Brake Pad Quality Assurance with Ultrasound
In this study several non-destructive test methods have been applied to as-manufactured automotive brake pads. The primary emphasis of our study is the formulation and development of ultrasonic methods where 4 independent velocity modes are measured on each pad. For two of the measurements, the ultrasound is propagated in-the-plane of the pad, while in two other measurements the ultrasound is propagated through-the-thickness (out-of-plane). Over 300 pads from 5 different manufacturers have been tested. In many cases, the ultrasonic data is compared with other testing methods including conventional compressibility tests, modal analysis, and hardness testing. In some cases, measurements have been made of several different batches of materials to test long term consistency of the material properties in the production environment. In other studies the production process has been deliberately altered to help establish specific cause and effect relationships.
Brake noise, a major source of warranty costs, is a complex problem involving a myriad of design and processing variables that include friction material properties, rotor design, caliper design, and vehicle suspension. The mechanical properties of friction materials are thought to play an important role in braking system noise performance1-4. Although brake noise is minimized through proper design and appropriate friction material formulation, the realization of noise-free brakes requires that the manufacturer maintain these properties in production. Process variations can significantly alter friction material properties. This can adversely influence noise performance. Better measurement tools applied at the point of manufacture are desirable.
This NSF Small Business Innovation Research project was designed to enhance the automotive friction material manufacturing industry’s productivity and efficiency by providing a superior measurement method for quality, consistency and the quantification of noise influencing material properties. Friction material manufacturing is subject to intra-material as well as inter-batch inconsistency that existing methods are unable to adequately quantify at the point of manufacture. These inconsistencies adversely affect customer satisfaction, contribute to lost business and consume engineering and testing resources. In this program, studies were conducted to relate ultrasonic data to friction material processing variables and to forge a relationship between ultrasonic test data and brake noise performance. Ultrasonic measurement can be implemented as both part of a control scheme to improve the manufacture of friction materials and/or as a quality assurance method to ensure that noise-prone components do not enter the marketplace.
Summary of Accomplishments
This program demonstrated the ability to use ultrasonic methods to obtain characterization data on as-manufactured, automotive friction materials. Over 300 brake pads of 7 different configurations from 5 different manufacturers were non-destructively measured. Measurements on production pads demonstrated significant variation in both the average value and spatial uniformity of friction materials from various manufacturers. Specific process studies related measured ultrasonic characteristics to variations in manufacturing. Laboratory experiments identified ultrasonic coupling methods, signal processing schemes and analysis methods suitable for rapid, automated testing. A mechanically scanned prototype system was assembled and used to test brake pads. Comparison of automated test results with those obtained using manual test methods established the feasibility of the automated testing scheme. The survey studies established the need; the process and performance studies demonstrated relevance/sensitivity; the prototype proved the feasibility.
The use of ultrasound to determine mechanical properties of materials is based on the fundamental relationship between the ultrasonic velocity and the material elastic constants. These methods have been described in a number of books and review articles5-7. For automotive friction materials, the primary application for ultrasonic measurements has traditionally been the determination of the material elastic constants. This data is used for input into computer models of braking system performance which speeds the development of new materials. The current test is destructive; requiring the removal of the friction material from its steel backing plate. By measuring sound speeds of longitudinal and shear waves along various directions, the elastic constants are determined. This method is described in detail in the testing specification SAE J27258.
The rationale underlying this work is to exploit the methods used for elastic property measurements to develop ultrasonic-based non-destructive methods which can be used for quality assurance and improvement in friction material manufacturing methods. Successful tests must be non-destructive, automated, and readily adapted to the numerous configurations of as-manufactured components. Although it is currently not possible to measure the complete set of elastic constants on as-manufactured brakes, it is possible to measure the relevant ultrasonic velocities both in the plane of the pad and through the thickness of the pad. Because the elastic modulus is proportional to the square of the velocity, it is plausible that magnitude and uniformity of the measured velocity is useful as a potential process quality assurance measurement.
The symmetry of both drum brake segments and disc brakes is transversely isotropic, owing to compression molding manufacturing methods. The elastic properties are, to a good approximation, isotropic in the direction perpendicular to the pressing direction (i.e. through-the-thickness), but are 4 to 5 times more compressible in the pressing direction due primarily to flow and orientation of fibers7. Figure 1a shows the coordinate definition for a typical disc brake pad where the “3” direction (through-the-thickness) is along the pressing direction. The “3” direction is also the direction which force is applied in a braking application. The “1” and “2” directions are in the plane of the pad and perpendicular to the pressing direction. Figure 1b shows the elastic constant matrix for this coordinate definition and transversely isotropic symmetry showing that there are 5 independent elastic constants.
The ultrasonic velocities that in part contribute to all of the diagonal elements of the elastic constant matrix can be determined non-destructively. Specifically, the elastic constant is directly related to the density times the square of the appropriate ultrasonic velocity. Measurement of the compression velocity, V33 through-the-thickness and V22 in the plane yield the C33 and C22 while measurement of the shear velocities V32, and V21 yield the diagonal constants C44 and C66.
With the velocity in (Km/sec) and the density, r in (g/cc) the above product yields modulus in GPa We are using conventional tensor notation to designate the velocity where the first sub-index is the propagation direction and the second sub-index is the wave polarization. The through-the-thickness velocities should be inversely related to the material compressibility while the in-plane velocities will be more closely related to the pad flexural and torsion vibration modes.
Figure 1 a) Coordinate definition for disk brake pad; b) Elastic constant matrix for transversely isotropic symmetry with the “3” direction as the unique axis
As-Manufactured Brake Pad Measurements
The strategy for this development entailed immediately implementing non-destructive laboratory methods similar to those used previously for destructive laboratory-based measurements of friction material elastic constants (SAE J2725). Although these methods are labor intensive, they have the advantage of being readily adapted to the large variety of brake pad geometries. Because the methods are non-destructive the test results and test samples can be used to evaluate and validate the automated methods implemented in the later stage of this program. Additionally, these results are used to demonstrate the potential of the ultrasonic method to friction material vendors, Tier 1 brake systems suppliers, and original equipment manufacturers.
The measurement process begins by generating a scanning template illustrated in Figure 2. The numbered, circular regions Figure 2a show the location of regions where the through-the-thickness velocities, V33 & V32 are measured. For a standard measurement, a sensor with a “footprint” of 12 millimeters in diameter is used. Multiple measurements (eight regions in this case) are made on each pad. Thus it is possible to measure the spatial uniformity within as-manufactured pads. As shown in the micrograph, the steel backing may contain structures (attachment clips and spigot holes) which preclude measurements on specific pad areas. The location of these features as well as the shape of the pad varies from one brake design to the next, requiring that the measurement method to be flexible. For laboratory measurements, a viscous, water-soluble organic coupling compound, (IMS-SWC), was used to promote ultrasonic transmission. Figure 2b shows the propagation path for the in-plane velocity measurement, V22 & V21. Sensors are placed opposite one another and ultrasound is transmitted across the entire width of the pad. Only one trajectory is shown in Figure 2b. However, in general 3 to 4 measurement zones are possible along the entire length of the pad.
Figure 2 Example measurement template for a brake pad; a)Numbered circles indicate the measurement locations appropriate for the V33 and V32 velocities. b) Propagation path for the in-plane velocity measurement, V22, & V21
In this program, most of the effort has been directed at measuring the through-the-thickness velocities V33 and V32. These velocity modes are related to the friction material compressibility which is known to influence noise as well as other brake performance parameters9,10. For the through-the-thickness modes, the average pad velocity and the spatial variation of the velocity is measured. For completeness, the in-plane longitudinal and shear modes V22 and V21 are also measured. It may be of interest to measure the attenuation or absorption of ultrasound and the load-dependence of the velocity (friction materials exhibit non-linear elastic behavior). These parameters might be useful measures of brake pad quality. These parameters are best measured using a fully automated system and are beyond the scope of this Phase I effort.
Figure 3a shows the configuration used for the semi-automated laboratory measurement of the through-the-thickness modes, V33 & V32. Brake pads are “sandwiched” between a pair of sensors and manually positioned to one of the eight regions shown in Figure 2a. The time-of-flight, ToF, of the ultrasonic wave propagating through the entire component is recorded. The ultrasonic velocity and thickness of the steel backing is known and well controlled. Thus, it is relatively easy to remove the influence of the steel from the measured ToF to yield a measure of the friction material velocity. The in-plane modes, V22 and V21 are measured using the same test fixture (Fig 3b) with the pad rotated out of the plane. These test methods have been described in detail in a recent publication11.
The first step in our analysis is to determine the measurement error (repeatability) when applied to as-manufactured brake pads. This was accomplished by making repeat measurements on 10 production pads.. Figure 4 shows the results obtained for 2 modes, the through-the-thickness mode V33 (Figure 4a) and the in-plane mode, V22 (Figure 4b). For the V33 mode, the mean value for the 6 repeat measurements along with the standard deviation (error bars) are plotted for each of the 10 batch #6 samples labeled “6-1” to “6-10”. Similar results for the 6 repeat measurements for the V22 mode are shown in 4b. These results quantify the measurement error for our test method which varies from +/- 0.2% to 0.5% for the through-the-thickness modes and +/- 0.3% to 0.7% for the in-plane modes. Essentially identical results are obtained on the shear modes V32 & V21.
One other variable which needs to be considered in the measurement process is the coupling pressure. Due to their structure, friction materials display non-linear elastic properties. The non-linear behavior arises at very low loads, well below the yield point of the materials. For ultrasonic velocity measurements, this translates into a strong dependence of the velocity on load. For the measurements described above, the coupling force was controlled at 4.0 MPa. For a fully automated system it is possible to measure the velocity at multiple loads. For our semi-automated work we have chosen to measure at a single load of 4 MPa.
Figure 3a) Configuration for through-the-thickness (V33,V32) measurements and b) in-plane (V22,V21) on as-manufactured brake pads
Figure 4 Measurement repeatability for V33 and V22 on 10 production pads from batch #6