Synthesis Of Disc Brake Squeal Quieting Experience

Transportation Research Record 983at 100 ft reached a maximum level and remained constant for approximately 6 sec.3. At a train speed of 90 mph the noise level at100 ft reached a maximum level and remained constantfor approximately 3 sec.4. The noise level behind the noise wall inMontebello was 74 dB(A) when the passenger trainspeed was 65 mph.5. Because of continuously welded rail construction of the trackage in Montebello, the noise levelis expected to increase less than 4. 5 dB (A) if thespeed is increased to 90 mph.6. The noise wall in Montebello along SycamoreStreet provides approximately 13 dB(A) attenuation.437. Other noise from sources ( such as switchingof freight cars, freight trains, and engines; passing cars; motorcycles; delivery trucks; and airplanes) ranged from 71 to 82 dB(A). The highesttrain whistle noise recorded was 99 dB(A). Theselevels were measured behind the wall at the intersection of Spruce and Sycamore streets in Montebello.Publication of this paper sponsored by Committee onTransportation-Related Noise and Vibration.Synthesis of Disc Brake Squeal Quieting ExperienceMICHAEL A. STAIANOABSTRACTDisc brake squeal produced by WashingtonMetropolitan Area Transit Authority transitcars precipitated a number of on-car andlaboratory tests. The results of these testsare summarized along with the experience ofother investigators in both the transit andautomotive fields. Experience indicates thatsome brake systems are prone to squeal whileothers are not, and that the benefit ofsystem modifications varies dramatically fordifferent disc brake systems. Modificationsthat hold the most promise for reducing thepropensity of a brake system to squeal aresufficiently damping the brake pad backplate, disc rotor, or caliper; altering discrotor stiffness, caliper mass, or caliperstiffness; reducing (if possible) the brakepad friction coefficient;and, possibly,eliminating brake pad grooves.gained in attempting to reduce the brake squeal onWMATA cars and to place this experience in the context of other work.WASHINGTON METRORAIL EXPERIENCEThe disc brake assembly used by the WMATA Metro carsis similar in configuration to automotive caliperdisc brakes with ventilated rotors (Figure 1). Thedisc is 20 in. in diameter and 3.6 in. thick andweighs 137 lb (1). The kidney-shaped brake pads areactuated by t side-by-side pistons. The brakesystem was designed by the Abex Corporation with twoassemblies fitted per axle for a total of eightassemblies per car. (Author's note: Equipment sup--The existence of brake squeal has been recognized asa problem for some time primarily because of itsoccurrence in highway vehicle drum and disc brakesystems. Rail transit cars have also exhibited brakesqueal from their tread and disc brakes. In theUnited States three transit systems use disc brakes:the San Francisco Bay Area Rapid Transit District(BART), the Chicago Transit Authority (CTA), and A). Of these, only WMATA has experienced problems due to brake squeal. The squeal propensity ofthe WMATA brake system has limited the optionsavailable for replacement friction pad materials,and at one point squeal sound levels became sosevere that a considerable public outcry ensued. Thepurpose of this paper is to document the experienceCaliper ----.;:LJ(Piston Sidel.-··-.,,.)·. . /Bracket andLink AssemblyCaliper(Piston SidelFIGURE 1 WMATA/Abex disc brake system.Truck Wheel

Transportation Research Record 98344pliers are identified in this paper for clarity ofpresentation. Their reference constitutes neitherendorsement nor criticism.)From the time WMATA initiated rail service, itsbr . e 5Y!:t.e!!'. e h! iteC ! pe eit ,7 t ::q l { }.The BART brake system with similar design parameters, on the other hand, has never exhibited squeal(.!) The squeal experienced early in the Metrorailhistory was occasional in nature and relatively lowin amplitude and frequency (Figure 2a). However,when the original asbestos-bearing Abex 45109 fric-pads, squeal became common, and its amplitude andfrequency increased considerably (Figure 2b) (]).To correct the deteriorated squeal behavior, aseries of on-car modifications were evaluated for animmeaiate solution (ii. Laboratory investigationsusing a constant-speed dynamometer were also undertaken for long-term improvements Ci,il. The on-carmodifications included a variety of replacement pads,md system modifioationo fitted to teat trains runduring nonrevenue hours. The on-car tests found onepad--the Abex l389b--to be relatively squeal-free.This pad was selected as a suitable replacement padwith acceptable squeal behavior.SQUEAL fll .ECHANISMThe mechanism of brake squeal generation confoundedinvestigators until two key observations were made(1):01020304050l. Squeal is a function of the direction of discrotation (i.e., squeal might occur for only clockwise rotation and not for counterclockwise rotation), and2. Deceleration measured during a squealing stopis somewhat higher than that measured for an essentially identical stop without squeal.(a) Judiciary Square station, January 27, 197601020304050TIME (s)(b) McPherson Square station, February 20, 1981FIGURE 2 Typical brake squeal time histories.These observations suggested that squeal is related to an asymmetry of the contact area of the padwith the disc, such that the contact area is in a"sprag position" with respect to the backplatepiston contact point. This sprag configuration, or"digging in" of the pad to the disc, results inelastic deformation of the pad, which then reducesthe binding forces and releases the pad to its original position from which the cycle can be repeated.This theory was tested by grinding a pad to an exaggerated asymmetry and varying the backplate-pistoncontact point. Squeal was observed consistent withthe theory for the con f iguration shown in Figure 3Cl) fore tan-l µ, whereµ is the pad friction coefficient. Further analysis indicated that squeal wasposs ibl e for O e tan-1 µ Rotadon.Contact PointFIGURE 3 Disc, pad, and piston in squeal configuration (7).

45Staianomodel consists of three elements (the disc, pad, andcaliper), with each possessing rotational and translational stiffness and mass. Element rotations areabout an axis parallel to the disc radius and displacements are lateral (i.e., normal to the plane ofthe disc) Disc mass and stiffnesses are effectiveor equivalent values estimated for an equivalentbeam from expected disc modal behavior during squeal.When the equations of motion and boundary conditions are applied to the disc, pad, and caliperelements, a set of equations is obtained yieldingsolutions for the displacements of disc, pad, andcaliper and the rotations of the disc and pad. Thesesolutions have the form y Y ezt, where z can beimaginaryi thus,Fundamental to the sprag-slip theory is the presence of the asymmetrical pad-disc contact area. Theasymmetry is believed to arise from nonuniform padwear, working of the backplate by the piston, orthermal distortion of the pad (1, ). A squeal-pronecondition and slight backplate dishing have beenobserved after simply heating a pad to about 250 Cand allowing it to cool. Thus heavy braking andsubsequent cooling is likely to cause such distortions, as has been observed of squeal in in-serviceconditions. At low brake pressures the contact areais quite small due to the distortioni at high brakepressures the pad is flattened-out, thereby reducingthe asymmetry. The result is less squeal at highbrake pressures. The requirement that pad asymmetryexist probably accounts for the appearance of squealin the Knorr BB1 pad in the WMATA brake system onlyafter a period of revenue service.Investigators have noted that the frictionalforce on the face of the pad is resisted by a restraining force at the pad backplate resulting in amoment tending to rotate the leading edge of the padtoward the disc and causing more rapid wear at theleading edge (l). This moment also tends to shiftthe contact patch toward the leading edge--to a spragg ing" position. This is shown in Figure 4.From translational equilibrium, all forces acting onthe pad can be estimated in terms of the brake pressure force (N) and the friction coefficient. Fromrotational equilibrium, the offset (a) of the paddisc effective contact point to the backplate-pistoneffective contact point must be a µt, where t isthe pad thickness. It is interesting to note thatthe maximum leading contact point offset (b) forsqueal is equal to a. (Note that, even for theWMATA/Abex brake system, this offset is quite small.Specifically, for a new pad, b 0.25 in. for apad whose half-length is approximately 6 in.)ezt ext eiwt ext [cos(wt) i sin(wt)]Therefore, the solutions are simple harmonic if x O, sinusoidal and damped if xis negative, and sinusoidal and diverging (i.e., unstable) if x is positive (15). Squeal is the self-excited oscillationthat occurs when the brake system parameter valuescombine to yield an unstable solution to the displacement equations (i.e., positive x). In the literature the magnitude of positive x has been definedas the squeal propensity and w/2 has been calledthe "squeal frequency. EFFECT OF BRAKE SYSTEM DESIGN ON SQUEALThe fundamental squeal mechanism is understood.Simplified experimental and analytical models provide results that agree mutually and agree reasonably with specific real brake systems. However,behavior for one brake system may not be representative of that of another. Furthermore, the best current analytical model in the literature is still aconsiderable simplification of a complex system.Industry practice (both rail transit and automotive)for solving squeal problems remains basically trialand-error modification of actual components. However, the analytical results are useful in guidingthe types of modifications that may ultimately provesuccessful.ANALYTICAL DESCRIPTION OF SQUEALThe sprag-slip theory was supported and refined by aseries of investigations ( -14). A mathematicaldescription of a complex system such as a disc brakeassembly requires a considerable number of simplifying assumptions. Expression of sprag-slip for a real(automotive) disc brake assembly, for example, consisted of a 6 degree of freedom (df) lumped-parameter model in which system damping was ignored (B).The interaction of the inner and outer pads was alsoignored, such that the model was simplified to asingle pad pressed against the disc. [Other work hasdemonstrated that the interaction of the pads doesaffect the parameter ranges that will result insqueal. However, the general trends indicated by thesingle-pad model are representative (14).] Thus thisBrake PadClearly, the brake pad plays a significant role inthe squeal mechanism. When the original Abex 45109pad was replaced by the Knorr B81 pad on the Metrorail cars, squeal worsened considerablyi when theKnorr 881 was replaced by the Abex 1389b, squeal wasvirtually eliminated (see Table 1). Interestingly,the presence of even one squeal-prone pad in aBRAKE PRESSURE FORCE, N/BRAKE PAD/µNI','//1 /;'/'FRI CT!ON FORCE, µN(1)/"aNFIGURE 4 Forces acting on brake pad.I'/""I'''DISC ROTATIO