Professional Documents
Culture Documents
www.elsevier.com/locate/apacoust
Received 20 November 2000; received in revised form 18 May 2001; accepted 14 June 2001
Abstract
Brake squeal, which usually falls in the frequency range between 1 and 16 kHz, has been one of
the most difficult concerns associated with automotive brake systems since their inception. It cau-
ses customer dissatisfaction and increases warranty costs. Although substantial research has been
conducted into predicting and eliminating brake squeal since the 1930s, it is still rather difficult to
predict its occurrence. In this paper, the characteristics and current difficulties encountered in
tackling brake squeal are first described. A review of the analytical, experimental and numerical
methods used for the investigation of brake squeal is then given. Some of the challenges facing
brake squeal research are outlined. # 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
Brake squeal has been one of the most difficult concerns associated with auto-
motive brake systems since their inception. Research into predicting and eliminating
brake squeal has been conducted since the 1930s [1,2]. Initially drum brakes were
studied due to their extensive use in early automotive brake systems. However, disc
brake systems are used more extensively in modern vehicles and have become the
focus of brake squeal research.
Figs. 1 and 2 show a typical disc brake system with a ‘‘fist type’’ caliper design. A
disc brake system consists of a rotor that rotates about the axis of the wheel. The
caliper assembly is mounted to the vehicle suspension system through an anchor
* Corresponding author.
E-mail address: z2278180@octarine.itsc.adfa.edu.au (A. Papinniemi).
0003-682X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0003-682X(01)00043-3
392 A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400
bracket. The caliper housing can slide on the anchor bracket through the two pins.
Brake pads with moulded friction material can also slide on the anchor bracket. A
piston can slide inside the caliper housing. When hydraulic pressure is applied, the
piston is pushed forward to press the inner pad against the rotor and in the mean
time, the housing is pushed in the opposite direction to press the outer pad against
the rotor, thereby generating a braking torque.
Like all the other applications with friction interface, noise and vibration are
inherent by-products of brake application. Brake noise and vibration has been clas-
sified according to its frequency as judder, groan, hum, squeal, squelch and wire
A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400 393
brush [3]. The squeal noise that is particularly annoying usually falls into a fre-
quency range from 1 to 16 kHz.
Brake squeal is generated by the vibration of an unstable vibration mode of the brake
system. In this condition the brake rotor can act as a loudspeaker since it has large flat
surfaces that can readily radiate sound. The occurrence of brake squeal is a concern
since it causes significant discomfort to the vehicle occupants and leads to customer
dissatisfaction and increased warranty costs. Unfortunately, the large body of research
into brake squeal has failed to provide a complete understanding of, or the ability to
predict its occurrence [1–26]. This is partly because of the complexity of the
mechanisms that cause brake squeal and partly because of the competitive nature of
the automotive industry, which limits the amount of cooperative research that is
published in the open literature.
Although a comprehensive review of brake squeal was conducted by Yang and Gib-
son in 1997 [4], it was focussed to some degree on the material aspects of a brake system.
The objective of this paper is to outline the characteristics and current difficulties
encountered in tackling brake squeal and to review the analytical, experimental and
numerical methods used for the investigation of brake squeal.
One of the biggest contributors to brake squeal is the friction material, since
squeal excitation occurs at the friction interface, and it normally takes approxi-
mately 12 months to finalise a friction material selection. This certainly makes it
very difficult to predict a priori the propensity of a brake system to squeal. Also,
often in the design of a brake system, priority is given to requirements such as
braking performance, cost and ease of manufacture. The common practice for the
different components of a brake system to be manufactured by different suppliers
further complicates matters. The large number of vehicles produced means that even
a low squeal propensity found during initial testing of a brake system can become a
major concern once a vehicle is in production due to a much larger population size.
Modifications towards the end of development phase will have two potential risks:
(1) leading to production delays and increased costs to both the brake and vehicle
manufacturers and (2) leading to products not fully validated with potential field
warranty concern.
The most significant complication in brake research is the fugitive nature of brake
squeal; that is, brake squeal can sometimes be non-repeatable. There are many
potential squeal frequencies (unstable modes) for a brake system. Each individual
component has its own natural modes. The number of modes for a rotor within
human hearing range may be up to 80. The modal frequencies and modal shapes of
the rotor, caliper, anchor and pad will change once these parts are installed in-situ.
During a brake application, these parts are dynamically coupled together resulting
in a series of coupled vibration modes, which are different from the component free
vibration modes. The addition of the friction coupling forces at the friction interface
results in the stiffness matrix for the system containing unsymmetric off-diagonal
394 A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400
coupling terms. From the stability point of view, this coupling is considered to be
the root cause of the brake squeal. A brake system may not always squeal given the
‘‘same’’ conditions. Alternatively, small variations in operating temperature, brake
pressure, rotor velocity or coefficient of friction may result in differing squeal pro-
pensities or frequencies. Figs. 3 and 4 show the percentage occurrence of brake
squeal obtained at PBR Automotive Pty Ltd using a Rubore drag type noise
dynamometer and an AK noise matrix for various brake pressures and temperatures
respectively. It can be seen from Fig. 3 that there is no simple relationship between
the percentage occurrence and frequency of the brake squeal and the brake pad
pressure. Similarly, the influence of temperature on both the occurrence and fre-
quency of the brake squeal is quite complex (Fig. 4).
Due to the above-mentioned difficulties in designing a noise free brake system,
efforts to eliminate brake squeal have largely been empirical, with problematic brake
systems treated in a case by case manner. The success of these empirical fixes
depends on the mechanism that is responsible for causing the squeal problem. The
most fundamental method of eliminating brake squeal is to reduce the coefficient of
friction of the pad material [5–7]. However, this obviously reduces braking perfor-
mance and is not a preferable method to employ. The use of viscoelastic material
(damping material) on the back of backplate can be effective when there is sig-
nificant pad bending vibration [8,9]. Changing the coupling between the pad and
rotor by modifying the shape of the brake pad has also been found effective [10,11].
Other geometrical modifications that have been successful include modifying caliper
stiffness [12,13], the caliper mounting bracket [14,15], pad attachment method [16]
and rotor geometry [17,18].
Fig. 3. Variation of occurrences of brake squeal with frequency and brake pad pressure.
A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400 395
The earliest research into brake squeal suggested that the variation in the friction
coefficient with sliding velocity was the cause [19]. Not only is there a difference
between the static and dynamic coefficient of friction, but it was thought the drop in
kinetic friction with increased sliding velocity could lead to a stick-slip condition and
produce self-excited vibration. However, squeal has been shown to occur in brake
systems where the coefficient of kinetic friction is constant [20], and has led to ana-
lysis of the geometrical aspects of a brake system.
Spurr proposed an early sprag-slip model that describes a geometric coupling
hypothesis in 1961 [6]. Consider a strut inclined at an angle to a sliding surface as
shown in Fig. 5(a). The magnitude of the friction force is given by
L
F¼
1 tan
where is the coefficient of friction and L is the load. It can be seen that the friction
force will approach infinity as approaches cot . When =cot the strut ‘sprags’
or locks and the surface can move no further. Spurr’s sprag-slip model consisted of a
double cantilever as shown in Fig. 5(b). Here, the arm O0 P is inclined at an angle 0
to a moving surface. The arm will rotate about an elastic pivot O0 as P moves under
the influence of the friction force F once the spragging angle has been reached.
Eventually the moment opposing the rotation about O0 becomes so large that O00 P
replaces O0 P, and the inclination angle is reduced to 00 . The elastic energy stored in
O0 can now be released and the O0 P swings in the opposite direction to the moving
396 A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400
The frequencies of a squealing brake are highly dependent on the natural fre-
quencies of the brake rotor [17]. Consequently it is of fundamental importance to be
able to determine the vibration modes of the rotor. Not only will an understanding
of the vibration modes of the rotor help predict how a brake system may vibrate,
but it is also necessary in developing countermeasures to eliminate the problem. The
existence of in-plane modes in addition to the bending modes is a further complica-
tion, and there is evidence that the in-plane modes can be the cause of some type of
brake squeals as well as the bending modes [18].
Fig. 5. (a) Single strut rubbing against moving surface; (b) sprag-slip system.
A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400 397
Fig. 6. (a) Experimental bending mode shape; (b) FEA bending mode shape.
398 A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400
Finite element analysis (FEA) has been used in the analysis of brake squeal. Modal
analysis of brake components is an area where FEA can be readily applied. Fig. 6(b)
shows a finite element model of a brake rotor. The model, consisting of 8700 Tet92
solid elements, was developed using a commercial finite element code ANSYS 5.6.
Unfortunately, the coupling between brake components leads to vibration modes that
differ to those found for the individual components. Therefore, the real interest
among researchers is to be able to model an entire brake system.
The critical aspect in the modelling of a complete brake system is the coupling
between components, particularly the rotor/pad interface. The contact stiffness itself is
adjusted using experimental results, but the more difficult aspect is to introduce the
tangential friction coupling. Liles included friction coupling between rotor and pad as
off diagonal terms in the stiffness matrix and used a complex eigenvalue analysis to
assess the stability of a brake system [5]. Once the model was developed, the effect of
varying parameters such as friction coefficient, pad geometry and caliper stiffness,
could be determined. Dihua and Dongying also used a similar approach to improve
the design of an anchor bracket [14]. The work of these, and other, researchers has
shown that it is possible to create models that incorporate the friction coupling
between the rotor and the pad. However, there has been little experimental evidence to
verify the accuracy of these models. They may be useful for studying the effect of
varying parameters within the brake system, but their ability to model the important
friction interface is limited. As small variations in operating temperature, brake pres-
sure, rotor velocity or coefficient of friction may result in differing squeal propensities
or frequencies (Figs. 3 and 4), an accurate prediction of brake squeal using numerical
methods requires an accurate determination of material properties (particularly for the
friction material) under different operating conditions. Furthermore, proper model-
ling of the boundary conditions especially where the coupling between various
components is important remains a challenge.
Presently, research into brake squeal is focused on specific brake systems or genera-
tion mechanisms. The challenge for the future is to be able to develop general techni-
ques and guidelines to eliminate brake squeal during the design stage. Given the
complexity of the mechanisms that generate brake squeal, it appears that general
guidelines are some way off in the future. For the present, the reduction of squeal noise
for specific brake systems is achievable, with the additional knowledge acquired in
each case adding to the overall understanding of brake squeal.
Theoretical analysis of brake systems is difficult given the complexity of the
mechanisms and the lack of an adequate model for the friction interface that causes
brake squeal. However, this should not limit the development of simplified models as
valuable insight can be gained. Understanding obtained by studying simplified
models can assist in the interpretation of experimental results and the development
of improved computational tools.
A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400 399
The application of FEA to brake squeal appears to offer some promise. Com-
mercial software packages are being continually refined with improved modelling
features and the friction coupling capabilities are improving. The rapid development
in computer aided engineering systems should make it feasible to analyse every
aspect of a brake system from braking performance to vibro-acoustic analysis, thus
allowing brakes to be designed with minimum propensity to squeal and desirable
braking performance.
Experimental methods will still play an important role for a number of reasons.
Firstly, they offer more effective analysis tools than numerical or purely theoretical
methods. Secondly, diagnosis of the cause of brake squeal problems can often only
be found by experimentation. Finally, the verification of solutions to squeal pro-
blems, and the applicability of FEA models, can only be achieved through experi-
mental means. Ultimately the future elimination of brake squeal will be confirmed
though experimental results and the final testing of brake systems.
Acknowledgements
This study forms part of a project funded by the Australian Research Council
under the SPIRT scheme and the industry partner is PBR Automotive Pty Ltd.
References
[1] Lamarque PV, Williams CG. Brake squeak: the experiences of manufacturers and operators and
some preliminary experiments. Research report no. 8500 B. The Institution of Automobile Engi-
neers, 1938.
[2] Mills HR. Brake squeak. Research report nos. 9000 B (1938) and 9162 B. The Institution of Auto-
mobile Engineers, 1939.
[3] Lang AM, Smales H. An approach to the solution of disc brake vibration problems. Institute of
Mechanical Engineering, C37/83, 1983, p. 223-31.
[4] Yang S, Gibson RF. Brake vibration and noise: reviews, comments, and proposals. Int J of Materials
Product Technol 1997;12(4-6):496–513.
[5] Liles GD. Analysis of disc brake squeal using finite element methods. SAE Paper, No. 891150, 1989.
[6] Spurr RT. A theory of brake squeal. I Mech E Auto Div Proc No. 1, 1961/62. p. 30-40.
[7] Bergman F, Eriksson M, Jacobson S. Influence of disc brake topography on generation of brake
squeal. Wear 1999;225-229:621–8.
[8] Hoffman CT. Damper design and development for use on disc brake shoe and lining assemblies.
SAE paper No. 880254, 1988.
[9] Lewis TM, Shah P. Analysis and control of brake noise. SAE paper No. 872240, 1987.
[10] Flint J. Instabilities in brake systems. SAE paper No. 920432, 1992.
[11] Nossier TA, Said MAR, El Nahas NS, Abu El Fetouh G. Significance of squeal in disc brake design.
Int J of Vehicle Design 1998;19(1):124–33.
[12] Matsui H, Murakami H, Nakanishi H, Tsunda Y. Analysis of disc brake squeal. SAE paper No.
920553, 1992.
[13] Murakami H, Tsunada N, Kitamura T. A study concerned with the mechanism of disc brake squeal.
SAE paper No. 841233, 1984.
[14] Dihua G, Dongying J. A study on disk brake squeal using finite element methods. SAE paper No.
980597, 1998.
400 A. Papinniemi et al. / Applied Acoustics 63 (2002) 391–400
[15] Baba H, Okade M, Takeuchi T. Study on reducing low frequency brake squeal from modal analysis
of mounting bracket. SAE paper No. 952697, 1995.
[16] Fieldhouse JD, Newcomb TP. An investigation into disc brake noise using holographic inter-
ferometry. Institute of Mechanical Engineering, C427/11/213, 1991.
[17] Nishiwaki M, Harada H, Okamura H, Ikeuchi T. Study on disc brake squeal. SAE paper No.
890864, 1989.
[18] Matsuzaki M, Izumihara T. Brake noise caused by the longitudinal vibration of the disc rotor. SAE
paper No. 930804, 1993.
[19] North MR. A survey of published work on vibration in braking systems. MIRA Bulletin No. 4,
1969.
[20] Hulten J. Brake squeal—a self exciting mechanism with constant friction. SAE Truck and Bus
Meeting, Detroit, MI, USA, 1993.
[21] Jarvis RP, Mills B. Vibrations induced by dry friction. I Mech E Proc 1963/64;178(1/32):847–57.
[22] Earles SWE, Soar GB. Squeal noise in disc brakes. Paper C101/71. I Mech E Symposium on Vibra-
tion and Noise in Motor Vehicles, 1971.
[23] North MR. Frictionally induced, self excited vibrations in a disc brake system. PhD thesis, Lough-
borough University of Technology, 1972.
[24] Millner N. An analysis of disc brake squeal. SAE paper No. 780332, 1978.
[25] Felske A, Hoppe G, Matthai H. Oscillation in squealing brakes—analysis of vibration modes by
holographic interferometry. SAE paper No. 780333, 1978.
[26] Fieldhouse JD, Newcomb TP. Double pulsed holography used to investigate noisy brakes. Optics
Lasers Engng 1996;25:455–94.