Wear 252 (2002) 26–36 On the nature of tribological contact in automotive brakes Mikael Eriksson, Filip Bergman, Staffan Jacobson∗ Materials Science Division, The Ångstrom Laboratory, Uppsala University, Box 534, 751 21 Uppsala, Sweden Received 1 November 1999; received in revised form 30 August 2000; accepted 15 August 2001 Abstract The tribological contact in automotive brakes involves dry sliding contact at high speeds and high contact forces. The commonly used organic binder-type brake pad friction materials are extremely inhomogeneous and exhibit very low bulk strengths. Despite the low strength, the specific contact surfaces that form during the use render the pads very good friction and wear characteristics. This paper gives a general view of the contact situation of organic binder brake friction materials against cast iron discs, with special emphasis on many mechanisms for contact surface variations and the corresponding variations of the coefficient of friction. © 2002 Published by Elsevier Science B.V. Keywords: Tribological contact; Organic binder; Plateaus 1. Introduction In the automotive industry, the friction system of brake pad against a cast iron disc or drum has an enormous technical significance. Yet, the basic knowledge about the contact and friction mechanisms on a microscale of this system is limited and the number of scientific publications on the subject is small. The demands on the friction behaviour are high and manifold. The friction coefficient should be relatively high but most importantly stable. It should keep a stable level irrespective of temperature, humidity, age, degree of wear and corrosion, presence of dirt and water spraying from the road, etc. In addition to these safety requirements come the requirements for long life and high comfort, i.e. absence of vibration and squeal noise. The tremendous technical importance of this friction system and its significant deviations from most other tribological contact situations motivates a study on the particular nature of the tribological contact in automotive brakes. Further, this study also has bearings on brake squeal. In a review on brake noise and vibration, Crolla and Lang  stated that the key to further progress in analysing the brake noise problem and its fruitful application to real brake design lies in developing better finite element models and, in particular, in finding an accurate way of representing the frictional coupling terms at the rubbing interface. ∗ Corresponding author. Tel.: +46-18471-3088; fax: +46-18471-3572. E-mail address: [email protected] (S. Jacobson). 0043-1648/02/$ – see front matter © 2002 Published by Elsevier Science B.V. PII: S 0 0 4 3 - 1 6 4 8 ( 0 1 ) 0 0 8 4 9 - 3 The objective of the present paper is thus to increase the general understanding of the friction situation in a microscale, based on experiments in a disc brake rig. The experimental procedures and the rig, comprising a complete front wheel (wheel, brakes, suspension) driven by an electric motor have been presented in some detail in . 1.1. The contact characteristics of automotive brakes The main purpose of a car brake is to reduce the speed. In this process, the kinetic energy is transformed into heat by the frictional work. Car brakes experience dry sliding contact at roughly 50% of the speed of the car. A typical front brake pad is about 8 cm long and 5 cm wide and the brake disc has a diameter of 28 cm. The pad covers around 10–15% of the corresponding rubbing surface of the disc. During normal, relatively soft brakings the force pressing the pad against the disc is about 5 kN, resulting in a nominal pressure at the pad surface just above 1.2 MPa. In extreme situations, the pressure could be close to 10 MPa. During hard brakings, the power dissipation on a brake pad easily exceeds 30 kW. These high power densities result in very high surface temperatures and thus put special demands on the friction materials. 1.2. Typical composition of organic brake pad friction materials The lining materials of automotive brakes are usually composites formed by hot compaction of coarse powders M. Eriksson et al. / Wear 252 (2002) 26–36 including many different components (typically 10–20). These components include a: • binder, that holds the other components together and forms a thermally stable matrix. Thermosetting phenolic resins are commonly used, often with the addition of rubber for increased damping properties; • structural materials, providing mechanical strength. Usually metal, carbon, glass, and/or kevlar fibres are used and more rarely different mineral and ceramic fibres; • fillers, mainly to reduce cost but also to improve manufacturability. Different minerals such as mica and vermiculit are often employed. Barium sulphate is another commonly used filler; • frictional additives, added to ensure stable frictional properties and to control wear rates of both pad and disc. Solid lubricants such as graphite and different metal sulphides are used to stabilise the coefficient of friction, primarily at elevated temperatures. Abrasive particles, typically alumina and silica, increases both the coefficient of friction and the disc wear. The purpose of the latter is to give a better defined rubbing surface by removing iron oxides and other undesired surface films from the disc. 1.3. Materials in the brake discs The material in the discs is usually pearlitic grey cast iron with 3–4 wt.% carbon. This material contains free graphite in the shape of small flakes in a pearlitic matrix. Besides having desirable thermal properties, grey cast iron has sufficient mechanical strength, satisfactory wear resistance, good damping properties, it is cheap, and it is also relatively easy to cast and to machine. 1.4. Limitations of the present investigations The present paper is based on investigations of a limited number of organic binder pad materials only. The nature of tribological contact for materials outside this narrow selection may be different. 27 All observations are part of a project on brake squeal generation. As brake squeal usually occurs at low speeds and at moderate brake pressures the test equipment was designed primarily for this application. This means that the brakes have only been exposed to braking under conditions of relatively low speeds (0–25 km/h) performed in lab atmosphere with ambient humidity and almost constant temperature (20–25 ◦ C). The limited top speed puts a restriction to the attainable brake pad histories. The history of the brake pad is shown to be important for the friction properties and thus its tendency to generate squeal. 2. General view of the contact situation During braking, the pad surface is in continuous contact against the disc, while each part of the disc surface experiences intermittent contact. As for all dry sliding systems, the friction force is transmitted via the area of real contact. Conventional lining materials are relatively compliant. Despite this, only a limited fraction of the pad surface is in contact with the disc. This area of real contact is confined within a number of contact plateaus, scattered over the pad surface, (see Fig. 1). These plateaus can often be observed by the naked eye as shiny spots against a duller background. In a microscope or with a 3-D profilometer it can be observed that the plateaus are relatively flat, often rising a few microns over more irregular surroundings, (see Fig. 2). The plateaus show signs of wear and sliding contact, often including a pattern of parallel grooves in the direction of sliding, (see Fig. 3). The area of real contact is small compared to the size of the plateaus. The location and size of the points in real contact will fluctuate irregularly due to the associated deformation and wear and due to the passing countersurface (i.e. the disc) not being perfectly flat and smooth. The contact plateaus are typically 50–500 m wide and (when inspected after unloading) they rise typically a few Fig. 1. General view of the brake pad area and its division into contact plateaus and areas of real contact within the contact plateaus. 28 M. Eriksson et al. / Wear 252 (2002) 26–36 Fig. 2. Appearance of a used brake pad surface showing contact plateaus and the surrounding rougher surface. The same area viewed with different techniques: (a) 3-D profilometric micrograph (white light interference profilometry); (b) SEM in secondary electron mode; (c) SEM in back scatter electron mode. The central steel fibre appears light. microns over the surroundings. This size range can be compared to 25–100 m that are typical diameters of the hard lining components. Measurements show that the contact plateaus typically exhibit hardness values considerably higher than the mean hardness of the pad composite (e.g. 3000 MPa compared with 200 MPa). This is due to the plateaus typically being formed by metal fibres surrounded by compacted softer constituents, as shown in Fig. 2c. The total number of plateaus is on the order of 105 and they typically constitute 15–20% of the pad area. The average pressure on the plateaus is thus some five times higher M. Eriksson et al. / Wear 252 (2002) 26–36 Fig. 3. The grooved appearance often found on the contact plateaus (SEM). than the average pad pressure. Still these pressures (typically in the range 6–50 MPa) are very low compared with the hardness of the harder components of the pad. Thus, the total load on the pad can be carried by a very small fraction of the total plateau area. 2.1. Formation and disintegration of the contact plateaus The lining components show very different wear properties. Steel fibres (and other structural components and abrasive particles) have a relatively high resistance to sliding wear against the disc. Thus, following a short running-in, the metal fibres will stand slightly higher than less wear resistant constituents, forming the central part of the contact plateaus. During steady state wear, the more wear resistant constituents will carry a large proportion of the load [3,4], and will in this way protect the adjacent material. The higher proportion of the load carrying will subsequently lead to the wear resistant components having a higher influence on the coefficient of friction than indicated by their volume fraction . Therefore, despite softer components filling the majority of the lining volume, the contact situation is dominated by metal to metal contact with respect to both friction and wear. The topography composed of rather flat contact plateaus rising over irregular surroundings must be the result of both 29 Fig. 5. Outlines of the contact plateaus on the same area of a pad, studied after 1 and 1000 m sliding at low pressure against a cast iron disc, investigated in a light optical microscope. Many of the contact plateaus remain, with small changes in shape and size. removal of fragments and collection of material around the plateaus. During steady state wear, the linear removal rate must be the same on top of the contact plateaus as on the surrounding “low lands”. The lower areas show no signs of sliding contact. It is thus believed that the removal of material from these regions is due to either a mechanical “crushing” action performed by wear debris trapped between the pad and the disc, or due to fragmentation preceded by mechanical degradation and decomposition of the organic pad constituents. The removal of organic binder material may even be due to emission of CO, CO2 or other gaseous decomposition products. The deterioration processes are accelerated by the repeated temperature cycling, sometimes up to very high temperatures in combination with occasional tribological contact. The contact plateaus can grow by processes involving agglomeration and compaction of pad wear debris around the nucleation site , (see Fig. 4). The tendency for the debris to stick to the plateau depends on the temperature, the humidity and also on the prevailing state of shear and normal pressures. The contact plateaus may also shrink due to disintegration and removal of the compacted debris caused by changes in Fig. 4. Schematic drawing of the mechanism proposed for slow contact area growth due to accumulation and compaction of debris. 30 M. Eriksson et al. / Wear 252 (2002) 26–36 the contact situation. The contact plateau will finally disintegrate when the protective fibre or particle has been worn away. However, the plateaus exhibit a relatively long life. When performing repeated inspection of a specific area of the pad, the same plateaus could be recognised, with only slow changes of shape and size, (see Fig. 5). 3. Contact surface variations The size of the area of real contact between the pad and the disc, and also the composition of the outermost surface layers within this area, is far from constant but will vary due to changing pressure, changing temperatures, deformation and wear. The contact pressure may vary on different time scales and both locally and globally, due to different processes. 3.1. Rapid global processes Naturally, a change in the braking force will result in a corresponding change of the elastic compression of the pad. The actuation and variation of the braking force can be quick due to manual “fine tuning” during braking or due to ABS brake power variations, etc. The associated pressure changes are global over the brake pad (the average compression of the pad varies) and on a time scale of 1/10 s. Ideally the resulting braking power should be proportional to the pedal force. A quick brake pressure increases thus momentarily results in a corresponding elastic compression of the pad. This compression will result in: • more contact plateaus becoming engaged, as shown in Fig. 6a; • an increased load on the already engaged contact plateaus, which will result in a higher area fraction of real contact within these plateaus, (see Fig. 6b). 3.2. Rapid local processes The pad compression, and hence the pressure, may also vary locally over the pad surface. Rapid local pressure variations are caused by vibrations in the brake system, such as brake squeal, see e.g. [7–9]. Brake squeal vibrations are associated with bending and wave motions of the pad and disc. These deformations result in local pressure variations over the contact surfaces, often on a time scale of milliseconds or less. The mechanisms for contact area variations are the same as in the global processes. 3.3. Slow processes In addition to the rapid mechanisms of contact area variation, a number of slow mechanisms are operating. These become important in determining the nature and size of the contact area if sufficient time (or sliding distance) is allowed. The slow processes are due to different kinds of wear and accumulation of debris, and to temperature variations. They typically appear on time scales of seconds or more, such as during long, low decelerating brakings or as the accumulated result of numerous brakings. Slow processes appear both on a microscale, that is on the scale of individual contact plateaus or smaller, and on a macroscale, that is on a scale involving numerous plateaus. The slow processes, which are discussed below, include: • • • • formation, growth and disintegration of contact plateaus; shape adaptation on a micro level; shape adaptation on a macro level; thermally induced deformation on a macro level; Fig. 6. Illustration of the mechanisms for rapid contact area variation: (a) the elastic loading and unloading of contact plateaus, i.e. the number of engaged contact plateaus increases due to the elastic deformation of the pad; (b) the area fraction of real contact within individual plateaus increases due to local deformation. M. Eriksson et al. / Wear 252 (2002) 26–36 • thermally induced surface property variations; • “contamination and cleaning” processes. 3.3.1. Formation, growth and disintegration of contact plateaus The growth and disintegration of contact plateaus involve agglomeration and compaction of pad wear debris around a wear resistant nucleus, as discussed in a previous section and illustrated in Fig. 4. 3.3.2. Shape adaptation on a micro level When the load on a contact plateau is increased, the small areas of real contact within the plateau will flatten plastically and by wear. These processes result in an increased area of real contact against the disc. When the load is decreased, the wear and deformation of the points in real contact will tend to reduce their contact with the disc. 3.3.3. Shape adaptation on a macro level The disc is continually worn, chiefly by the harder components in the pads. This wear will initially polish the disc surface, to make it better adapted to the pad. The individual contact plateaus on the pad will correspondingly experience milder contact conditions along the less rough sliding path. Due to the inhomogeneous structure of the materials, the continuous wear on both the disc and the pad will not be evenly distributed. However, the mutual adaptation to the shape of the countersurface will result in wavy surface. On the disc, the waves will form concentrical circles. In a steady state situation, the matching between the two parts is perfect, and each individual contact plateau will experience a smooth ride. However, small misalignments or movements between brakings or due to other changes, will result in mismatched surfaces and initially a reduced area of real contact. 3.3.4. Thermally induced deformation on a macro level During moments of high and increasing temperature, the pad surface will be hotter than the interior and the back plate. This will result in convex bending of the pad and hence an uneven pressure distribution, (see Fig. 7). The pressure reduction on the leading and trailing edges will result in a corresponding uneven distribution of wear; i.e. the pad will become thinner in the centre. When returning to a lower temperature the pad will straighten out. Now, the uneven wear during the bent 31 situation will result in reversal of the uneven pressure distribution. A similar behaviour may also be present for the mating part of the friction couple, the disc. Due to thermal instability, the disc may buckle along the sliding direction causing areas with locally high temperatures, known as hot spots. 3.3.5. Thermally induced surface property variations The properties of any surface depend on the prevailing temperature. When the disc and pad are heated during braking, this will affect both the chemical reactivity on their surfaces, the mechanical properties (thermal softening, etc.), the structure of the pad (decomposition of polymer constituents, etc.), the tendencies for smearing and sticking of wear debris on both surfaces, etc. Both the composition and the tribological properties of the surfaces are affected. 3.3.6. “Contamination and cleaning” processes Unloaded contact plateaus are exposed to different “contamination” processes, including oxidation, smearing out of wear debris and road dust, etc., which will change their composition. When the sliding contact is continued the plateaus will be subjected to a “cleaning” process involving the removal of the less wear resistant surface layers. This cleaning results in an increased degree of metallic contact. The corresponding processes occur on the disc surface. The slow contact surface variation processes are responsible for: • the slow increase of µ during long brakings; • the µ hysteresis reported for brakings under varying pressures; • the µ increase during running in of a new disc or a new pad; all effects which will be further discussed in the following section. 4. Friction coefficient variations The contact area variations are corresponded by different processes of friction coefficient variations. It can be stated that the friction generally increases with an increased area of real contact. However, the distribution of load over the different phases will also affect the friction level. Fig. 7. Mechanism for uneven pressure distribution and uneven wear due to thermally induced distortion of the pad and plate: (a) a hot surface will give the pad a convex bend; (b) when returning to lower temperatures the pad will straighten, but wear during the bent situation will result in an uneven shape and a corresponding uneven pressure distribution. 32 M. Eriksson et al. / Wear 252 (2002) 26–36 Fig. 8. Coefficient of friction during the running in of a new brake disc. The increase of the average friction coefficient level coincides with a gradual smoothening of the disc. Experimental data: program of varying brake-line pressures 3–25 bar, repeated brakings from 3 to 1 rps. Pad material TX4005. The initial running-in of a new disc is associated with a slow increase of the coefficient of friction. The as-manufactured disc surface has a spiral ridge pattern resulting from the turning operation. During the first brakings, this ridge is gradually worn off, resulting in a much smoother surface. Correspondingly, the coefficient of friction against the pads increases, (see Fig. 8). This effect is believed to be due to the fact that all continuous sliding of any individual contact plateau is made impossible by the topography (the rotating spiral) of the new disc. The discontinuous contact impedes the growth and mutual surface adaptation of the contact areas. At this point the mean friction coefficient during each braking was about 0.55. The slow increase of the average friction is believed to be due to the slow contact area growth processes described in the previous section. Specifically, the first formation of the contact plateaus involves a more rapid wear of the less wear resistant components in the pad. This will gradually transfer a larger proportion of the load (and contact area) to the more wear resistant components. These components, in this case brass fibres exhibit a higher friction coefficient against the disc. The repeated escalation of the friction coefficient during each individual braking and the corresponding fall between the brakings is treated in the next section. 4.2. The initial running-in of the pad 4.3. Repeated adaptation processes An unworn pad surface is rather flat, as typically formed by hot pressing. The described typical topography will be formed during the first number of brakings. The initial establishment of contact plateaus has been studied by measuring the friction development when running in a set of pads. The pads were prepared by grinding against 80 grit SiC paper to establish a macroscopically flat surface with a rather rough microtopography. The running in test was performed as a series of brakings using a constant braking pressure and a 2 min of idle time between consecutive brakings to allow cooling. An interesting tendency was found (see Fig. 9). During each individual braking, the friction coefficient was found to increase substantially. Initially, when braking with the freshly ground pads the mean friction value was 0.4, with a starting value just above 0.3. For each new consecutive braking both the mean value and the start and stop values increased. The increase was most rapid during the first four brakings, but continued at a lower pace for some 35 brakings. Further to the slow running-in of the pad and disc, each individual braking shows a form of a running in procedure resulting in a friction increase. This has been observed to be a very common behaviour and is exemplified by Fig. 9. Here, despite a constant braking pressure, the coefficient of friction escalates during each individual braking and decreases the corresponding level between the brakings. This behaviour can be due to the friction induced temperature rise, due to “cleaning” of the contact plateaus (removal of oxide films, contaminants, etc.), or finally due to geometrical adaptation necessary to compensate for minor changes of pad position and alignment. (The pads are not fixed and the disc is not ideally flat, thus a small misalignment or movement from a previously formed steady state contact position will require some wear or deformation to restore a close adaptation of shape. Even larger friction falls between brakings have been observed when the pad has been dismounted and remounted between consecutive brakings). 4.1. The initial running-in of the disc M. Eriksson et al. / Wear 252 (2002) 26–36 33 Fig. 9. Evolution of the friction coefficient of individual brakings from 3 to 1 rps at a constant 10 bar brake-line pressure. Each braking lasted for 20 or 30 s and was preceded by a 120 s idle time. Pad designation MD611B: (a) the first braking using the ground pads; (b) the first four individual brakings; (c) all 40 brakings. 4.4. Friction hysteresis The slow processes for contact area variations will cause a slow response to pressure changes. During pressure increase, the area of real contact will be smaller than the steady state size, and hence the friction coefficient will be lower. Correspondingly, during pressure decrease the area of real contact will be unproportionally large and the friction coefficient thus higher than in steady state at that pressure. This influence from the immediate brake pressure history on the momentary friction coefficient constitutes hysteresis behaviour of the friction coefficient. A simple experiment with continuously varying brake pressure demonstrates this phenomenon (see Fig. 10). For one of the sets of pads, the friction coefficient is clearly lower when the brake pressure is increasing than when it is decreasing (see Fig. 10a and b). This effect is believed to be due to the relatively slow adaptation of the friction surfaces to the prevailing conditions. When the load is continuously increasing, the real area of contact must increase correspondingly. New contact plateaus are formed and the size of the existing plateaus increases by the rapid contact area variation processes. However, also the slow processes operate, including cleaning of contact plateaus, shape adaptation on the micro level, contact plateau growth, etc. During the time it takes for these processes to form the contact area typical for the prevailing higher pressure, the friction will be lower. Correspondingly, when the normal load is decreasing the elastic unloading of contact plateaus is a very rapid process. However, the area of real contact between the cleaned and microscopically adapted spots will reduce more slowly. When the load on a spot in real contact is reduced, wear and deformation will result in a relatively slow decrease of the contact area, and a corresponding slowly decreasing friction. However, this hysteresis behaviour is very dependent on the friction characteristics of the pad material. The results from a pad of a different formulation, exhibiting very little friction hysteresis, are shown in Fig. 10c and d. 34 M. Eriksson et al. / Wear 252 (2002) 26–36 Fig. 10. Hysteresis in the coefficient of friction with continuously changing normal load (brake-line pressure). Braking at 2 rps constant speed: (a) and (b) hysteresis effects involving higher coefficient of friction for decreasing than for increasing normal load. Standard pad to the Volvo 850 TX4005: (c) and (d) less obvious hysteresis displayed by a non-commercial pad formulation MD631D; (e) and (f) brake-line pressure variations and the resulting disc temperatures. (e) Corresponds to the friction curves in (a) and (c), while (f) corresponds to (b) and (d). A lining material that quickly adapts to new contact conditions will show a smaller hysteresis than does a slowly adapting pad and vice versa. In order to separate the effect of varying the pressure from the influence of the resulting temperature changes, the pressure variations were performed in two ways. The test was first started from low load, increasing to maximum and then going back as shown in Fig. 10e. The second run used the reversed pressure curve, starting from the maximum pressure, decreased and finished by increasing back to maximum (see Fig. 10f). In this way it has been made clear that for the lining material showing hysteresis, the coefficient of friction is higher when reducing than when increasing the pressure, irrespective of whether the temperature is higher or lower. The design of the test equipment, a disc brake using pressure control to adjust the brake pads, minimises the effects of self locking as well as any thermomechanical or viscoelastic effects. Thus, the measured brake pressure always correlated well to the actual normal load, making the studied hysteresis a true frictional effect. 4.5. Effects of reducing the pad area The exact size of the pad area has been found to have a very limited influence on the friction . This is as expected from fundamental friction understanding: the size of the nominal contact area has very little influence on the M. Eriksson et al. / Wear 252 (2002) 26–36 35 friction level as long as it is very much larger than the real contact area. The only effect of decreasing the nominal contact area will be to bring the areas of real contact correspondingly closer, and in the particular case of brake linings probably collected within fewer contact plateaus. 4.6. Effects of disc surface topography In an investigation on the influence of the disc topography on the generation of brake squeal , it was found that shot blasting the disc surface reduced friction coefficient. Shot blasting produces a rough surface, basically characterised by small pits scattered over a relatively flat surface. Here, the pits produced by the angular SiC particles (φ, 1–2 mm) where typically 100 m wide and 20 m deep. The presence of these pits reduced the area of the sliding surface on the disc. For each individual contact plateau on the pad, the passage of a blast pit results in partial or total loss of contact. Prior to the shot blasting, the friction coefficient of the tested system was about 0.6. After the blasting, the initial friction level was about 0.3 (see Fig. 11). The initial topography of the disc surface was coarse, with sharp ridges surrounding the pits. After some 300 brakings the friction coefficient reached 0.4, all traces of the ridges were worn away, and the area fraction of the flat surface was about 70%. The fraction of flat disc contact area continued to grow as the disc was gradually worn. The coefficient of friction, however, increased very slowly. By the end of the test (after 1260 brakings), while the disc was very smooth (no roughness could be felt when touching with the finger tip) and the nominal contact area fraction was almost 90%, the friction was still 25% lower than prior to the shot blasting. Fig. 11. Development of the coefficient of friction and nominal fraction of contact area of the disc during a sequence of brakings against a shot blasted disc. In another investigation only a sector of the disc was shot blasted . Here it was found that the friction fell during each passage by the sector, and re-established quickly after the passage (see Fig. 12). The low friction against the dimpled surface and the quick re-establishment of the friction when returning to the unaltered surface are supposed to be due to changes of the real contact area. The frequent encounters between the contact plateaus and the pits in the disc surface will damage the contact plateaus, thus preventing them from growing to their normal steady state size. Further, it will obstruct the shape adaptation on the micro level. Since the damage and loss of adaptation here is on a small scale, the re-establishment to steady state conditions will be quick compared to the running-in procedures and the repeated adaptation process. Fig. 12. Coefficient of friction and overlap between the pad and the shot blasted sector of the disc. 36 M. Eriksson et al. / Wear 252 (2002) 26–36 Fig. 13. Overview of the discussed mechanisms for contact surface variation. 5. Conclusive summary Acknowledgements The contact situation between a cast iron brake disc and an organic brake pad is specific for this material combination and unlike most other tribological contacts. It is a result of the wide diversity in mechanical properties for the pad components. The pad area is divided into numerous contact plateaus (occupying some 20% of the area) surrounded by lowlands. The lowlands are constantly out of sliding contact. The area of real contact is concentrated to small spots confined within the plateaus. The plateaus have a relatively long life while the areas of real contact are constantly shifting due to wear and deformation and surface roughness on the disc surface. Due to the described contact surface variations, the friction coefficient will vary. Apart from high frequency scatter, the friction level has been found to shift between 0.3 and 0.6, due to the different mechanisms described in the paper. It has been shown that the contact surfaces between the pad and the disc may vary with respect to size, properties and composition. These variations are due to a multitude of processes operating on different scales in time and size. An overview of the processes is given in Fig. 13. It can be concluded that the nature of tribological contact in automotive brakes is manifold and intricate. Numerical simulations of brake behaviour and brake properties are very useful in the development of new brake systems. A relevant model of the friction is of vital importance in these simulations. However, such a model will have to be extremely complex to cover only the most important phenomena and still probably not suitable for numerical work. It is expected that in the foreseeable future, careful experimental work will still be needed to develop better brakes and their friction materials. Deep insights into the nature of the tribological contact will render this process more rational and efficient and in the long run build a useful base for relevant friction modelling. The authors gratefully acknowledge the financial support from the Swedish Board for Technical Development (NUTEK), Volvo Technological Development for providing test materials and Claes Kuylenstierna for valuable discussions. References  D.A. Crolla, A.M. Lang, Brake noise and vibrations—the state of the art, in: Proceedings of the Leeds–Lyon Symposium on Tribology 17, 1991.  F. Bergman, M. Eriksson, S. Jacobson, A software based measurement system for test and analysis of automotive brake squeal, TriboTest 5 (3) (1999).  N. Axén, S. Jacobson, A model for the abrasive wear resistance of multiphase materials, Wear 174 (1994) 187–199.  S. Jacobson, Applications of a new model for the abrasive wear resistance of multiphase materials, composites and coated materials, in: Proceedings of the Austrib’94, Perth, Australia, 1994.  N. Axén, I.M. Hutchings, S. Jacobson, A model for the friction of multiphase materials in abrasion, Tribol. Int. 29 (6) (1996) 467–475.  M. Eriksson, F. Bergman, S. Jacobson, Surface characterisation of brake pads after running under silent and squealing conditions, in: Proceedings of the Nordtrib’98, 1998, Ebeltoft, Denmark.  J.O. Hultén, Friction Phenomena Related to Drum Brake Squeal Instabilities, ASME Paper DETC 97/VIB-4161, September 1997.  J.O. Hultén, Drum brake squeal—a self-exiting mechanism with constant friction, in: Proceedings of the SAE Truck and Bus Meeting, SAE Paper 932965, Detroit, MI, USA, 1993.  F. Bergman, M. Eriksson, S. Jacobson, The effect of reduced contact area on the occurrence of brake squeals for an automotive brake pad, J. Automotive Eng., 1999, in press.  F. Bergman, M. Eriksson, S. Jacobson, Influence of disc topography on generation of brake squeal, Wear, 1999, in press.  M. Eriksson, F. Bergman, S. Jacobson, A study of initialization and inhibition of disc-brake squeal. J. Automobile Eng. 1999, in press.