent circular frequency.

The equality of h(x) and h(t) entails ω·t = Ω·x , and with the relationship x = v·t the time-
dependent circular frequency follows:

v
ω = v ⋅ Ω = 2π ⋅ (1.2-3)
L

The next step in the description of the road unevenness is the transition to a non-sinusoidal,
but still periodic, unevenness (Fig. 1.2-2).

Fig. 1.2-2: Periodic pattern of unevenness

This unevenness can be represented as Fourier series as follows:

∞
h(x ) = h0 + ∑ ĥk ⋅ sin(Ω ⋅ x + εk ) (1.2-4)
k =1

or:

∞
h(t) = h 0 + ∑ ĥ k ⋅ sin(ω ⋅ t + ε k ) (1.2-5)
k =1
The Road as the Source of Excitation 7

with: ĥk Amplitude

εk phase shift;

2Π
Ω= , ω = v·Ω
X

X period length

When the individual amplitudes ĥ k of the Fourier series are plotted versus the frequency, a
discrete amplitude spectrum (line spectrum), belonging to the periodic unevenness, results
(Fig. 1.2-3).

Fig. 1.2-3: Line spectrum of a periodic pattern of unevenness

For the description of real roads there has to be made another step, because they do not
show a periodic type of unevenness, but a variating (stochastic) type of unevenness. The
representation of the Fourier series in a complex equation, results as following:

∞
h(x ) = ∑ ĥk ⋅ e j k Ω x (1.2-6)
k =1

∞
h(t ) = ∑ ĥ k ⋅ e j k ω t (1.2-7)
k =1
8 The Road as the Source of Excitation

x
2
Ω
∫ h( x )⋅e
− ik Ω x
with: ĥk = ⋅ dx (1.2-8)
2π x
−
2

On the assumption that the regarded period length is very large, the distance between the
frequencies in the amplitude spectrum ∆ω becomes very small as a consequence.

In the boundary condition ‘ X → ∞ implies that ∆Ω → 0 and the Fourier transformation

changes into a Fourier integral as:

∞
1
2π −∫∞
h(x) = ĥ(Ω )⋅e j Ω x dΩ (1.2-9)

with the continuous amplitude spectrum

∫ h(x)⋅ e
- jΩ x
ĥ(Ω) = dx (1.2-10)
−∞

∞
1
h(t) = ∫
2π − ∞
ĥ(ω)⋅e j ω t dω (1.2-11)

∞
ˆ ω) = 1 ˆ
h( ∫
−∞
h(t) ⋅ e - j ω t dt =
v
⋅ h(Ω ) (1.2-12)

1.2.1 Spectral Density of the Road Unevenness

For theoretical investigations of the vehicle oscillations caused by road unevenness, the
knowledge of the unevenness as a function of time or the distance travelled is usually not
very important. It is much more interesting to find out which kinds of excitation appear
statistically during the process of travel on a roadway with a characteristic unevenness. That
means which kinds of amplitudes and which kinds of frequencies are excited by road
unevenness at certain fixed distances.

The resulting square average value is defined as follows:

T
2 1 2
T ∫0
x (t) = x (t) dt (1.2-13)
The Road as the Source of Excitation 9

The characteristic line of unevenness results as follows:

1
x ∞ ĥ ( Ω )
h 2 ( x ) = ∫ h2 ( x ) dx = ∫ lim dΩ (1.2-14)
x0 0
x →∞ x

or:

2
x
1 2
∞ ĥ ( ω)
T ∫0
h2 ( t ) = h ( t ) dt = ∫ lim dω (1.2-15)
0
T →∞ T

The limit values occurring here indicate that these simple expressions give a result only for
very large time intervals T or distances X. The expressions

2
ĥ(Ω)
Φ h (Ω ) := lim and (1.2-16)
X →∞ X

2
ĥ(ω)
Φ h (ω) := lim (1.2-17)
T →∞ T

are called circular path frequency or periodical spectral power density (power density
spectrum).

With X = v ⋅ T (see above) similarly the following connection results for the unevenness
spectra:

1
Φ h (ω) = ⋅ Φ h (Ω) (1.2-18)
v

If one measures the power density spectra Φ h (Ω) of different roads and applies this on a
double-logarithmic scale, then similar characteristic curves result for all types of roads (Fig.
1.2-4).

In this presentation the power density spectra can be approximated by straight lines, which
then can be described by the following equation:

−w
 Ω 
Φ h (Ω) = Φ h (Ω 0 ) ⋅   (1.2-19)
 Ω0 
10 The Road as the Source of Excitation

Fig. 1.2-4: Power spectral density of unevenness dependent on circular course frequency.

Here Φ h (Ω 0 ) represents the power spectral density dependent on Ω 0 (reference circular

path frequency), which is usually selected as Ω0 = 10 -2 cm -1 = 1 m-1 . This corresponds to a

reference wavelength of L0 = 2π Ω0 = 6.28 m . Φ h (Ω 0 ) is also called the degree of
unevenness of the road.

W defines the gradient of the straight line and is also called undulation (in German:
“Welligkeit”). The undulation of the roadway varies depending on the road design between
1.7 and 3.3 (standardized road: w = 2) /7/.

The degree of unevenness (and undulation) are considered as assessment criteria for the
condition of a road. An increase of Φh(Ω 0 ) corresponds to a larger unevenness of the
roadway, while an increasing value of ‘W’ means a higher proportion of long waves in the
The Road as the Source of Excitation 11

spectrum. The relationship between the curve of the spectral density and an increase of the
degree of unevenness Φ h (Ω 0 ) or undulations ω is shown in Fig 1.2-5.

Fig. 1.2-5: Power spectral density dependent on the circular path frequency

For road unevenness of wavelengths L between 0.15 m to 80 m there result excitation

frequencies Ω from 0.06 Hz up to 267 Hz (see eq. 1.2-3) in the case of driving speeds v
between 5 m/s ( ≈ 20 km/h) and 40 m/s ( ≈ 160 km/h).

1.2.2 Data Acquisition for Road Unevenness

For the determination of the spectral power density of the road unevenness, one proceeds
from the longitudinal section of the road. For the measurement of road unevenness different
measuring procedures were developed, which are adapted partially to both road construction
as well as for the application in the Automotive technique. About 70 different unevenness
measuring procedures are available.

An exact measurement of the road surface is possible only with subgrade level and levelling
instrument. Because of the high demand on time in the measurement of longer distances
and the disturbance by the flow of traffic etc. these procedures are less important in technical
application. Those so-called geometrical unevenness measuring instruments were
developed mainly for application in road construction.

The measuring procedures used in automotive technology operate according to the principle
of the dynamic unevenness measurement. During these measuring procedures the reference
level is not fixed, but derived continuously from an average value of the road unevenness,
formed over a sufficiently long period of time. Fig 1.2-6 shows a test vehicle equipped with
devices for the measurement of road unevenness.
12 The Road as the Source of Excitation

Fig. 1.2-6: Instrumentation for unevenness measurements (ika)

The measuring job is divided into two sub-problems:

1) Measurement of the change in distance ∆z between roadway surface and points of

reference at the vehicle bodywork

2) Measurement of the vertical structure movement zA at the points of reference

The change in distance ∆z is entered continuously for each track with the help of a
contactless laser measuring system. The measuring system is based on the triangulation
principle, Fig. 1.2-7.
The Road as the Source of Excitation 13

Fig. 1.2-7: Operational principle of the contactless distance measurement

With the help of a laser beam, incident at a certain angle on the road surface, a light spot is
produced, which is received through a lens on the surface of a light-sensitive photodiode.
According to the angle of incidence of the laser beam, the spot of light on the surface of the
photodiode shifts as illustrated, and its movement is proportional to the vertical distance
between the point of reference and the road surface. Thereby it can be converted into a
measuring signal.

The measurement of the displacement of the vehicle body zB is managed by acceleration

adaptors fastened to the vehicle bodywork perpendicularly over the points of reference for
the distance measurement.

The low-pass effect of the vehicle suspension system results in accelerations &z&B of max. 1 to
2 g, and so sensitive acceleration sensors can be selected accordingly, which record low-
frequency structure accelerations with sufficient accuracy.

The displacement of the vehicle body zB is determined continuously by the double integration
of the vehicle body acceleration &z&B . The characteristic of unevenness h(t) or h(x) results
from the addition of the displacement of the vehicle body zA and the change in distance ∆ z.

The procedure to define the determination of the unevenness spectrum Φ h (Ω)

metrologically from a recorded measuring signal h(t) is schematically represented Fig 1.2-8
/38/.
14 The Road as the Source of Excitation

Fig. 1.2-8: Metrological determination of an unevenness spectrum

From the measured signal which can be analyzed with a bandpass filter with mean frequency
ω0 a frequency range  ∆ω is filtered. Then the band-pass filter signal generated is squared
and the average value is calculated over a certain period of time. From this average value,
defined in Fig 1.2-8 as h 2 B , the value of the power density spectrum for ω0 is determined by
division with the filter bandwidth ∆ω.

By shifting the bandpass filter center frequency the power density spectrum for the entire
frequency range of interest can be determined.

From the time-frequency-dependent power density spectrum the spectral unevenness

density can be calculated as a function of the circular path frequency Ω by simple
multiplication with the driving speed v (see Eq. 1.2-18). In this case it is assumed that the
measurement of the characteristic of unevenness h(t) was determined at constant driving
speed v.
Components of the total Suspension System 15

1.3 Components of the total Suspension System

1.3.1 Tires

In the following section, the vehicle tire is considered only from the technical point of view of
the suspension and damping. Further representation of the characteristics of vehicle tires
can be found in chapter 2.2. The transverse characteristics of the tire are described there,
e.g. generation and transfer of side forces.

• Tire as a spring

The wheel load is transferred to the tire by the rim and pressure is applied to this contact
area. The reaction force, which opposes the load on the tire, consists of different
components. Fig. 1.3-1 shows the proportions of the load carrying components of a tire as a
function of compression.

Fig. 1.3-1: Structure of the spring characteristic of a pneumatic tire/41/

The component I corresponds to the proportion of the load carrying capacity of the solid
rubber fabric body, due to elastic deformation. The component II represents the so-called
16 Components of the total Suspension System

circular retaining force of the compressed air, which reinforces the tire in its walls (balloon
effect). The component III represents the very small proportion of the air compression.

The component IV represents the principal part, which is based on the adaption of the road
contact area A to the vertical load Gw:

Gw ≈ pi ⋅ A

(with pi inflation pressure, that means the pressure in the wheel road contact area A
corresponds to the inflation pressure.)

Due to the tire compression an oscillatory subsystem develops in the wheel or in the
unsprung masses. In this way, the spring characteristic of a tire influences natural frequency
of the wheel or axle. Also the inflation pressure has decisive influence on the suspension
system beside the structure of the tire.

Fig. 1.3-2: Spring characteristic of a Truck tire with four different air pressures

The spring characteristics indicate a linear characteristic in the operating range. By the
modification of the vertical load FW as function of the compression zW (lowering of the wheel
axle) the spring rate cR can be determined.
Components of the total Suspension System 17

dFW
cW = (1.3-1)
dz W

Since the force characteristic is not a constant function of compression, the spring rate is
variable. Therefore a constant spring rate can be considered only as an approximation for a
certain area of the curve , where the static wheel load and the compression is appropriate to
Fig. 1.3-3.

Fig. 1.3-3: Linearisation of the tire spring rates.

In the operation range, the spring rate is given by

FWstat
cW = (1.3-2)
z 'Wstat

If the tire compression is increased or reduced over zdyn, the force approximation changes
around

Fdyn = cw ⋅ zdyn (1.3-3)

As measure for driving safety and protection of the road, the ratio of total load to the static
load is indicated as:

F FWstat + Fdyn
= (1.3-4)
FWstat FW stat
18 Components of the total Suspension System

With the equation (1.3-2)

F cW
=1+ ⋅ zdyn (1.3-5)
FWstat FWstat

FWstat
The quotient is according to Fig. 1.3-3 equal to the sub-tangent with the length z ' Wstat
cW

z 'Rstat = FWstat c W (1.3-6)

This parameter is a measure for the additional load relative to the static load with a
compression zdyn, e.g as a result of unevenness. To all tires (passenger cars and trucks)
applies approximated:

z 'W stat = 1.7 ... 2.9 cm

The lower value refers to diagonal tires and the upper value to radial tires, that means the
latter are softer and with a certain tire compression zdyn, induce smaller additional loads Fdyn.
With this reference value, the approximate wheel natural frequency can be determined:

cW FW stat
ωeW = ≈ (1.3-7)
mW z ′W stat ⋅ m W

The wheel natural frequency is usually approximately feW = 8 -14 Hz. This estimation
proceeds from a static spring characteristic of the tire. In driving condition, the radial spring
rigidity of the tire is further influenced by a multiplicity of parameters. Due to the viscoelastic
behavior of the rubber, the spring rigidity increases e.g. with increasing excitation frequency
when excited by road unevenness (chapter 1.4).

Besides, the respective operating point of the tire also influences the spring rate. While e.g.
with increasing driving speed the influence of the centrifugal force at the tire circumference
causes an increase of the spring rigidity, the side forces generated at the tire due to the side
slip angles (Chapter 2.2.4) lead to a lateral displacement of the tire road contact area. This
causes a more uneven load of the tire shoulders and thus a reduction of the spring rate.
Components of the total Suspension System 19

Fig. 1.3-4: Radial spring stiffness of a truck tire as a function of vehicle velocity and side slip
angles/17/

• Absorption capacity of the tire

Besides the suspension the absorption capacity of the tire becomes a special meaning, since
it is in direct relation to the tire comfort. The term absorption capacity characterizes thereby
the characteristic of the tire to absorb unevenness, which are short in relation to the tire
contact length, without vertical movement of the axle/9/. The tire behaves in relation to such
obstacle more softly than in relation to an equivalent high compression of the entiretire
contact patch. The absorption capacity is influenced mainly by the rigidity of the sidewall.

• Tire damping

The tire damping is a material damping. It is a cause for the temperatur rise of the tire and
increases with the number of plies and the material mass in the tire. The damping. factor kW
is thereby no constant, it depends on several influence parameters like the radical spring
rigidity of the tire, Fig. 1.3-5.
20 Components of the total Suspension System

Fig. 1.3-5: Damping factor kw of a tire as a function of rolling speed and excitation frequency

The influence of the excitation frequency by unevenness can be described after experimental
testing by the following absorption rule/20.

λ
kw = (1.3-8)
ωe

ω circular frequency of the excitation

The constant λ and the exponent e do not depend on the tire design only, but particularly
from the rolling speed v.

With high rolling speeds e is ≈ 0, i.e: kw depends only on λ or v and is independent thereby
of ω (see Fig. 1.3-5).

For estimation of the order of the tire damping rate DW the following approximation formula
for the damping factor factor is suitable.

0,1⋅ c R
kw ≈ (1.3-9)
ω

The tire absorption is important particularly in the resonance area ( ω = ω eW ). With the
approximation formula (Gl.1.3-8):
Components of the total Suspension System 21

2
kW 0,1⋅ c W 0,1⋅ ω eW
D R res = = = 2
= 0,05 (1.3-10)
2 ⋅ m W ⋅ ω eW 2 ⋅ m W ⋅ ω eW ⋅ ω eW 2 ⋅ ω eW

A damping rate of DWres of 0,05 is not sufficient for the damping of the wheel masses. For a
satisfying absorption of the wheel oscillations a damping rate of DW of approximately 0.4 is
required, that means that the hydraulic vibration damper-for the body (see chapter 1.3.3)
must take over the main part of the wheel damping at the same time.

1.3.2 Body Springs

Those sections of wheel suspension systems, which give restoring force under elastic
deformation, are described as body springs in this chapter. Besides conventional coil, leaf
and torsion bar springs gas springs can also be mentioned.

1.3.2.1 Leaf Springs

Leaf springs represent the classic leaf spring form, which were already used with carriages.

The substantial advantage of leaf springs in relation to other spring design concepts consists
of the fact that they cannot be used as spring elements only, but as design elements for the
coupling of body and axle (in particular for the guidance of the axle) as well. Multilayer leaf
springs possess damping characteristics due to frictional forces between the leaves.

In connection with rigid axles, conventional leaf springs nowadays are used for a few car
types (Station wagon and offroad vehicles) only. Front and rear suspension systems of
trucks are usually designed like this. Fig. 1.3-6.

Fig. 1.3-6: Leaf spring guided front axle of a truck (Daimler-Benz)

22 Components of the total Suspension System

Without any special measures the spring characteristic of the leaf spring is linear. The
relationship between the deflection and the load at the end of a clamped bending beam is
known from the strength theory.

F ⋅ l3
f= (1.3-11)
3 ⋅E ⋅ J

With l: length of the bending beam,

E : elastic modulus

J : moment of inertia

b ⋅ s3
J= (1.3-12)
12

Fig. 1.3-7: Deflection of a leaf spring

Thus, one receives following spring rate for one half of a leaf spring system corresponding to
Fig 1.3-7:

F 1 b ⋅ s3
c= = ⋅ 3 ⋅ E ≈ const (1.3-13)
f 4 l
Components of the total Suspension System 23

In order to ensure the guidance characteristics of the axle leaf spring, the width b and the
thickness s can be varied for the definition of the leaf spring rigidity c only within certain
limits. Usually relatively large spring lenghts are necessary in practice. Truck leaf springs are
up to 1800 mm long.

Since the bending force decreases accordingly from the spring center to the ends, one
arranges springs from different length in layers in such a way that a high and even degree of
utilization of material is achieved. The bending stress corresponds to the trapezoidal one leaf
spring Fig. 1.3-8. So this kind of leaf spring is called the trapezoid spring.

Fig. 1.3-8 trapezoidal spring

Due to the linear spring characteristic two main disadvantages of the leaf spring result:

• Spring travel

In the application of high loads the available spring travel can often reduce ∆zReserve to no
longer sufficient values (recommended values for passenger car: ∆zReserve 50mm, truck:
∆zReserve 70mm)

• Natural frequency ≠ const.

Since the spring rigidity does not change with loading, the body natural frequency in the
unloaded condition is larger than the natural frequency in the loaded condition because of
m B,loaded > m B,unloaded . The acceleration of the body mass is described in chapter 1.4.

These disadvantages of multi layer leaf springs can be reduced by a progressive behavior of
the leaf spring characteristic. This can be achieved by the different pre-load of the individual
leaves and/or by the parallel addition of auxiliary springs, which come into use only when the
load exceeds a certain limit, Fig. 1.3-9.
24 Components of the total Suspension System

Fig. 1.3-9 Leaf springs with auxiliary springs

The natural frequency remains constant, if the upward gradient of the spring characteristic
changes proportionally to the load:

c loaded mBloaded
= (1.3-14)
c empty mBempty

The difference in spring travel between unloaded and loaded condition decreases in case of
such a progressive spring characteristic. So the entire spring travel can be reduced, which is
important for the space requirement in the wheel housings, the entrance height of buses etc.,
Fig. 1.3-10.
Components of the total Suspension System 25

Fig. 1.3-10: Spring rate of a leaf spring with additional springs

Another disadvantage of the leaf spring consists of the fact, that both in the suspension and
between the leaves of the multi layer leaf spring, dry friction occurs (Fig. 1.3-11). This affects
the suspension characteristics unfavourably.

Fig. 1.3-11: Spring characteristics of a leaf spring with dry friction (rear axle of a commercial
van)
26 Components of the total Suspension System

The disadvantages can be moderated by the use of shackles instead of sliding shoes in the
suspension or by the application of plastic layers between the spring leaves.

A small number of leaves in the spring influences the decrease of the frictional forces
positively. This is possible, not by the use of additional spring leaves but by the use of
stronger leaves which can withstand a higher bending force in the spring centre. This leads
to a higher degree of utilisation of material. The bending stress is constant along the length
of the leaf, if the spring halves indicate a parabolic profile in the longitudinal section. Leaf
springs with such leaves are called parabolic springs. The development of the conventional
laminated leaf spring to the parabolic spring is illustrated in Fig. 1.3-12:

conventional
trapezoidal spring,
m = 128 kg

improved trapezoidal
spring, m = 94 kg

parabolic spring,
m = 61 kg

Fig. 1.3-12: Parabolic spring

These three springs have the same length, spring rate and nominal load. The advantage of
the parabolic spring, apart from the decrease of friction, is the decrease of the spring weight
by around 50% (minimization of the unsprung masses). The main advantage of the leaf
spring is the combination of suspension and axle location, which is partially lost with the
transition to parabolic springs with few leaves. During braking, the parabolic springs tend to
be affected by the S-impact, which must be prevented by an additional torque support.
Otherwise, changes in the running speed between the roadway and the tire can result in
brake-hopping (Fig. 1.3-13).
Components of the total Suspension System 27

Fig. 1.3-13: Leaf spring with torque support

1.3.2.2 Torsion-bar Springs

Torsion-bar springs are mainly used for the suspension of passenger cars and commercial
vans. Torsion-bar springs are rods made of spring steel, that are predominantly stressed by
torsion. The shafts are clamped at one end and free to twist at the other, so that the shaft
can be twisted flexibly by a moment acting in the direction of the axle.

For the use of torsion-bar springs as vehicle suspension elements, the elastic twist of the bar
is converted into a reciprocating motion with the help of a crank, which is located to the
swivelling free end and which applies the twisting moment, Fig. 1.3-14.

Fig. 1.3-14: Crank mechanism of a torsion bar

28 Components of the total Suspension System

The strut of the axle or the suspension builds up the crank arm. The torsion-bar springs are
arranged normally along the bearing axle facing the body of the strut, at the opposite end of
which the vertical wheel force applies at as the outside load.

Between the twisting angle ϕ and the torsional moment Mt of a torsion bar with circular cross-
section, the following context exists:

Mt ⋅ l
ϕ= (1.3-15)
G ⋅ Jp

with G: shear modulus

Jp: polar surface moment of inertia

L: length of torsion bar

The polar moment of inertia of a full bar with a circular cross-section applies:

π ⋅ d4
Jp = (d : staff diameter) (1.3-16)
32

So the torsional stiffness ctor of such a bar follows :

G π ⋅ d4
c tor = ⋅ (1.3-17)
l 32

Related to the crank end the spring stiffness follows approximately:

∆z ≈ r ⋅ ∆ϕ

1
∆F ≈ ⋅ ∆M t
r

∆F 1
c= = ⋅ c tor (1.3-18)
∆z r 2

While the spring characteristic of a torsion bar is linear, the effective spring stiffness at the
crank end depends on the kinematics of the strut arrangement. In the equation 1.3-18
corresponding trigonometric relations have to be used for exact calculations.

At the surface of a torsion bar with circular cross section the shear stress amounts to:

16 ⋅ M t
τt = (1.3-19)
π ⋅ d3
Components of the total Suspension System 29

The torsion bar diameter d cannot be selected freely for the achievement of a suitable
torsional rigidity ctor. A minimum diameter has to be specified as a function of the expected
maximum twisting moment and the admissible shear stress.

In practice, to achieve a suitable spring stiffness, relatively large torsion bar lengths are
necessary.

Torsion bar springs are combined both with transverse and longitudinal control arms and in
some cases even with semi-trailing arms. They are arranged predominantly parallel to the
vehicle floor.

For front wheel suspensions torsion-bar springs are used predominantly in combination with
wishbones and arranged in the vehicle’s longitudinal direction, Fig. 1.3-15. With the axle
shown in the picture a vertical adjustment is intended at the end of the bars.

Fig. 1.3-15: Torsion-bar suspension (VW T4)

1.3.2.3 Coil Springs

The most common spring used in passenger cars is the coil spring. It can be imagined as a
wound torsion bar. The spring material is stressed predominantly by torsion.

If you consider D/2 as the lever arm of the spring force F, which affects the wound torsion
bar, its value follows as half of the diameter of the coil spring (D/2). The spring stiffness is
shown in equaition 1.3-18:

4
c≈ ⋅ c tor (1.3-20)
D2

G π ⋅ d4
with c tor = ⋅
l 32
30 Components of the total Suspension System

The total length l of the wound torsion bar follows approximately as:

l ≈ i ⋅ π ⋅D (1.3-21)

with: i number of turns

So you receive the spring stiffness of a cylindrical coil spring with circular wire cross section
G ⋅ d4
from Eq.1.3-20 and Eq.1.3-21: c ≈ (1.3-22)
i ⋅ 8 ⋅ D3

The normally linear characteristic of the coil spring can be changed into a progressive
characteristic by the variation of the overall diameter, the wire thickness and the gradient,
see Fig. 1.3-16:

Fig. 1.3-16 Coil springs

The effect is based on the fact that with increasing spring load some of the windings are
forced to contact each other, whereby the effective length l of the wound torsion bar
shortens, see equation1.3-22.

The combination of all three measures results in the so-called mini block spring, shown in
Fig. 1.3-17. Apart from its progressive behavior, the mini-block spring also has the advantage
of an extremely low overall height, since the coils partially lay one into another when they are
stressed by load.
Components of the total Suspension System 31

Fig. 1.3-17: Miniblock spring

Coil springs can essentially compensate higher forces towards their longitudinal axis in
comparison with their transverse axis. As a consequence they are used similar to torsion
bars in combination with the struts of the wheel suspension, which support the force
components not portable by springs, Fig. 1.3-18.

Fig. 1.3-18: LSA rear wheel suspension (Porsche 911 Carrera)

32 Components of the total Suspension System

The spring is connected on the one hand with the body and on the other hand with the strut,
at whose axle-facing end the wheel load acts as an external force.

Depending on the kinematics of the wheel suspension and the arrangement of the springs,
there exists a ratio i for independent suspension, which is represented by the relation
between the variation of the force affecting the body spring ∆f and the displacement of the
tire contact patch ∆zW:

∆f
i= (1.3-23)
∆z W

The spring ratio i is usually smaller than 1 and not constant, but changes depending on the
instantaneous position of the transmission components of the wheel suspension, thus
dependent on the instantaneous condition of compression.

Between the wheel load FW and the spring load FS the following equilibrium exists with the
leverage I

FW
FS = (1.3-24).
i

By that the so-called wheel specific spring stiffness of an independent suspension for a
certain compressed condition can be indicated

dFW d (FF ⋅ i )
c radbezogen = =
dz W dz W

dFF di
= ⋅i + ⋅ FF
dz W dz W

dFF df di
= ⋅ ⋅i + ⋅ FF
df dz W dz W

di
= c ⋅ i2 + ⋅ FF (1.3-25)
dz W

A progressive characteristic of the suspension can also be obtained possibly by appropriate

configuration of the wheel suspension kinetics.
Components of the total Suspension System 33

1.3.2.4 Gas Springs

With the springs regarded so far, the springing medium was solid and the spring absorbed
energy by deformation. With the springs considered in this paragraph, the springing medium
is gaseous and the spring absorbs energy by a variation of volume. Fig. 1.3-19 shows the
structure of an air spring in principle.

Fig. 1.3-19: Piston cylinder gas spring

The theoretical spring length hth is a characteristic dimension which is given by the quotient
of the active volume V (inclusive of additional volume) and the effective surface area A
affected by the gas pressure:

V
h th = (1.3-26)
A

The spring force F is given by

F = (p - pa) ⋅ A (1.3-27)

With: p: gas pressure

pa: external pressure

Considering the gas equation p ⋅ v n = const. with n as the polytropic exponent, the spring
rigidity of an air spring can be calculated from the relation:

dFF
c (f) = (1.3-28)
df
34 Components of the total Suspension System

1 V (f )
c (f ) = A ⋅ n ⋅ p (f ) ⋅ ; h th = (1.3-29)
h th A

The polytropic exponent is situated between n=1 (isothermal, slow spring movement) and
n=1.4 (adiabatic, quick spring movement). Fig. 1.3-20 shows, that with finite hth the spring
action changes for both quasi-static (Fstat) and dynamic (Fdyn) movements.

Fig. 1.3-20: Spring force in relation to the compression of a gas spring

The relation between variation of the spring force and the variation of the displacement
defines the spring rigidity c. As a reference value, a line for hth = ∞ , c = 0 is added. A small
spring rigidity (low natural frequency) requires a large hth, that means a large spring volume.

The oscillation of a mass m on an air spring features the following natural frequency:

c c⋅g
ωe = =
m ( p − pa ) ⋅ A
with equation 1.3-29 follows:

g⋅n⋅p
ωe = (1.3-30)
hth ⋅ ( p − pa )

If one considers a spring ratio i (here simplifying di / dzW = 0), then you receive in accordance
with equation 1.3-24 and equation 1.3-25:

c wheel rel. g⋅n⋅p

ωe = = (1.3-31)
GW / g hth ⋅ ( p − pa )
Components of the total Suspension System 35

With relatively small spring diameters follows: p>>pa. The equations for the spring rigidity and
the natural frequency are simplified to:

g⋅n⋅m
c≈ ; m⋅g ≈ p⋅A
h th
g⋅n
ωe = (1.3-32)
h th

The theoretical piston-cylinder gas spring is used in motor vehicles only in a modified
specification. In practice, gas springs are differentiated between hydro-pneumatic springs
and gas bellows or u-type gas bellows.

• Hydropneumatic spring

Hydropneumatic spring elements are used by Citroen in different car types, Fig. 1.3-21.

Fig. 1.3-21: Hydropneumatic spring (Citroen)

The spring action is transferred by the piston, first to a fluid and then to a gas. Fluid and gas
are separated by an impermeable rubber membrane.

The presence of the fluid permits to a large extent wear-free and low-friction sealing between
piston and cylinder. Additionally there is a possibility of providing the spring elements with an
integrated hydraulic damper and a hydraulic level regulator. The hydro-pneumatic air spring
with level control has the following characteristic:
36 Components of the total Suspension System

Gas weight = const.

In hydro-pneumatic springs, the compression ∆z with the increase of load is balanced only
by pumping of fluid. The theoretical spring length is given by (A ≈ const.):

h th ≈ h th1 − ∆f (1.3-33)

With the change of the operating point of an air spring n=1 has to be fixed. The Gas equation
simplifies to

p1 ⋅ V1 = p ⋅ V with A = const.

p1 ⋅ hth = p ⋅ (hth1 − ∆f ) (1.3-34)

The spring stiffness of a hydro-pneumatic suspension follows from equation (1.3-29):

1
c (f ) ≈ A ⋅ n ⋅ p 2 (f ) ⋅ (1.3-35)
h th1 ⋅ p1

As a result, the spring stiffness increases in this case with the square of the spring load (the
body mass which has to be cushioned):

c load F m
> load = load (1.3-36)
c empty Fempty m empty

By the compression-conditioned decrease of hth the natural frequency of the structure

oscillations rises with the increasing load.

Hydro-pneumatic spring elements are used predominantly at the rear axle of vans, estate
cars and heavy limousines, since they enable an effective level control. Also partially carrying
installations are used in combination with steel springs (parallel connection), Fig. 1.3-22. The
advantage of this combination consists of the fact that setting the natural frequencies of the
two spring types against each other leads finally to a suspension with an almost constant
natural frequency.
Components of the total Suspension System 37

Fig. 1.3-22: Partially carrying spring cylinder with hydraulic level regulation and adjustable
damping (Sachs)

• Gas spring bellows

Fig. 1.3-23 shows the two designs of bellows, the gas spring bellow and the U-type bellow.
The gas spring bellow is composed, similar to a pneumatic tire, of rubber material reinforced
by textile fabric.
38 Components of the total Suspension System

Fig. 1.3-23: Gas spring bellows

The surface area of the pneumatic spring affected by the positive pressure, which entails a
force to the vehicle structure, is called effective surface area Aw. In contrast to the hydro-
pneumatic gas spring mentioned above, the effective surface area Aw of the gas spring
bellow changes proportionally to its travel. In Fig.1.3-23, the effective surface area is
represented by the effective diameter dW for both designs.

The load carrying capacity of the spring is determined from the product of positive pressure
and effective surface area. In the particular case of the U-type bellows the possibility of a
direct influencing control is given, because the effective surface is defined by the outer
contour of the piston, Fig. 1.3-24 and Fig. 1.3-25.
Components of the total Suspension System 39

Fig. 1.3-24: Characteristics of the load capacity of an air spring bellow

Fig. 1.3-25: Characteristics of air springs with u bellows

Besides, the general possibility exists of implementing an additional volume formed by an air
chamber (increase of hth) to realise a less progressive spring rate.

Also gas spring bellows enable a level control of the vehicle bodywork with increasing load.
The spring compression by a static load is balanced by the additional pumping of gas. So the
40 Components of the total Suspension System

theoretical spring length hth remains constant. The feature of a gas spring bellow with
pneumatic regulation is therefore:

Gas volume = const.

Thus, for the spring stiffness of a gas spring bellow applies:

1
c (f ) = A ⋅ n ⋅ p (f ) ⋅ (1.3-37)
h th

In this case, the spring stiffness is directly proportional to the spring load:

c bel F m
= bel = bel (1.3-38)
c leer Fleer m leer

As a consequence, the natural frequency of the body oscillation does not change here with
loading (hth = const. ⇒ ωe = const. ).

Gas springs (in most cases U-type bellows) are mainly used in buses (constant entrance
height) and trucks (high additional load compared to the dead weight). A spring element with
fully carrying pneumatic U-type bellow for a passenger car is shown in Fig. 1.3-26.

Fig. 1.3-26: Gas spring and damper unit with U-type bellow
Components of the total Suspension System 41

Apart from a level control in this case an infinitely variable pneumatic damping control is
implemented (soft characteristic favourable for travelling comfort, stiff characteristic
favourable for driving stability).

The defined outline of the rolling piston influences the spring characteristic by a variation of
the effective surface area AW. Additionally, the auxiliary spring (bumper) causes a highly
progressive behavior of the spring characteristic in compression.

The responsiveness of the pneumatic spring is influenced particularly by the orientation of

the spring bellow cord angle. Drastic improvements result, if the resistance carrier is
arranged axially to the spring direction. However this orientation of cords requires an outside
guidance of the pneumatic spring bellow implemented by the use of a supporting tube, in
order to be able to compensate tangential forces. A final overview about the most important
features of steel springs and pneumatic springs is given in the fig.1.3-27 and 28.

Steel spring

1.) Driving comfort (Fig. 1.3-27)

The driving comfort changes in order to the loading condition. The driving comfort is
measured as the natural frequency and is proportional to c/F and changes with const. spring
stiffness c and variable load F.

2.) Total spring travel (Fig. 1.3-28)

The total spring travel results from the addition of the static and the dynamic compression:

s = sstat + sdyn

A steel spring cannot be dimensioned any soft, since the total spring travel would be too
large. For this, the necessary space is not available in the wheel housing. The driving
comfort however is directly proportional the spring softness.

3.) Level control

Due to static compression, a steel spring suspended vehicle has different body level
positions dependent on the loading status.

4.) Space requirement

Modern design methods of coil springs cause an ever decreasing space requirement. The
overall height is minimized to a small number of coil diameters by the use of spiral coil
springs.
42 Components of the total Suspension System

Air spring

1.) Driving comfort

The travelling comfort remains almost constant, independent of the load. With constant
effective surface area, the natural frequency is proportional p/pÜ

2.) Total spring travel

The total spring travel results as: s tat = s dyn .

The static spring travel’s value is zero, since, depending upon load the air pressure is varied
with constant vehicle height. So the pneumatic suspension can bedimensioned more softly
than the steel spring suspension.

3.) Level control

An air-suspended vehicle has a constant level position due to the loss of the static
compression.

4.) Space requirement

U-Type bellow air springs always require additional components (covers and rolling pistons).
In this case, the length of the rolling piston is proportional to the half of the pneumatic
spring’s travel. The block dimension of the pneumatic spring corresponds to the height of its
components.
Components of the total Suspension System 43

mloaded
relative load =
munloaded

Fig 1.3-27: Natural frequencies of different suspension systems dependent on the additional
load

Fig 1.3-28: Differences between steel and air springs

44 Components of the total Suspension System

1.3.3 Vibration Dampers

Vibration dampers serve both to guarantee driving safety of a vehicle and to optimise the
travelling comfort (Fig. 1.3-29).

Driving safety is highly influenced by the road grip of the wheels. Oscillations of the wheel
masses, which are also called unsprung masses together with the proportionate masses of
the wheel suspension components, have to be minimised by damping, because they are
suspended only by the tire spring but not by the vehicle body suspension(cR >>cA).

A satisfactory driving comfort requires on the one hand, that the amplitudes of the body
oscillations are small, and on the other hand, that the body accelerations which are induced
by the damping forces are as small as possible. This implies a rather weak damping.

Fig. 1.3-29: Functions of the shock absorber

During the selection of a damper therefore a compromise between hard damping for safety
and a soft damping for comfort has to be implemented.

Vibration dampers or shock absorbers differ basically in the type of friction, which causes the
transformation of oscillation energy into heat, Fig. 1.3-30.
Components of the total Suspension System 45

Fig. 1.3-30: Different damper designs

While shock absorbers of vehicles in the 20s and the 30s operated with mechanical friction,
liquid dampers became generally accepted in the course of development, since the
exponential relationship between damping force and the spring rate given renders a better
adaptation of the absorption characteristics to the vibratory vehicle system. It enables an
asymmetrical configuration of rebound and compression stage at relatively small effort and
predominantly, these dampers show a better responsiveness (damper force application at
very low compression speeds).

The first shock absorbers based on liquid friction were the lever absorbers. In these
absorbers oil is pressed through a valve by a piston which is moved by a lever, that is often
applied as a transverse arm. This type of damper is rarely used nowadays. Today, there are
almost exclusively used telescopic shock absorbers.

Apart from the omission of wearing parts such as levers, bearings and operating cams, these
dampers indicate the advantage of the possibility of a more precise adjustments in
absorption in comparison to lever-type dampers, since spring travel and plunger lift mostly
correspond. Due to the large plunger lift the liquid recirculation is bigger. The direct absorber
can therefore operate with a much lower pressure (compared with the lever absorber), which
facilitates the design of the absorber valves and has a favourable effect on its life span.

The telescopic absorbers can be differentiated into mono tube and twin-tube dampers. Fig.
1.3-31 shows the configuration and principle of the two absorber systems:
46 Components of the total Suspension System

Fig. 1.3-31: General assembly of different damper designs

In types of absorbers there is a piston moving in a cylinder filled with liquid while overcoming
the flow resistance, which features throttling elements. The absorbed mechanical labour is
converted into heat and transferred to the environment over the absorber’s external tube. If
the absorber surface is not large enough for the heat dissipation, the surface must be
enlarged. Otherwise the absorber would overheat, which could cause damage to the rubber
parts.

In the conventional twin tube absorber, the liquid volume displaced by the immersing piston
rod is pressed through the bottom valve into the casing tube which serves as the
compensational volume and re-aspirated when the piston rod exits again. In the rest
condition the liquid pressure is equal to the ambient pressure.

In order to suppress cavitation occurrences at the valves, considerable absorption work can
be achieved by those oil flows, which are notaspirated into an enlarging part of the working
room.

During the compression of the twin-tube absorber, most of the absorption work is achieved
by the oil flow that is pressed through the bottom valve into the casing tube due to the
intrusion of the piston rod. The other oil flow is leading into the upper part of the working
room without any resistance.

On the other hand, during the dilatation of the conventional twin-tube absorber, the
absorption work is compensated by the oil flow, which is leading through the piston valve
Components of the total Suspension System 47

from the upper part of the working room to the lower part of the working room. Accordingly,
the oil flow that is running back from the compensational volume due to the returning piston
rod, is not prevented by any resistance.

In the mono tube absorber, the liquid volume displaced by the piston rod is compensated by
compressing a gas volume included in the absorber. This gas volume is either above the
surface of the oil and is separated from the hydraulic oil by an impact plate, or is separated
from the oil volume by a dividing piston, if the gas is placed in the lower part of the absorber.

The gas force exerted on the absorber piston surface by the gas pressure must be bigger
than the maximum damping force, as otherwise the gas volume would be compressed with
sudden movement of the piston, while in the part of the absorber work space facing the gas
volume, the pressure would drop to 0 bar. With sudden reversal of the direction of the piston
rod movement, this would cause a brief breakdown of the absorber and cavitation in the area
near to the piston valves. The gas pressure usually amounts to 30-40 bar. The gas pressure
acting on the cross section of the piston rod results in a driving out force of the piston rod,
which must be considered during the design procedure of the body suspension.

The advantage of the twin tube absorber is its cost and its life span, while the tendency
towards cavitation, unfavourable heat dissipation, its relatively large diameter and the fact
that it can generally be installed only perpendicularly orpettily inclined, are the
disadvantages.

The main disadvantage of the mono tube absorber in comparison to the twin tube absorber is
the fact that the precision required in production makes it more expensive and that its life
span may be shorter because of the critical piston rod sealing.

• Examples of design

The constructional design of the telescopic absorbers are different from one manufacturer to
another, but also in addition to the field of application.

Fig. 1.3-30 shows a telescopic damper designed according to the principle developed by De
Carbon in the 40s. A dividing piston separates the gas volume from the in order to prevent
the foaming of the oil.
48 Components of the total Suspension System

Fig. 1.3-32: Example of a mono tube absorber

This dividing piston enables any arbitrary position of installation, but it does not influence the
function of the absorber. The foaming of oil and gas can also be prevented by other
measures. For example, an impact disk already reduces the foaming of oil and gas, by
deviating and decelerating the high speed oil jets escaping from the piston valves and
rushing up the cylinder wall.

A sectional view of the twin tube absorber is shown in Fig. 1.3-31. The working volume, in
which the piston moves, is totally filled up with oil, and the storage volume between the
working cylinder and the container pipe is filled with oil up to an amount of 2/3. This tube
space serves as balancing volume for the oil displaced by the immersing piston rod. The
absorption valves - base valves and piston valves - consist of a system of spring washers
and valve bodies with throttle bores. The valves are held in position against the valve seat by
a weak coil spring and can also execute the function of a non return valve at the same time.
Components of the total Suspension System 49

Fig. 1.3-33: Design of a twin tube absorber

The gas pressure twin tube absorber was developed specially for the application in
suspension struts. With spring suspension and shock absorber strut wheel suspensions, the
piston rod of the absorber is not only used for the transfer of the absorber forces, but also for
wheel control at the same time. In order to keep a deflection of the piston rod caused by the
transverse forces within limits, larger piston rod diameters (~ 25 mm) are necessary
compared to absorbers or spring carriers exclusively stressed by expansion and
compression (~12 mm).

Accordingly, during retracting and emerging the piston rod, the oil volumes flowing from the
work space to the balancing volume and the other way round are also bigger. For this fact,
the danger of cavitation at the non return valves in the piston and at the base of the absorber
becomes bigger too. This leads to a brief breakdown of the absorber. A significant
improvement of the operational behavior of these absorbers could be obtained if the oil
reservoir is exposed to a relatively small gas pressure (usually 6-8 bar).
50 Components of the total Suspension System

Fig. 1.3-34: Typical designs of absorbers

By that, one receives an exact function of the absorber without having to accept the
disadvantage of a mono tube absorber, that the response of the absorber is impaired by the
clamping friction due to the piston rod seal which has to be dimensioned with high surface
pressing because of the high internal pressure.

• Operational behavior and damper characteristics

As previously mentioned, the absorbers used today operate based on hydraulics without any
exception. The damping force FD is a function of the compression velocity, and so follows the
relation:

FD = -sign ( z& rel ) ⋅ k ⋅ | z& rel |n (1.3-39)

with :

n absorption exponent

k absorption factor

“sign ( z& rel )” indicates the direction of the absorber force. The absorber characteristics
describe the functional relation between the absorber force and compression velocity. For
the determination of the absorber characteristics of an implemented absorber, the absorber
forces can be measured on a testing machine with a pushing crank mechanism at constant
Components of the total Suspension System 51

travel and varying test speeds, that cause different max. piston speeds in each case, Fig.
1.3-33.

Fig. 1.3-35: Damper work chart of the absorber operation for the determination of the
absorber characteristics

For scheduling the absorber characteristics, the max. traction and thrust forces are
registered as functions of the max. piston speed.

The sequence chart is closely coupled with the absorber characteristics. Fig. 1.3-36 shows
the possible cases of design.

The degressive process characteristics “3” has to be assigned to the sequence chart with the
largest surface, and so the dimensioning indicates the highest mean absorption. In
comparison of “3” to the designs “1” and “2”, its gradient at the zero point is higher, which
indicates a relatively large damping force with small piston speeds. This has an unfavourable
effect on the absorption capacity of the suspension on small unevenness, which however
brings certain advantages concerning the roll and pitch absorption.

The progressive absorption “1” has the advantage that the forces around the zero point are
small and so, favours a smooth rolling comfort, even for hard tires. The piston forces,
significantly rising with higher piston speeds, cause an increase of wheel and body
absorption, which positively affects the contact to the ground on bad road surfaces.

The mean absorption is smaller than in the case of the degressive dimensioning “3”, as
shown in the sequence chart.
52 Components of the total Suspension System

Fig. 1.3-36: Damper force depending on relative velocity at different damping coefficients

The relation between expansion and compression stage is different depending on the
manufacturer and the targeted application of the vehicle. Fig. 1.3-37 shows the influence of
the damper layout on the wheel load fluctuations.
Components of the total Suspension System 53

Fig. 1.3-37: Influence of damper layout on wheel load variation

A relation between rebound and compression stage of 1 enforces axle oscillations to fade
away very quickly. The wheel load fluctuations achieve a minimum for this tuning, which
means a better road grip of the wheels. This layout is however not favourable regarding the
suspension comfort.

Both the relation between rebound and compression stage and the absolute magnitude of
the absorber constant “k” are usually investigated based on tests. With that, an optimum
damper design is determined for the spectrum of the possible loading conditions of a vehicle
and for all possible road surfaces .

In order to reduce the conflicting aims between a high driving safety and max. attainable
comfort as far as possible, some vehicles are provided with an adjustable chassis
absorption.
54 Components of the total Suspension System

Fig. 1.3-38: Double piston absorber with discontinuously variable absorber characteristics

The main parts are a couple piston and a direct current motor placed in the hollow piston rod.
This motor actuates two rotary valves situated in the piston rod, which are both equipped
with individual piston valves. By that, two independent characteristics are realizable, Fig. 1.3-
39. In order to make the adjustment in rebound and compression direction effective, a spring
loaded check valve is assigned to each rotary valve for the rebound stage and a soft one for
the compression stage.
Components of the total Suspension System 55

Fig. 1.3-39: Absorber characteristics with the adjustments ‘Sport’ and ‘Comfort’

Fig. 1.3-40: Adjustable absorber with switching valves at the inside

56 Components of the total Suspension System

The adjustment times for the ADC-1 system are in a time range of 30-200 ms, which does
not permit reacting to single obstacles. Adjustable absorbers with switching valves at the
inside enable significantly shorter switching times (20-100 ms) and also a reaction on single
obstacles in a limited extend. Fig. 1.3-40 shows an adjustment system fully integrated in the
absorber. The control of the piston valves is performed by electromagnets (solenoid valves).

Continuously adjustable absorbers without fixed characteristics which could be implemented

by proportional valves, for example, are still in the development stage at present.

1.3.4 Seats

Spring-cushioned seats originally have only one function, the ‘upholstery’. With humans
placed on it, the seat forms an oscillatory system, whose natural frequency is in the range of
2,5 – 5 Hz, depending on the spring rigidity of the bolster, Fig. 1.3-41.

Fig. 1.3-41: Amplification functions of different vehicle seats

The location of increased resonance between seat ( &z& S ) and body acceleration ( &z& A ) of
&z& S
> 3 may reduce the suspension comfort significantly, in particular, if the body natural
&z& A
frequency is high, as in case of construction machines, agricultural tractors and partly of
trucks, since body and seat resonance frequency are close to each other in this case. A
remedy can be obtained here by the application of suspended seats, Fig. 1.3-42.
Components of the total Suspension System 57

Fig. 1.3-42: Suspended seat

1.3.5 Evaluation of Oscillations by Humans

The quantitative evaluation of the oscillation comfort requires a standard level for oscillation
exposure (and a reference road). For this, serial investigations have lead to the guidelines
VDI 2057 and ISO 2631. In these guidelines, there exists a difference between the
perception intensity and the duration of exposure.

• Perception intensity of sitting humans

The human body represents an oscillatory system, which can be approximated by a backup
system for humans sitting in the vehicle according to Fig.1.3-43.

Since humans are oscillatory systems themselves, it is impossible for them to evaluate the
effects of oscillation only by their intensity, but they tend to perceive oscillations of the same
intensity but of different frequencies variably. As a consequence, there exists a frequency-
dependent evaluation between the physically measured values and subjective perception,
which is different for individual parts of the body and different directions of action.
58 Components of the total Suspension System

Fig. 1.3-43: Backup system of a seated person

If one considers at first only the excitation of the body by the seat and if the average friction
R occurring in the bolster between back and backrest is added to the velocity-dependent
absorption, then the following results are given for this simple damped oscillator due to
appropriate tests with different test persons:

fe ≈ 5s − 1, D ≈ 0,45

and the appropriate amplification function, see Fig. 1.3-44.

V = zk ( f ) z s ( f ) (1.3-40)

or, for sinusoidal vibration simulation:

zk ( f ) &&
V = && zS ( f ) (1.3-41)
Components of the total Suspension System 59

.. 1 + 4 D² η²
.. z
zS v = ..K V= -
zS (1- η²)² + 4 D² η² ..
zS für K = const.

..
K ~ z K = const.
1

D ~ 0,4

4 8
fe ~ 5 Hz f e. 2 excitement frequency
KFZ_II_1-3_eng

Fig. 1.3-44: Amplification function of head acceleration during vertical excitation of seat

If the vertical acceleration of the head &z&K is considered as a measure for the perception
intensity, then you will receive for the perception intensity:

zK ( f ) = V ( f ) ⋅ &&
K ≈ && zS ( f ) (1.3-42)

with the amplification function V for a simple oscillator follows:

1 + 4 D2η2
K ≈ &&
zS ⋅ (1.3-43)
2
(1 − η )2
+ 4D η 2 2

f
with: η =
fe

This entails the excitation &z&S ( f ) , which leads to a constant perception intensity K*:

1
&z& S ( f ) ≈ (for K* = const.) (1.3-44)
V( f )

Fig. 1.3-45 shows the curves of equal perception intensity determined for sitting and standing
humans in the VDI specification 2057. For comparison, the reciprocal amplification function
(eq. 1.3-44) is included.

The curves are divided into three areas, which were linearised for the simplification of the
analysis .
60 Components of the total Suspension System

Fig. 1.3-45: Curves of equal perception intensity KZ of sitting and standing humans
according to VDI 2057 and ISO 2631.

Since all investigations done up to now show a correspondence in the fact that the
perception of humans begins at approx. 0.8 to 1 Hz and ends at approx. 70 to 100 Hz, the
curves are applied in a frequency range between 1 and 80 Hz. The first segment of the curve
is from 1 to 4 Hz. Within this segment the KZ values are calculated as follows:

&z&
KZ = 10 ⋅ ⋅ f / Hz (1.3-45)
m / s2

With &z& as the effective value in a narrow frequency band ∆f around the frequency f.

In the range between 4 and 8 Hz humans have the highest oscillation sensitivity:

&z&
KZ = 20 ⋅ (1.3–46)
m / s2

Within the range between 8 and 80 Hz the KZ value is calculated as follows:

&z& 1
KZ = 160 ⋅ 2
⋅ (1.3-47)
m / s f / Hz

While the perception of an exposing oscillation occurs immediately, the stress depends,
apart from the intensity of the oscillation, also on the duration of exposure. Fig. 1.3-46 shows
the evaluated vibration intensity (effective value of vertical accelerations evaluated according
to the eqations 1.3-45 to 1.3-47) as a function of the duration of exposure with equal
intensities for the criteria ‘wellness‘, ’performance’ and ‘health’.
Components of the total Suspension System 61

As a consequence, durations of exposure between 1 and 10 minutes lead to the same

stress. If the exposure continues for a longer time, the stress remains the same, if the square
of the evaluated vibration intensity K decreases with respect to the exposure time t (Fig. 1.3-
45 with double logarithmic scale)

A K-value of 20 is considered as limit value, which impairs the wellness of humans also for
short durations of exposure. With the same stress after 1.5 hours, impairments of
performance increase, and a daily duration of exposure of more than 4 hours may lead to
health damage.

Fig. 1.3-46: Evaluated vibration intensity as a function of the duration of exposure with equal
stress for the criteria ‘wellness‘, ’performance’ and ‘health’.

• Perception intensity of lying humans

The oscillation exposure on lying humans is of interest, for example, in long distance trucks
with sleeping cabins or in ambulance and rescue cars. A huge number of tests results
already available points to the fact that, because of the direct oscillation effect on the head,
the acceleration acting here can be applied as a measure for the oscillation stress of lying
62 Components of the total Suspension System

humans without evaluation.The curves of equal perception intensity for lying humans
prescribed in the VDI specification 2057 are given in Fig. 1.3-47.

Fig. 1.3-47: Curves of equal perception intensity for lying humans based on VDI 2057

The frequency range is also divided here into three areas for the exposure executed
perpendicularly to the direction of the spinal column:

&x&
1 ≤ f ≤ 3,15Hz KXL = 15 ⋅ (1.3-48)
m / s2

&x&
3,15 ≤ f ≤ 6,3Hz KXL = 4,8 ⋅ ⋅ f / Hz (1.3-49)
m / s2

&x&
6,3 ≤ f ≤ 80Hz KXL = 30 ⋅ (1.3-50)
m / s2
Single Wheel Suspension Model 63

1.4 Single Wheel Suspension Model

To describe the vertical dynamics of motor vehicles, different mechanical models can be
used. They vary from the simple single-mass system up to spatial vehicle oscillation models.
In the following section, a few models are presented.

Of central importance to all model approaches is the model of the vertical tire characteristics,
Fig.1.4-1.

a b c d

kW kW
cW cW kW c Wp
cW c Ws

z Fz

Fig.1.4-1: Tire models (vertical dynamics)

In the simplest case the tire can be modelled as a linear spring (a). The tire absorption
mentioned in chapter 1.3.1 is considered in the models after Voigt-Kelvin (b), Maxwell (c) and
Gehmann (d).

Below, the four modeling approaches will be compared with regards to their dynamic
behavior. For this purpose, the Laplace transformations

Fz (s)
Model a = cW (1.4-1a)
z(s)

Fz (s)
Model b = cW + kW ⋅s (1.4-1b)
z(s)

Fz (s) kW ⋅s
Model c = (1.4-1c)
z(s) k
1+ W ⋅ s
cW
64 Single Wheel Suspension Model

 c Wp 
c Wp + k W ⋅ 1 +  ⋅ s
Model d
Fz (s)
=  c Ws  (1.4-1d)
z(s) k
1+ W ⋅ s
c Ws

of the four models are presented in the frequency domain.

Frequency progressions of different vertical dynamic tire models

2

1
amplitude

Hooke
0 Voigt-Kelvin
Maxwell
Gehmann
-1
-1 0 1 2
10 10 10 10
Exciting frequency [Hz]

135

90
phase [°]

45

-45
-1 0 1 2
10 10 10 10
Exciting frequency [Hz]
KFZ_II_1-4_eng.ppt

Fig.1.4-2: Frequency responses of the tire models according to Fig.1.4-1 (the amplitude
response shows the normalised response z)

The behavior of Hooke’s approach (a) expresses itself in a constant amplitude response
without phase. In comparison, the tire models with damping elements (b)-(d), show
frequency dependent responses, whereby it becomes evident that the Maxwell model is
inapplicable as a tire model. The static spring rate disappears, while over the entire
frequency range the standardised absolute value of the amplitude is smaller than unity.

The dynamic spring stiffening, which occurs with rubber tires as the excitation frequency
increases, can be simulated with the models of Voigt Kelvin (b) and Gehmann(d), as the
respective amplitude responses show. As this stiffening transitions into saturation, the
Gehmann model describes the tire behavior in a more realistic manner.

In chapter 1.3.1 (Fig. 1.3-5) it was shown that, with tires at standstill (slow rolling), tire
damping decreases with increasing excitation frequency. This effect is also reproduced by
model (d), as the phase response of the Gehmann model shows. For high frequencies the
phase response of model (d) goes down to zero.
Single Wheel Suspension Model 65

Overall, the Gehmann model shows the best approximation of the vertical-dynamic tire
behavior. For fundamental investigations, as carried out in the following, also the Voigt Kelvin
model can be used.

1.4.1 Single Mass Model

The simplest vehicle model is the single mass model as shown in Fig.1.4-3. The mass
corresponds to the share of the body mass, with which is the wheel is loaded. The axle mass
is connected as unsprung mass to the body. The task of the suspension - e.g. with
construction machines or dump trucks - is taken over by the tire. Only the tire damping acts
as a shock absorber.

ZB
mB

CW kW

ZE

F Wdyn

Fig.1.4-3: Single mass suspension model

The system is described by the following equation of motion:

mB ⋅&z&B = −k W ⋅ (z& B − z& E ) − c W ⋅ (z B − z E ) (1.4-2)

kW c
&z&B = − ⋅ (z& B − z& E ) − W ⋅ (z B − z E ) (1.4-3)
mB mB

The natural frequency ωe as well as the Lehr absorption D can, in case the base excitation is
disregarded, i.e. by solving the homogeneous portion of this differential equation by means of
the approach z = z 0 ⋅ e ωt , be derived as:
66 Single Wheel Suspension Model

cW kW kW
ωe = and D= = (1.4-4)
mB k krit 2 ⋅ m B ⋅ ωe

Thereby the following correlation exists between undamped natural frequency ωe, damped
natural frequency ωe m.D and damping D:

ωe m.D. = ωe 1− D 2 (1.4-5)

To determine the motion of the body as well as the spring and damper forces during arbitrary
excitation (e.g. a measured road surface profile) computers are most suitable, particularly
when non-linear characteristics (e.g. tire lift off, kinked shock-mount and feather/spring
characteristic curves) are to be considered.

If one uses a swept sine (sine of constant amplitude and varied frequency) as excitation
signal, the amplification function can be determined from the peak values of body amplitude
and excitation amplitude:

zB
V (f ) = (1.4-6)
zE

Note: The amplification function z B /z E for the body amplitudes normalised to the excitation
amplitudes is identical to the amplification function &z&B /&z&E for body acceleration normalised to
the excitation accelerations, as follows from the double differentiation of a sine oscillation
with the frequency f = 2 ⋅ π ⋅ ω :

&z&E (t) = −ω2 ⋅ z E (t)

&z&B (t) = −ω2 ⋅ z B (t)

and with

z B ( t ) &z&B ( t )
=
z E ( t ) &z&E ( t )

The amplification function of the single mass suspension model is shown in Fig.1.4-4 for tire
data in as per paragraph 1.2.1.

Due to the negligible self damping behavior of the tires, a pronounced resonance peak
occurs. The natural frequency – resulting from proportionate body and axle mass and the tire
spring rate – which is approximately 3 - 4 Hz, is situated in a frequency range of high human
sensitivity to vibration.
Single Wheel Suspension Model 67

10
D = 0,05
5

ZB
ZE
1
0,5

amplification factor
0,2

0,1

0,05

0,02

0,01
0,1 0,2 0,5 1 2 5 10 s-1 20
frequency ω/2 π

Fig.1.4-4: Amplification function of the single-mass suspension system

1.4.2 Dual-Mass Equivalent System

Common motor vehicles not only feature tire (wheel) but also body suspension. The simplest
equivalent system, which nevertheless possesses essential features of a real vehicle
suspension, is represented by the dual-mass equivalent system described below. It is arrived
at by reduction on the basis of a four-wheel vehicle, with the share of the body mass
apportioned to the wheel considered substituting the body mass. The influence of mass
coupling, among other things, is neglected.

Fig. 1.4-5 shows the structure of a dual-mass equivalent system. The system consists of the
proportional body mass, a wheel or axle mass, the body springs and dampers as well as the
tire (wheel) suspension and damping.
68 Single Wheel Suspension Model

Fig. 1.4-5: Dual-mass suspension model

The differential equations describing the system result from the formulation of the force
balance on body mass (equation 1.4-7) and wheel mass (equation 1.4-8):

mB ⋅ &z& B = −k B ⋅ (z& B − z& W ) − c B ⋅ (z B − z W ) (1.4-7)

m W ⋅ &z& W = −k B ⋅ (z& W − z& B ) − c B ⋅ (z W − z B ) − k W ⋅ (z& W − z& E ) − c W ⋅ (z W − z E ) (1.4-8)

These differential equations are coupled via body suspension or body damping. For
approximate determination of the two natural radian frequencies ωe and the damping values
D, coupling of the two differential equations is to be neglected, so that only the homogenous
parts of the differential equations are considered.

Thus, for the body mass mA follows:

mB ⋅ &z&B + k B ⋅ z& B + c B ⋅ z B =0 (1.4-9)

from which follow:

cB
ωeB = (1.4-10)
mB
Single Wheel Suspension Model 69

kB
DB = (1.4-11)
2mB ωeB

for natural radian frequency ωeA and damping DA.

For the wheel mass mR from :

m W ⋅ &z& W + (k W + k B )⋅z& W + (c W +c B )⋅z W =0 (1.4-12)

follow natural radian frequency ωeR and damping D at:

c W + cB
ωeW = (1.4-13)
mW

kB + k W kB + k W
DW = =
2 mW ωeW 2 mW (c W + c B )

Furthermore, through formulation of the force balance at the tire contact center and with the
aid of the equations 1.4-6 and 1.4-7, an expression can be determined for the tire forces
acting on the roadway, i.e. an equation for the dynamic wheel load FWdyn :

FWdyn = −k W ⋅ (z& W − z& E ) − c W ⋅ (z W − z E ) = m B ⋅ &z&B + m W ⋅ &z& W (1.4-14)

With the aid of this equation, a method of indirect measurement of dynamic wheel load can
be derived through the measurement of body and wheel mass acceleration and knowledge
of the masses.

In the single-mass equivalent system, the magnification function was determined by means
of a sliding sinusoid acting on the system as excitation signal zE and calculation of the peak
values of the body amplitude zA. It is also possible to excite the system with the aid of a
synthetically generated roadway, and from this to infer the magnification function. This
method is demonstrated for the dual-mass equivalent system.

This method uses simulation to determine the spectral power density of body acceleration
φ Z&& A (ω) . The spectral density of the excitation amplitude φ ZE (ω) is determined by the
roadway. It must be converted to the spectral density of the excitation acceleration φ Z&& (ω) .
E

From the combination between power density and the quadratic mean value of the amplitude
in the time domain

∫Φ
2
z E (t) = ZE (ω) dω (1.4-15)
0
70 Single Wheel Suspension Model

and the analogous relation for the exciter acceleration

∞
&z&E 2 (t) = ∫Φ
0
&&
Z E
(ω) dω (1.4-16)

as well as from the interrelationship between exciter acceleration and exciter amplitude

&z&E (t) = −ω2 ⋅ z E (t) (1.4-17)

follows the relation between the power density spectrum of the

exciter amplitude and the exciter acceleration:

Φ Z&& E (ω) = ω 4 ⋅ Φ ZE (ω) (1.4-18)

The magnification function V can then be determined on the basis of the following
relationship:

2
 &z&B  φ Z&&B (ω)
V =   =
2
(1.4-19)
 &z&E  φ Z&&E (ω)

Fig. 1.4-6 illustrates above relationships.

Single Wheel Suspension Model 71

z A / &&
Fig. 1.4-6: Determination of the magnification function && zE on the basis of the spectral
density of the body acceleration φ Z&& (ω ) and the spectral density of the excitation
A

amplitude φ ZE (ω )

With the aid of the relationships explained above, the influence of various roadways (degree
of irregularity φh(Ω0) and waviness w) on the spectral density of body acceleration φ Z&& (ω )
A

can be discussed.

For example, a roadway with a very high share of short-wave excitations (low waviness w)
would on the basis of equation 1.4-17 result in a high density of the excitation acceleration in
the range of high frequencies. This again would on the one hand overrate the resonance
peak in the range of the wheel natural frequency ( fnW ≈ 12 Hz ) and on the other hand
underrate the range of the body natural frequency ( fnB ≈ 1 Hz ).

Analogous to the magnification function for the body amplitudes, a magnification function for
the dynamic wheel load variations relating to the exciter amplitudes FW ,dyn / zE can be
72 Single Wheel Suspension Model

determined. This magnification function is given here as a standard value for the static wheel
load FW ,stat (Fig. 1.4-7). It follows direct from the spectral density of the exciter amplitude
φ ZE (ω ) :

2
 FW dyn  φ FR dyn /FR stat (ω)
V =
2  = (1.4-20)
F ⋅ z  φ zE (ω)
 W stat E 

Fig. 1.4-7: Determination of the magnification function FRdyn / (FRstat ⋅ zE ) from the spectral

density of excitation amplitude φ ZE (ω ) and dynamic wheel load φFR dyn / FR stat (ω )

1.4.2.1 Parametric Study - Automobile Suspension

In the following we are investigating how modification of essential parameters of the dual-
mass model affects body acceleration, which even without seat suspension and vibration
measurement is considered by rough human assessment a criterion of suspension comfort,
and relative (specific) dynamic wheel load, i. e. road grip of the wheels (driving safety).

The oscillation equations of the dual-mass model, equation 1.4-7 and equation 1.4-8, are
mapped by means of the MATLAB program and excited by a synthetically generated road
signal. On the basis of the simulation results, the spectral power densities of body
Single Wheel Suspension Model 73

acceleration and dynamic wheel load are determined. The road signal used for simulation is
shown in Fig. 1.4-8 as path-time function as well as time-frequency-dependent power
density.

Fig. 1.4-8: Synthetically generated road signal with w = 2.14 and φh(Ω0) = 3.7 ⋅ 10-6 m3

Original variant is a quarter vehicle with the following parameters:

R = 150,000 N/m; A = 21,000 N/m; kW = 100 Ns/m; kB = 1500 Ns/m; mW = 40 kg; mB = 400 kg

Fig. 1.4-9 shows how in this vehicle the road signal from Fig. 1.4-8 affects the vertical
quantities of motion of wheel and body. Damping and reduction of the motion parameters
between road and body can be clearly seen. For better perception, the curves of body and
wheel have been vertically displaced.
74 Single Wheel Suspension Model

Fig. 1.4-9: Vertical motion, speed and acceleration of road, wheel and body

• Variation of the wheel mass mW

In Fig. 1.4-10 the wheel mass is varied first. With regard to body resonance, this has neither
an effect on the position of the body natural frequency nor on the intensity of the body
resonance.
Single Wheel Suspension Model 75

Fig. 1.4-10: Parameter variation with varying wheel mass

With regard to wheel resonance, however, the elevated relative (specific) dynamic wheel
loads resulting from a greater wheel mass are conspicuous. This is due to the fact that a
larger mass has to be stabilized by unchanged dampers. Consequently, a wheel mass as
small as possible is to be aimed at with regard to the dynamic wheel load variations, and
therefore also with regard to driving safety.

• Variation of the stiffness/spring rate of the tire (tire rate) cW

In Fig. 1.4-11 the spring rate of the tire (tire rate) is varied. A softer tire rate tends to reduce
wheel natural frequency and the dynamic wheel loads, i.e. road grip is improved. Softer tires
76 Single Wheel Suspension Model

would thus clearly improve driving safety. Realization of such tires is, however, only possible
within narrow limits due to the associated increase in rolling resistance and flexing energy.

Fig. 1.4-11: Parameter variation with varying stiffness/spring rate of the tire

• Variation of the stiffness/spring rate of body suspension cB

In Fig. 1.4-12 the spring rate of the body suspension is varied. A softer body suspension rate
reduces body natural frequency, and relative damping is increased as a result. Body
acceleration and relative (specific) dynamic wheel load are reduced.
Single Wheel Suspension Model 77

A softer body suspension rate would thus have a positive effect with regard to the criteria
driving safety and ride comfort. Realization, however, is only possible to a limited degree
because of effects such as body inclination due to cornering, brake dive, major ride height
changes due to loading, and especially because of the considerable spring travels then
needed.

Fig. 1.4-12: Parameter variation with varying stiffness/spring rate of body suspension
78 Single Wheel Suspension Model

• Variation of body damping kB

In Fig. 1.4-13 body damping is varied. Stiffer body damping has a positive effect in the range
of the natural frequencies of both body acceleration (suspension comfort) and the dynamic
wheel loads (road grip of the wheels, active safety). In the range outside the points of
resonance, however, it is especially the body acceleration which is positively influenced by a
soft damper.

Tuning of body damping consequently always requires a compromise to be made between

the requirements of the various frequency ranges.

Fig. 1.4-13: Parameter variation with varying body damping

In Fig. 1.4-14 the results gained on the basis of Figs. 1.4-10 - 1.4-13 with regard to driving
safety and riding comfort are summarized in a table. Analogous to body and wheel natural
frequencies, a difference is made between long-wave and short-wave effects.
Single Wheel Suspension Model 79

Fig. 1.4-14: Effects of design changes effected on the suspension system on driving safety
(dynamic wheel loads) and riding comfort (body acceleration);
↑ = positive, ↓ = negative

In addition to the "conventional" ways of influencing suspension behavior dealt with so far,
"unconventional" solutions are being presented below. In contrast to passive systems, the
force F acting here on the body does not depend via a single characteristic on the
compression travel z and the compression rate z& of the components. Fig. 1.4-15 presents a
summary of controlled suspension systems.

Adaptive systems, unlike the passive system, can switch between various component
characteristics. Here too, however, the force direction remains determined by the signs of
compression travel and compression rate.

In the semiactive state, the switching frequencies exceed the characteristic vibration periods
of wheel and body. Switching from one characteristic to the other can thus take place at a
speed permitting each point located in between to be reached dynamically.

Adaptive and semiactive systems only need energy for the control of actuators and the
electronics.

Only with active suspension does the force F between body and wheel become independent
of the compression movement of the wheel. Positioning force, however, requires external
energy supply.
80 Single Wheel Suspension Model

Fig. 1.4-15: Classification of controlled suspension systems

• Adaptive damper control

The results of parameter variation regarding damper stiffness (Fig. 1.4-13) suggested that
body acceleration as well as wheel load variations can be minimized for a wide frequency
range if a relatively high level of damping is available in the body and wheel resonance
ranges and if the level of damping is low in the range outside these points of resonance.

A clear improvement of conventional suspension systems can thus theoretically also be

achieved by means of a damping system that depends on the excitation frequency of the
system, Fig. 1.4-16.

Fig. 1.4-16: Simulation results with a single-wheel suspension model for ideal frequency-
dependent damper force adjustment /21/
Single Wheel Suspension Model 81

The improvement potential attainable by means of such damper control is limited by the fact
that a definite frequency selection is not possible due to the randomness of road irregularity.
Apart from that, the improvement potential also depends on the realizable range of variation
of damper characteristics and on the potential extent of reduction of system life.

The EDC system (Electronic Damper Control) of BMW was developed in cooperation with
Boge and VDO. Here, the damper characteristics are adapted in three steps by means of
control electronics. The EDC system identifies road excitation by means of an acceleration
sensor mounted to the body.

The sensor signals are processed by the ECU in a manner allowing determination of
separate characteristic values of excitation with regard to body and axle natural frequencies.
Depending on vehicle load and the intensity of instantaneous vibration excitation, different
threshold values will result in graded damper adjustment.

System efficiency is illustrated by the conflict diagram between comfort and driving safety,
Fig. 1.4-17.

Fig. 1.4-17: Comparison of suspension behavior of a test vehicle with different damper
settings of the EDC system illustrated by the conflict diagram comfort/driving
safety /21/

The adaptive damping system (ADS) of Mercedes-Benz, whose system functions are
comparable to those of the EDC system, effects irregularity-dependent damper setting in 4
steps according to a characteristic map considering a comparative value for the spectral
power density of the irregularities Φ(Ω 0 ) and a comparative value for the waviness w of the
road being traveled on, Fig. 1.4-18.
82 Single Wheel Suspension Model

Fig. 1.4-18: Map for damper force adjustment of the adaptive damping system ADS of
Mercedes-Benz /25/

An effective value H of the course of irregularity h(t) in the frequency range 0.5 - 20 Hz.
( Φ′ = H ( 0.5 − 20 Hz ) ) determined by exponential floating calculation is used as a

comparative value for the spectral power density Φ(Ω 0 ) .

The comparative value for waviness is formed as a quotient resulting from the respective
effective values for the body and axle resonance ranges.

(w' = H (0,5 - 2 Hz) / H (8 - 20 Hz) )

The course of irregularity h(t) is indirectly determined by approximation through the

measurement of wheel acceleration &z&W and body acceleration &z&B . The algorithm used can
be derived from equations 1.4-7 and 1.4-8.

z E = h(t) :

mW m
h(t) = z W + ⋅ &z& W + B ⋅ &z&B (1.4-21)
cW cB

• Active suspension

Due to its low power/weight ratio and its high power density, a hydraulic cylinder is suited as
a variable-length element for active ride-height control (Fig. 1.4-19).
Single Wheel Suspension Model 83

Fig. 1.4-19: Single-wheel suspension model with active suspension

Cylinder pressure is controlled by means of a quick-acting servo valve, which receives the
control signals from an overriding electronic control unit. Control logic can be fed with various
information about the driving condition, picked up by sensors on the vehicle. Active setting
force can be controlled in relation to wheel or body acceleration or via a pre-scan of the road
irregularities.

Unlike a fully active suspension, the active version of the hydropneumatic suspension
provides a spring action also without inflow and outflow of hydraulic oil. The control valves
supply or remove oil only if the system recognizes that it can improve vehicle handling
through active intervention. The advantage is that even under extreme driving conditions, the
energy input required will be significantly lower.

In Fig. 1.4-20 the spectral power densities of active hydropneumatic and fully active
suspension are compared with those of passive systems. Suspension comfort as well as
road grip of the wheels are clearly improved in a wide range of frequencies.
84 Single Wheel Suspension Model

Fig. 1.4-20: Theoretical riding comfort and safety potential of various suspension systems

Realization of motor vehicles with significantly improved suspension comfort is technically

possible. Drawbacks are seen in the higher design expenditure and the energy requirements,
which for a standard-size car amounts to between 7 kW (active hydropneumatic suspension)
and 20 kW (fully active suspension).
Single Wheel Suspension Model 85

1.4.2.2 Parametric Study of a Truck Suspension

The suspension characteristics of trucks differ substantially from those of a passenger car.

1,00E+01
PSD body acceleration [(m/s²)²/Hz]

LKW
Truck
1,00E+00 PK W
Car

1,00E-01

1,00E-02

1,00E-03

1,00E-04
0,10 1,00 10,00 100,00
fre que ncy [Hz]

1,00E +08
PSD dynamic differences in wheel load [N²/Hz]

LKW
Truck
1,00E +07
Car
PKW

1,00E +06

1,00E +05

1,00E +04

1,00E +03

1,00E +02

1,00E +01
0,10 1,00 10,00 100,00
fre que ncy [Hz]

Fig.1.4-21: Comparison of body acceleration and dynamic wheel load between trucks and
passenger cars

Fig. 1.4-21 shows the spectral power densities of body acceleration and dynamic wheel load
of a truck in comparison to a passenger car.

The simulation of the truck is based on the following data:

R = 3,000,000 N/m; A = 570,000 N/m; kW = 100 Ns/m; kB = 21,000 Ns/m; mW = 650 kg;
mB = 4500 kg

The body’s natural frequency of 1.4 Hz is significantly higher than that of the example
passenger car. Also, the magnitudes of body acceleration and the dynamic wheel load are
86 Single Wheel Suspension Model

larger. The suspension behavior of trucks is therefore less favourable than that of passenger
cars.

• Influence of frictional forces within springs

As body springs of trucks usually stacked leaf springs are used. The characteristic curves
show a hysteresis (see paragraph 1.2.1.1), resulting from the friction between the spring
leaves. This leads to a temporary blockade of the body spring during the course of an
oscillation, thereby turning the two-mass system into a single mass system according to
Fig.1.4-3 during these periods. Body and wheel mass then oscillate in conjunction on the stiff
and little damped tire spring. Fig.1.4.-22 shows the velocity difference between body and
wheel with and without friction. With the friction-affected model, the phases where the leaf
springs blockade and the velocity difference becomes zero are clearly visible.

Fig.1.4-23 shows, how the friction force affects the body acceleration of the example LKW. In
particular within the range of the highest oscillation sensitivity of humans between 4-8 Hz, the
friction causes a significant degradation of the suspension behavior.

0,50
zW-zB [m/s]

0,00
. .

-0,50

Ffric = 0 N
Freib
Freib
Ffric = 2400 N
-1,00
0,0 1,0 2,0 3,0 4,0 5,0 6,0

time [s]

Fig.1.4-22: Velocity difference between body and wheel

Fig.1.4-23: Parametric variation with different amounts of friction within the leaf spring
Single Wheel Suspension Model 87

1.4.3 Enhancement of the Model by Seat Suspension

Enhancing the two-mass suspension model, which has been discussed so far, by the seat
suspension, leads to a three-mass suspension model (Fig.1.4-24).

ZS
m Seat+ m G

ZB
mB

ZW

mW

ZE

FW
dyn

Fig.1.4-24: Structure of a three-mass suspension model

The added mass represents the mass of the sprung parts of the seat and the person sitting
on it. Due to the small amount of mass added with respect to the mass of the body, the
feedback on the body can generally be neglected. One can therefore begin with a two-mass
system and add a simple oscillator.

In the following figures, this enhancement is done by superposing on the example car from
paragraph 1.4.2.1 for two different seat masses. Fig.1.4-5 shows the amplification function,
and Fig.1.4-6 shows the spectral power density of seat acceleration resulting from the seat’s
filter effect.
88 Single Wheel Suspension Model

2,50
mS = 30 kg

zB
zS
2,00 mS = 20 kg

amplification function
1,50

1,00

0,50

0,00
0,10 1,00 10,00 100,00
frequency [Hz]

Fig.1.4-5: Amplification function of a three-mass suspension model

Fig.1.4-6: Seat acceleration of a three-mass suspension model

Single-Track Suspension Model 89

1.5 Single-Track Suspension Model

1.5.1 Double-axle Vehicle with bending resistant Structure

In single-track suspension models, the body is not seen as a mass point (lumped mass) but
as a massive beam. In the simplest case, we are dealing with a model of a two-axle vehicle
with a stiff body, i.e. a flexurally stiff beam (Fig. 1.5-1).

lBf lBr
zBf zB ϑ zBr
mB ; Θ B

cBf kBf cBr kBr

zWf zWr
mWf mWr

cWf kWf cWr kWr

zGf zGr

Fig.1.5-1: Single track suspension model

First we need to formulate the differential equations. The single-track suspension model for a
two-axle vehicle according to Fig. 1.5-1 has four degrees of freedom:

- lifting and pitching of the body,

- lifting of front and rear axles.

The equations applicable for the center of gravity of the body are:

mB &z&B = −k Bf ( z& Bf − z& Wf ) − c Bf ( z Bf − z Wf )

(1.5-1)
− k Br ( z& Br − z& Wr ) − c Br ( z Br − z Wr )

ΘB ϑ && = l k ( z& − z& ) + l c ( z − z )

B f Bf Bf Wf f Bf Bf Wf
(1.5-2)
− lr k Br ( z& Br − z& Wr ) − lr c Br ( z Br − z Wr )

Front and rear axles follow the equations:

m Wf &z& Wf = k Bf ( z& Bf − z& Wf ) + c Bf ( z Bf − z Wf )

(1.5-3)
− k Wf ( z& Wf − z& Gr ) − c Wr ( z Wr − z Gr )
90 Single-Track Suspension Model

m Wr &z& Wr = k Br ( z& Br − z& Wr ) − c Br ( z Br − z Wr )

(1.5-4)
− k Wr ( z& Wr − z& Gr ) − c Wr ( z Wr − z Gr )

Between the body movements above the axles zAv and zAh, the movement of the body center
of gravity zA and the pitch angle ϑ there are the following relationships:

z Bf = z B − l f ϑ (1.5-5)

z Br = z B + lr ϑ (1.5-6)

As for the single-wheel suspension model, natural radian frequencies and damping values
can be derived from the differential equations for the single-track suspension model as well if
coupling of the differential equations is neglected. Starting from the notion that all degrees of
freedom – except for the one concerned – are blocked, we get:

natural circular frequency damping factor

c Bf + c Br kBf + k Br
travel
mB 2 mB (c Bf + c Br )

c Bf k Bf
body front
mBf 2 mBf c Bf

c Br k Br
body rear
mBr 2 mBr c Br

2 2 2 2
lBf c Bf + lBr c Br lBf k Bf + lBr k Br
pitch
ΘB 2
2 ΘB (lBf 2
c Bf + lBr c Br )

c Bf + c Wf k Bf + k Wf
front axle
mWf 2 mWf (c Bf + c Wf )

c Br + c Wr k Br + k Wr
rear axle (1.5-7)
mWr 2 mWr (c Br + c Wr )
Single-Track Suspension Model 91

The axle-proportionate body masses result from the position of the centre of gravity:

lBr
mBf =mB (1.5-8)
lBf + lBr

lBf
mBr =mB (1.5-9)
lBf + lBr

With regard to suspension comfort, pitch natural frequency should be low. With specified
suspension stiffness for lifting suspension, however, target-specific influencing is difficult,
because the other influencing parameters are usually determined by different aspects
(center-of-gravity location, wheelbase) or more or less result from the vehicle concept
(moment of inertia).

1.5.1.1 Excitation by real unevenness Routing

Unlike the single-wheel suspension model, the single-track suspension model is excited by a
single road irregularity twice: that is, first on the front axle and then once more on the rear
axle. As a result, body acceleration of the two-axle model does not only depend on the
excitation frequency but also on parameters like driving speed, the body natural frequency
and the position of the measuring point in the vehicle.

In the following, excitation of the vibration system vehicle by the irregularity profile of the road
is performed on front and rear axles deferred by the time differential ∆t = l / v (l wheelbase, v
driving speed). The motion equations are solved with the aid of the MATLAB simulation tool.

• Influence of driving speed

While the power density spectra of body acceleration and relative (specific) dynamic wheel
load variation permit a differentiated assessment of a vehicle suspension, it is possible to
assess a trend also on the basis of the respective effective values.

This approach makes sense, for example, if for trend assessment those parameters are
varied that affect the entire power density spectrum, since then a value for suspension
comfort or road grip can be stated as a function of the varied parameter.

The effective value of body acceleration can be determined on the basis of the time curve of
acceleration:

T
1
&z&Beff =
T
⋅ ∫
0
&z&B 2 dt

The effective value of body acceleration above the rear axle with average load was
determined in this way in relation to the driving speed of a model vehicle, Fig. 1.5-2.
92 Single-Track Suspension Model

0.90

body acceleration zBr eff [m/s2]

0.80
T
.. 1 ..
0.70 zB eff =
T
⋅ ∫ z B 2 dt

..
0
0.60

0.50

0.40

0.30
10.00 15.00 20.00 25.00 30.00
velocity [m/s]

z Ah in relation to the driving speed

Fig.1.5-2: Body acceleration &&

• Influence of body natural frequency

For determining the influence of the spring rates of the body springs – and thus of the body
natural frequency – on body acceleration, these spring rates are gradually reduced. For the
model vehicle with average load and a driving speed of 22 m/s we thus obtain a curve as
shown in Fig. 1.5-3.

z Ah in relation to the body natural frequency

Fig.1.5-3: Body acceleration &&

For low body acceleration and good suspension comfort to be attained, it is of crucial
importance to realize soft body springs, i.e. a low body natural frequency. This, however,
requires a suitable vehicle concept, such as a low center of gravity, wide spring tracks.
These boundary conditions were neglected in the considerations according to Fig. 1.5-3.
Single-Track Suspension Model 93

• Measuring point location

In Fig. 1.5-2 and Fig. 1.5-3, the body directly above the rear axle was defined as the
"measuring point" of body acceleration. With otherwise identical conditions, the value of body
acceleration also depends on the location of the measuring point on the body in vehicle
longitudinal direction. This influence is shown in Fig. 1.5-4 for the original state.

0 .7 0
body acceleration zB eff [m/s2]

0 .6 0

0 .5 0
..

0 .4 0

0 .3 0

0 .2 0

0 .1 0

0 .0 0
1 .0 0 .8 0 .6 0 .4 0 .2 0 .0 -0 .2 -0 .4 -0 .6 -0 .8 -1 .0 -1 .2 -1 .4
p o s itio n o f m e a s u rin g p o in t [m ]

Fig.1.5-4: Body acceleration in relation to the location of the measuring point in the vehicle
(The 0-value marks the center of gravity)

A minimum amount of body acceleration is found in the area of the center of gravity of the
vehicle. With increasing distance from the center of gravity, the effect of the overlap of lift and
pitch oscillations of the body on the resulting acceleration increases.

1.5.1.2 Mass and Spring Coupling

For further considerations, it seems practical to split the mass of the beam-shaped body into
three single point (lumped) masses (Fig. 1.5-5).

Bf Br

CG
Bf Br
link

Fig.1.5-5: Configuration of the single-track suspension model with coupled mass mLink

Since total mass mB center-of-gravity location lAv*, lAh* and moment of inertia ΘB must remain
the same, the following three conditional equations follow:
94 Single-Track Suspension Model

* *
mBf + mBr + mlink = mB (1.5-10)

* * * *
mBf ⋅ lBf = mBr ⋅ lBr (1.5-11)

* *2 * *2
mBf ⋅ lBf + mBr ⋅ lBr = ΘB (1.5-12)

Out of this the so-called coupling mass mlink results too:

ΘB
m link = mB − * *
(1.5-13)
lBf ⋅ lBr

• Mass coupling

When a vehicle's front axle passes a bump, the shock/impact force acting on the body will
result in body movement composed of the lifting motion of the body center of gravity SA and a
rotary motion around the vehicle transversal axis (pitching motion) through the center of
gravity.

∆ϑ
∆Z P
SB

Fim pulse s
l Bf* l Br*

Fig.1.5-6: Shock excitation in the single-track suspension model

Associated with this body movement, there is a point of rest in the longitudinal center plane
of the vehicle, which is referred to as center of impact /1/. The location of this point can be
determined with the aid of the principles of linear and angular momentum:

t2

∫F
t1
impulse dt = mB ⋅ z& 2 − mB ⋅ z& 1 (1.5-14)

t2

∫ Fimpulse ⋅ lBf dt = ΘB ⋅ ϑ& 2 − ΘB ⋅ ϑ& 1

*
(1.5-15)
t1
Single-Track Suspension Model 95

In consideration of the geometric relationship

s ⋅ ∆ ϑ ≈ ∆z

(
s ϑ& 2 − ϑ1 2)
& = z& − z&
1 (1.5-16)

follows:

ΘB
s= *
(1.5-17)
lBf ⋅ mB

With the aid of this relation, the three potential cases regarding the value of the coupled
mass, which as a purely theoretical operand can also turn out to be negative, can be
illustrated:

1.Couple mass equals zero (mlink = 0):

For the special case that the coupled mass equals zero,

* *
ΘB = mB ⋅ lBf ⋅ lBr

applies and thus for the impact (shock) on the front axle

s = lBr*

In this case, the body above the rear axle remains at rest (unexcited). We have shock
decoupling.

2. Couple mass greater than zero (mlink > 0):

If the coupled mass is greater than zero, the following equation applies for the impact on the
front axle:

s < lBr*

The center of impact in this case is located between body center of gravity and rear axle.
This means that with identical total mass and identical wheelbase, the pitching motion of the
body is more intense, which is to be considered a drawback with regard to riding comfort.

3. Couple mass less than zero (mlink < 0):

If the relation between body mass, mass moment of inertia and center-of-gravity distances is
such that the coupled mass is less than zero, then the following equation applies for the
impact on the front axle:

s > lBr*
96 Single-Track Suspension Model

The center of impact in this case is no longer found between the axles. In the body
movement resulting from the shock force on the front axle, we then usually find the share of
lifting motion to be higher than that of pitching motion, which is to be considered favorable
with regard to riding comfort.

These considerations can be made in a similar way for a rear axle passing a bump.

• Spring coupling :

Spring coupling as realized, for example, by the BLMC hydrolastic springing system, acts in
the same way as mass coupling, mko < 0. By means of a hydraulic connection, compression
of the front axle (due to a bump) causes rebounding of the rear axle, a response which
counteracts the pitching motion of the body.

connecting line

rubber shear spring double valve

(shock absorber)

damper hole separating box

butyl cover plate butyl rubber

membrane with nylon
pushrod connection
to wheel control

pressure filling valve

connecting line

rear front

Fig.1.5-7: "Hydrolastic" compound suspension system in the Morris 1100, 1962

Spring coupling acc. to Fig. 1.5-8 is an alternative way of reducing pitching natural frequency
with unchanged lifting natural frequency. While all springs are involved in the lifting motion,
only the exterior springs are involved in the pitching motion.
Single-Track Suspension Model 97

Fig.1.5-8: Two-axle vehicle with spring coupling

The potential of improvement of suspension behavior by means of coupled masses or

springs is very limited, because the measures required are mostly hardly realizable or there
is no latitude of design with regard to the influencing parameters.

1.5.2 Two axle Vehicle with additional Degrees of Freedom

In the two-axle vehicle according to Fig. 1.5-1, the body was assumed to be rigid. As
opposed to cars and vans, the bending elasticity of the chassis frame as well as the elasticity
of engine and cab suspension (heavy commercial vehicles usually feature elastic cab
suspension for better isolation against vibration and torsion of the frame) in trucks cannot be
neglected, because these factors significantly influence suspension behavior in these vehicle
types. Fig. 1.5-9 shows a truck equivalent system with 15 degrees of freedom.
98 Single-Track Suspension Model

driver
15 degrees of freedom

cabin
motor+
gearbox

frame

wheel wheel

Fig.1.5-9: Structural model of a platform truck

Semi trailer truck

The articulated vehicle is a vehicle type that presents particular problems with regard to
suspension and driving stability. With this vehicle type, semitrailer and road tractor are linked
via a joint in the vertical and transverse axes, Fig. 1.5-10.

ms F stlr
CG stlr
cs ks
CG trc

cB 1 kB 1 cB 2 kB 2 cB 3 kB 3

mw mw mw
1 2 3
c w1 cw c w3
2

Fig.1.5-10: Structural model of a semitrailer unit

Since the semitrailer transfers the vertical load (fifthwheel load) FSattel to the road tractor,
adequately loadable, i.e. stiff suspension is required. The moment of inertia acting during the
pitching motion of the road tractor is, however, relatively small. The high moment of inertia of
the semitrailer has no effect due to the fiftthweel coupling. The consequence of this is a
Single-Track Suspension Model 99

correspondingly high pitch natural frequency resulting in high K-values. Some improvement
can be achieved by means of soft front springs of the road tractor.

It is a drawback of this measure that getaway jacking-up and brake dive are aided and in
extreme cases even dynamic pitch angle and center-of-gravity acceleration are increased.

By means of a number of additional measures and determined tuning development, the

suspension behavior of articulated vehicles has been significantly improved and adapted to
that of other commercial vehicles in the meantime.

• Tandem axles of commercial motor vehicles

In HD trucks, tandem axles are used as:

- powered rear axles of trucks and articulated vehicles,

- nonpowered rear axles of trailers and semitrailers

The following example shows the powered tandem axle of a 22-t semitrailer truck, Fig. 1.5-
11, the design of which is described by the equivalent system according to Fig. 1.5-12.

Fig.1.5-11: Swinging tandem axle with hydraulic damping (MB)

Fig.1.5-12: Equivalent system of a semitrailer truck with hydraulically damped tandem axle
100 Single-Track Suspension Model

Such tandem axles possess two degrees of freedom: homodirectional and heterodirectional
vertical motion of the two axles. Natural vibration is characterized in particular by high
dynamic axle loads, i.e. roadway load.
Two-Track Suspension Model 101

1.6 Two-Track Suspension Model

Using a two-track suspension model as shown in Fig. 1.6-1, the degrees of freedom

- Roll springing

- Distortion of the body

- Trampling of rigid axles

shall be dealt with in this chapter.

ϕ
f
ϕ distortion body
r

body
roll

tramp axle

Fig.1.6-1: Two-track suspension model (truck)

1.6.1 Roll Springing

Roll springing is to be regarded first for a vehicle with rigid front and rear axles (e.g. truck,
jeep), since this case is kinematically simple.

While cornering, the body is laterally displaced due to the centrifugal force. This deflection
takes place around an axis of rotation, which passes through the roll poles at the front and
rear axles and is called the Roll axis.

The positions of the roll poles, which depend on the kinematics of the wheel control or the
axle coupling to the body, will be dealt with in the chapter "wheel suspensions".
102 Two-Track Suspension Model

Fig.1.6-2: Equivalent system model for the suspension

The centrifugal force FLat,B , acting at the centre of gravity of the body, gives rise to a torque
Mϕ around the roll axis according to fig. 1.6-2:

Mϕ = FLat,B ⋅ ∆h ⋅ cos ϕ + mB ⋅ g ⋅ ∆h ⋅ sin ϕ (1.6-1)

with ∆h: vertical distance between the centre of gravity of the body SB and the roll axis

The inclination of the roll axis in the vehicle longitudinal plane is thus neglected. The torque
created by the body weight is also usually neglected with passenger cars, i.e. applies here:
sin ϕ << cos ϕ.

With these simplifications, the aligning torques applied by the body springs around the roll
axis are given by:

s Sf s
FLat,B ⋅ ∆h = 2 ⋅ ⋅ c Bf ⋅ fSf + 2 ⋅ Sr ⋅ c Br ⋅ fSr (1.6-2)
2 2

with: SSf,r: Spring track width in front, rear

fSf,r: Spring compression in front, rear

CBf,r: Body spring rigidity in front, rear

sS
With fF ≈ ϕ ⋅ one finally receives for the roll angle ϕ:
2

2 ⋅ ∆h
ϕ= 2 2
⋅ FLat,B (1.6-3)
c B,f ⋅ s S,f + c B,r ⋅ s S,r
Two-Track Suspension Model 103

The roll angle ‘ϕ’ is thus inversely proportional to the square of the spring track width. For a
small body inclination, the spring track width should therefore be as high as possible while
driving along curves.

1.6.1.1 Stabilizer and Compensating Spring

Anti-roll suspension (roll damping) can be stiffened without changing vertical suspension
through installation of a stabilizer spring, Fig. 1.6-3.

Fig. 1.6-3: Functional principle of stabilizer and compensating spring

During the rolling motion of the body, i.e. a heterodirectional compression motion of the
wheels, the stabilizer is torqued and thus provides a self-aligning torque around the roll axis,
which reduces body inclination. In case of a pure lifting motion of the body suspension on the
axle considered, the stabilizer will have no effect. A stabilizer track width sStab , to which the
stabilizer stiffness cStab relates, is defined analogous to the spring track width sF.

Stabilizer stiffness cStab then corresponds to the stabilizer force at the ends of the stabilizer
spring referred to half the differential compression of these ends.

For the roll angle ϕ thus applies:

2 ⋅ ∆h ⋅ FLat,B
ϕ= 2 2 2 2
(1.6-4)
c B,f ⋅ s F,f + c Stab,f ⋅ s Stab,f + c B,r ⋅ s F,r + c Stab,r ⋅ s Stab,r

Should the ratio of the shares of the rolling moment Mϕ supported by the stabilizers on front
and rear axles differ from the ratio of the shares supported by the body suspension, or if only
one axle has a stabilizer spring installed, then not only the roll angle will be reduced, but the
104 Two-Track Suspension Model

distribution of the wheel load differentials, which occur during cornering between the wheels
of the right and left vehicle sides, among front and rear axle will be influenced.

The influences on vehicle handling caused by this are dealt with in the chapter "Transverse
Dynamics".

The compensation spring is an antagonistic spring that has an effect counteracting that of the
stabilizer spring. As a purely vertical suspension element it has not effect on the rolling
motion of the body.

In the past, the compensating spring was used for axle designs whose wheel suspension
kinematics required wheel load differentials to be as low as possible for the so-called bump
effect (bottoming-out effect) to be suppressed (cf. chapter on wheel suspensions). The
stiffness of body suspension and thus the share of the rolling moment supported on the axle
considered could then be reduced accordingly. The compensating spring is no longer found
in modern wheel suspension systems.

Fig. 1.6-4 shows an automobile front axle (Opel Senator) as an example for the installation of
a stabilizer

Fig. 1.6-4: Stabilizer spring on an automobile front axle

1.6.1.2 Vehicle Conception and Suspension Characteristics

The parameter studies in the paragraph 1.4.2.1 (passenger car) and 1.4.2.2 (truck) show that
a necessary prerequisite for the achievement of good suspension characteristics is a low
body- eigenfrequency or sufficiently soft body springs.
Two-Track Suspension Model 105

On the other hand, the lower limit of body eigenfrequency or spring rates is determined in
particular by static and quasi-static effects such as level modification by load and body
inclination while driving along curves. While static level modifications could be eliminated by
load-sensitive regulating springs, curve inclination can be only be reduced in small limits by
means of stabilizers. The relationship between body inclination, body natural frequency and
body acceleration is therefore considered in the following.

According to paragraph 1.6.1.1, while driving along curves (steady-state circular-course

driving) a roll angle is generated:

2 ⋅ ∆h ⋅ FLat,B
ϕ= 2
(1.6-5)
(c B + c Stab ) ⋅ s S

By using centrifugal force

mB ⋅ v 2
FLat,B = (1.6-6)
r

arises

2 ⋅ ∆h ⋅ m B ⋅ v 2
ϕ= 2
(1.6-7)
(c B + c Stab ) ⋅ s S ⋅ r

With the equation

1 2 ⋅ cB
feB = ⋅ (1.6-8)
2π mB

For the body eigenfrequency of the stroke oscillation, resolved to mB

2 ⋅ cB
mB = 2
(1.6-9)
4π 2 ⋅ feB

The roll angle is given by

2 ⋅ ∆h ⋅ v 2 ⋅ 2 ⋅ c B
ϕ= 2
(1.6-10)
(c B + c Stab ) ⋅ s S ⋅ r ⋅ 4 ⋅ π 2 ⋅ f eB

and by solution feB is given by

v 1 1 ∆h 1
feB = ⋅ ⋅ ⋅ ⋅ (1.6-11)
π r ϕ sS
2
c
1 + Stab
cB
106 Two-Track Suspension Model

Note c B = c Bf + c Br , where c Bf or c Br r is the spring rate of one spring on the front or rear
axle.

The first three factors of this product depend on the respective driving condition. If a standard
case is assumed, which is characterized by a curve radius r = r0 = 50 m and a driving speed
v = v0 = 16 m/s, then the recalculation shows - in agreement with subjective feeling - that for
this case, a roll angle of ϕ = ϕ0 = 3° is acceptable. With this standardisation the body
eigenfrequency of a vehicle still depends on the various factors related to the design of the
vehicle, i.e. on the dimensions ∆h and sS as well as the relation of the spring constants cStab
and cB.

This relationship is represented in the lower section of Fig.1.6-5.

m r = 50 m
s2 v = 16 m/s
ϕ = 3° v = 20 m/s
Beff

1.0
..
Body-acceleration z

0.5
body-
eigenfrequency f e
0
0.5 1.0 1.5 2 s -1 2.5
1

4
s2
S m
∆h 5
C Stab
= 1 0.5 0.2 0
C
B

Fig.1.6-5: Influence of the height of the centre of gravity and spring track width on body
natural frequency and effective value of body acceleration

In the upper part of Fig.1.6-5, the relationship between body eigenfrequency and body
acceleration is represented for a vehicle with two axles at v=20 m/s. The characteristic curve
is well-known from fig. 1.5-3 for the single track model of the exemplary passenger car.

1.6.2 Distortion of the structure (twisting)

The closed bodies of vehicles such as sedans, transporters and buses can be assumed to
be rigid in terms of the handling of the vehicle suspension system, since the bending or
torsional rigidity of the body is much higher here than the rigidity in the suspension system.

In contrast, the torsional rigidity of the chassis in commercial vehicles with ladder-type
frames, Fig.1.6-6, can not be neglected. The frame is designed to be torsionally elastic here,
in order to reduce the stress of the material in the interconnecting points between transverse
Two-Track Suspension Model 107

and longitudinal members, i.e. riveting is used instead of welding at the joints and open
profiles are used instead of closed profiles for longitudinal and transverse beams.

Fig.1.6-6: Torsionally elastic framework of a light truck

For commercial vehicles used on construction-sites or off-road torsionally elastic framework

offers the advantage that larger surface irregularities can be absorbed without loss of traction
(wheels lifting off the surface of the road) than with sole use of the spring travel offered by
the body springs.

A computational handling of the twisting of the framework is relatively simple, if one divides
the body into two subsystems with individual centres of gravity, Fig.1.6-7.

m Br
∆ hr
c Br

m Bf s
Sr
c tor
∆ hf
c Bf

s Sf

Fig.1.6-7: Replacement system of a vehicle with a torsionally elastic framework

For steady-state cornering, the equilibriums of torques set up around the roll axis for front
and rear ends separately result in two equations for the two unknown roll angles ϕf and ϕr:
108 Two-Track Suspension Model

2
s Sf
mBf ⋅ a y ⋅ ∆h f = ϕ f ⋅ ⋅ c Bf + (ϕ f − ϕr ) ⋅ c tor (1.6-12)
2

2
s
mBr ⋅ a y ⋅ ∆hr = ϕr ⋅ Sr ⋅ c Br + (ϕr − ϕ f ) ⋅ c tor (1.6-13)
2

Since the difference between spring forces on one axle depends on the roll angle of the
appropriate body sub-system, the proportion of the total roll torque supported at the regarded
axle in vehicles with torsionally elastic chassis is influenced not only by distribution of the
body spring stiffness and stabilizer stiffness on the front and rear axles, but also by the
torsional rigidity of the framework. Accordingly, the wheel load differences during cornering at
the front and the rear axle depend additionally on the torsional rigidity of the framework.

The influence of these wheel load differences on the steering behavior during cornering is
treated in the section `Lateral Dynamics' in connection with the effects of stabilizer springs.

1.6.3 Rigid axle Tramp

Fig. 1.6-8 shows an equivalent model for the investigation of the vibrational behavior of rigid
axles.

Fig. 1.6-8: Axle equivalent system

Rigid axle excitation not only results in the vertical motion (bounce) discussed so far but also
in tramp motion due to irregular road roughness. The two forms of motion are usually (with
left/right axle symmetry) decoupled. Approximate vertical motion natural frequencies are
obtained from:
Two-Track Suspension Model 109

cBody + c W cW
ωe rec = ≈ (1.6-14)
mW mW

Only if the theoretical case is assumed that 2 mKR = 0, i.e. the mass of the entire axle is
concentrated on the two wheels, will also apply to tramp motion (axle tramp).

2 (s / 2 ) 2 ⋅ c W 2 (s / 2 ) 2 ⋅ c W
ωe Tr ≈ ≈
Θ 2 (s / 2 ) 2 ⋅ m W

cW
ωe Tr ≈ (1.6-15)
mW

In reality, tramp natural frequency exceeds vertical motion natural frequency, with the latter
being in the range of fe = 9 - 14 s-1 (with tandem axles and powered truck front axles also
lower).

Analogous investigation of damping D reveals that with tramp motion it is lower than with
vertical motion. Rigid axle tramp would therefore result in higher dynamic wheel loads than
vertical motion, provided the excitation of both forms of motion by the roadway is of about the
same magnitude. Usually, however, we find the (symmetric) vertical motion excitation to be
dominant with normal roadways.

The resulting dynamic wheel loads of the rigid axle are higher than for independent wheel
suspension. Moreover, the rigid axle is heavier.
110 Suspension Investigation Methods

1.7 Suspension Investigation Methods

The term riding comfort usually characterizes the vibrational comfort of a motor vehicle. The
low-frequent comfort impression in an automobile is not only the result of vibration excitation
of the automotive body due to road roughness but also of irregularities (wheel, tire) and the
gas and inertia forces of the engine, Fig. 1.7-1.

passengers

body
seat floor steering wheel

chassis

engine wheel
rubber bearings
bearing suspension

chassis subframe

rubber bearings rubber bearings rubber bearings

damper spring links

engine wheel
Motor wheel

tire

excitement single stochastic imbalance

obstacles unevenness
gas- and mass-
road unloaded radial
forces in engine
unevenness tire run-out

Fig.1.7-1: Paths of transmission of vibration excitation to the vehicle occupant

The forms of excitation listed cause a multitude of vibration phenomena perceived by the
vehicle occupant via floor, seat and steering wheel (driver) as body, chassis and engine
suspension vibration, shaking, jerking, or in the form of shocks (bumps).

Fig. 1.7-2 provides a more detailed overview of vibration phenomena determining riding
comfort. In addition to information about the frequency ranges involved, the excitations are
shown as well as the influences of individual assemblies of the vehicle on the vibration
Suspension Investigation Methods 111

phenomena.

influence of parameters
excitement

suspension/ damping
great influence

run-out (wheel, brake)

gas- and mass-forces

steering assembly
road unevennesses
unloaded radial tire

engine bearings
small influence

within engine

drive train
front axle

rear axle

exhaust
engine
cabin
body
vibration/ frequency
noise phenomenon range

body vibrations 1- 2
lateraljerking 4 - 10
abrupt load alteration 8 - 20
engine shake 10 - 13
axle vibrations 10 - 15
wheelfight 10 - 20
brake judder 15 - 25
idle shake/ drone 20 - 30
steering wheel shake 20 - 35
body shake 20 - 35
body drone 25 - 75
axle roughness 30 - 80
drive roughness 50 - 100
exhaust vibrations 50 - 200
driving noise 50 - 250
engine resonances 100 - 250

Fig.1.7-2: Vibrational phenomena / 27 /

A high level of riding comfort requires a low level of disturbing vibrations, shaking, jerking or
bumping (shocks).

Various methods of measuring and analysis are employed for the design and tuning of
vehicle suspensions. On the following pages, the most important methods will be described
and their advantages and disadvantages will be discussed.

• Calculating model

Simple calculating models, such as the single-wheel suspension model or the single-track
suspension model, were comprehensively dealt with in the preceding chapters. With modern
computer aid, significantly more complex models can be generated, which produce results of
much greater precision. Fig. 1.7-3 , for example, shows the FE simulation model of a
complete vehicle. This model not only considers the dynamic properties of chassis and drive
train but also the deformation behavior of the body.
112 Suspension Investigation Methods

Fig.1.7-3: FE complete-vehicle simulation model of the Audi A8

In addition to complete vehicles, also individual vehicle assemblies are subjected to

simulation studies. Fig. 1.7-4 shows an example of FE modeling of a front end with steering,
wheel suspension, axle bracket and wheels.

Fig.1.7-4: FE modeling of an automobile front end

Such a model can be used to define remedial measures against the vibration form wheelfight
(niggle) (cf. Fig. 1.7-2), which can be excited by unloaded radial tire run-out or wheel
imbalance.

Nevertheless, even very complex models involve a multitude of simplifications which

significantly affect the precision of calculating models and under certain circumstances may
even result in incorrect statements as to trends. The major benefits of calculating models are
represented by the fact that they can be employed at a very early stage of development and
that their transparence facilitates analysis and the implementation of target-specific
improvement measures.
Suspension Investigation Methods 113

• Excitation on the hydraulic pulse test rig

While computer models are applicable any time during vehicle development, vibrational
analysis based on a measurement on the hydraulic pulse usually requires a complete
vehicle. Thereby, the excitation can be done in form of a ‘real’ or simulated road unevenness
profile, or a sinusoidal signal. Fig.1.7-5 shows the testing of an ambulance on the hydraulic
4-post test rig at the ika.

advantages:
- good insight into complete suspension
system (analysis and diagnosis)
- reproducibility of measuring results
- shortening of measuring time by enlarging
amplitudes for endurance testing possible
- short conversion time
- vehicle does not have to be in running
condition

disadvantages:
- components must already exsist
- considerable effort with equipment KFZ II 1-7 eng.pp

Fig.1.7-5: Hydraulic 4-post test rig

While calculation methods can be employed at any stage of vehicle development, vibration
analysis by means of hydropulse testing as a rule requires a finished vehicle. Excitation can
be effected by means of a "real" as well as by means of a

simulated road spectrum, or by means of a sinusoidal signal. Fig. 1.7-5 shows the testing of
a motorized ambulance unit by means of the 4-ram hydropulser of the ika.

The lab environment of hydropulse measurement offers a number of advantages:

- good insight into the overall suspension system (analysis and diagnosis),

- reproducibility of measuring results,

- accelerated endurance testing through amplitude magnification,

- fast conversion,

- vehicle does not have to be ready for driving.

114 Suspension Investigation Methods

Problems are regularly encountered in hydropulse testing if the influence of the rolling wheel
and the longitudinal dynamics in the chassis play a role, Fig. 1.7-6.

change in
dynamic
spring rate
variation of
dynamic roll
radius

displacement of contact area

of wheel due to lateral forces

- Spring rate and damping factor of rolling wheel are

distinctively different from those of a resting wheel.
- Variation of the dynamic roll radius due to vertical
impression of the tire causes rotational accelerations of
the wheel and therefore lateral forces in the contact area.
- Unevennesses lead to local incline and therefore to a
displacement of the wheel-contact area. This also causes
lateral forces within the suspension.

Fig.1.7-6: Effects of the rolling wheel that cannot be simulated on the hydropulser

- Stiffness and damping of the rolling wheel differ greatly from the stiffness and damping of
the stationary wheel.

- Variation of the dynamic rolling radius as a result of vertical deformation of the wheel
causes rotational acceleration of the wheels and thus longitudinal forces in the tire-contact
center (footprint center).

- Unevenness results in local grades and thus in displacements of the tire-contact center
(footprint center). This effect also causes additional longitudinal forces in the chassis.

The considerable amount of equipment required and the fact that at least one prototype has
to be available are also disadvantages of hydropulse testing.
Suspension Investigation Methods 115

• Driving tests

Subjective assessment of the suspension behavior of a vehicle by means of an on-road test

is the most obvious and still also the most important method, because the driver can obtain
an immediate impression of the behavior of the vehicle.

The drawback of this method is the fact that assessment is only possible after design and
construction of at least one prototype, i.e. at an advanced stage of development that usually
allows only details to be changed. Moreover, analysis of a suspension evaluation composed
of many individual effects is very difficult and constitutes a poor basis for specific
improvement measures. On-highway testing is nowadays usually only used for fine-tuning at
a late stage of development.

Fig. 1.7-7 summarizes the methods discussed

technique mathematical hydropulser driving test

model
statement of certain effects are identical with
subjective neglected, reality,
assessments problems can be use of appropriate
clearly approached roads necessary

relevance of only as good as can be complicated good

measuring results the models used under certain
circumstances

reproducibility good medium

representative good within the good can be complicated

for limits of the model due to
analysis superposition of
both axles

Fig.1.7-7: Summary of methods discussed

internet
www.ika.rwth-aachen.de

5zb0411.cdr-pdf
www.fka.de