Efficiently Measuring the Behavior of Brake Materials - Correlation Between Benchtop and Dynamometer Tests

Using a full-scale dynamometer for the testing of friction materials for brake applications is a cost- and time-intensive method as it requires the rotor and brake pads to be in their final form (Figure 1). Bruker has developed a more rapid and cost-effective technique to screen materials used in automotive brake applications. This new benchtop method utilizes small samples made of friction materials, and subjects them to testing under real brake operating conditions.

The UMT TriboLab™ can be used to test such materials in real time and easily program industry-standard dynamometer protocols and standards, all while monitoring key parameters such as friction, sliding speed, vibration and temperature. This article discusses the technique and highlights its very good correlation to tests performed on the full-scale dynamometer following the SAE J2522 standard (known as AK Master).

Figure 1. Brake assembly of rotor and pad.

Evaluating Performance of Friction Materials for Automotive Brake Applications

Many challenges are involved in the development of new generation materials to be used for automotive brake applications, including formulation of materials that address cost reduction; expected (usually more demanding) performance; and compliance with new safety and environmental regulations. Brake materials are tested before on-vehicle stopping tests by fully developed methods such as the use of a full-scale dynamometer (see Figure 2).

Figure 2. Full-scale brake test protocols/standards are typically conducted on dynamometers. Photo Courtesy of Greening Test Laboratories.

In dynamometer tests, real pad or rotor can be tested under protocols simulating the conditions required to stop a vehicle. Different standards have been used by this specialized industry, such as the SAE J2522 developed by the AK Working Group, which represents European manufacturers of passenger car brakes. This test was designed to assess the effectiveness of the brake pad and the rotor system under different conditions of speeds, temperatures, pressures and deceleration.

To conduct any relevant test using the dynamometer, the rotor and pad need to be manufactured in the exact size and shape of a final product. Hence, it is a cost-intensive method, without even considering the time required to prepare materials and the queue time in accessing available dynamometers. In addition, this testing method requires other elements of the brake system in order to introduce further variables.

For instance, the effectiveness of the caliper can vary depending on design and as a result, the effect of the rotor and brake becomes difficult to separate.¹ Hence, it is necessary to have a fast screening method in the early stage of brake materials development, before the more laborious and expensive testing at the component level using the dynamometer.

Testing Brake Materials at a Reduced Scale

Bruker’s Brake Material Screening Tester for the UMT TriboLab (see Figure 3) was specifically developed to be a cost-effective and fast screening method and rank materials before performing the component level evaluation. With the techniques used, the tribological performance of small, friction material samples can be characterized in a precise and timely manner, while monitoring key parameters such as sliding speed, friction, wear, vibration and temperature.

Figure 3. UMT TriboLab Brake Material Screening Tester.

To scale down and properly simulate the brake system, critical physical parameters used by dynamometers must be consistently matched for protocols like the SAE J2522:

• Contact pressure between the pad and rotor
• Deceleration
• Sliding speed
• Initial temperature

Since the inertia of benchtop systems is typically smaller than a dynamometer or vehicle, deceleration is simulated by controlling the velocity of the motor as a function of time. During the different steps of the simulation, key parameters monitored include sliding speed, temperature of the pad, temperature of the rotor, torque and coefficient of friction.

Another key advantage of carrying out the tests at a smaller scale is the possibility of easily controlling environmental parameters, like temperature and relative humidity, but also the capability of collecting the debris released from the sliding contact interfaces during the test. Such brake wear particles can then be subjected to post-test physical and chemical characterization.

The traditional tests that fulfill industry standards typically are focused on deceleration tests, wherein the vehicle is simulated during stopping or speed-reduction conditions (e.g., in many of its test steps, the SAE J2522 uses snubs from 80 km/hour to 30 km/hour applying 3,000 kPa in the fluid line that applies the force to the caliper). The UMT TriboLab allows the deceleration force to be simulated in the same way, as illustrated in Figure 4.

Figure 4. Simulation of snub performed with the UMT TriboLab varying the speed from 2089 to 787 rpm (80-30 km/hour vehicle speed) in 5.5 seconds, and under an applied force of 300 N (0.75 MPa contact pressure). Force and speed are controlled, while torque and temperatures are monitored. Typically, torque increases while speed is reduced at constant contact pressure.

Correlation between Benchtop Test and Full-Scale Dynamometer

The minimum contact size of the coupons is critical in benchtop friction material testing. While it is possible to select the sliding speed and contact pressure from the real application or test protocol, there is a minimum sample size that can represent the composition and non-homogeneous morphology of the brake pad.

Here, the UMT TriboLab was used to simulate two different brake pad materials (A and B) that were previously evaluated on a full-scale dynamometer in accordance with the SAE J2522 protocol (courtesy of Greening Testing Laboratories). Figure 5 illustrates a typical report of the results obtained from SAE J2522, presenting the different steps and cycles for material "A", and showing the key variables such as friction, torque, pressure and temperature.

Figure 5. SAE J2522 graphic report of the material A test performed on the dynamometer. Courtesy of Greening Test Laboratories.

The samples were collected from the same pads, as small cylinders with a thickness of 6.35 mm [0.25"] and diameter of 12.7 mm [0.5"]. A real rotor was machined to the desired dimensions to obtain a cast iron disc so as to represent the rotor. For the setup, three samples of each material were positioned on a radius of 38 mm (Figure 6).

Figure 6. Samples that simulate the brake pad and rotor for testing with the TriboLab Brake Material Screening Tester.

The pad materials coupons were pressed against the cast iron disc, applying loads that simulate contact pressures employed in the SAE J2522 protocol. The area of the pad, the diameter of the piston and the pressure of the fluid need to be known to calculate the contact pressures applied at the full-scale test. The tire-rolling radius and the effective radius of the rotor were considered to calculate the linear speeds to be used on the benchtop scale test. Again, since a benchtop scale system has a very low inertia in comparison with the dynamometer, the deceleration was precisely controlled at the motor, using similar stopping-time values to those collected from the full-scale test.

The J2522 protocol includes 15 different main steps and many other sub-steps that aimed at evaluating the friction performance of the brake under various conditions. To perform this comparison, 8 of these steps (292 cycles) were selected:

• 6.1 Green µ characteristic (30 cycles): snub 80-30 km/hour, 380 N
• 6.2 Burnish (192 cycles): snub 80 to 30 km/hour, varying the applied load
• 6.3 Characteristic value 1: (6 cycles): snub 80-30 km/hour, 380 N
• 6.4 Speed/pressure sensitivity
  • 6.4.1 Speed/pressure sensitivity 40 km/hour (8 cycles): snub 40-5 km/hour, varying the load
  • 6.4.2 Speed/pressure sensitivity 80 km/hour (8 cycles): snub 80-40 km/hour, varying the load
  • 6.4.3 Speed/pressure sensitivity 120 km/hour (8 cycles): snub 120-80 km/hour, varying the load
• 6.5 Characteristic value 2: (6 cycles): snub 80-30 km/hour, 380 N
• 6.6 Cold application: 1 stop 40-5 km/hour, 380 N
• 6.9 Fade 1: 15 stops 100-5 km/hour, 500 N
• 6.10 Recovery 1 (18 cycles): snub 80-30 km/hour, 380 N

For the comparison, some of the key parameters were calculated after the tests:

• µ0P61 Average value of the friction values of the 1^st to 6^th application in section 6.3 (characteristic value 1)
• µv120 Average value of the friction values at 2000, 3000 and 4000 kPa in section 6.4.3 (speed/pressure sensitivity 120 km/hour)
• µ0P62 Average value of the friction values of the 1^st to 6^th application in section 6.5 (characteristic value 2)
• µT40 Friction value of the 1^st application in section 6.6 (cold application)
• µT40 Friction value of the 1^st application in section 6.6 (cold application)

The data obtained during the tests reveal a good correlation between the benchtop and full-scale tests, not only from the calculated average CoF, but also in the behavior of the torque with similar trend and shape. The results of the characteristic value 1 step for material "B" are shown in Figure 7. Compared to the dynamometer tests, more evidence of vibration (observing the noise on the CoF) is observed in the TriboLab tests. This could be because of differences in filtering the data, and due to the position of the sensor in the TriboLab tester, where the sensor is positioned very close to the tribological surfaces. Temperature values are different in magnitude, as temperature is not an intrinsic property, and relies on the mass of the system components.

Figure 7. Comparison of the SAE J2522 – 6.3 characteristic value 1. Results for material "B."

When observing another step of the SAEJ2522 protocol, wherein there is an increment in load in every cycle, a very good correlation is observed in terms of the coefficient of friction following the trend. Figure 8 presents the step 6.4.2 that aims to measure the sensitivity to brake-application pressure when the system is tested at snubs of 40-50 kmh.

Figure 8. Comparison of the SAE J2522 – 6.4.2 speed/pressure sensitivity 40 km/hour. Results for material "A."

A good correlation between the dynamometer and TriboLab results is also observed when the materials were tested at 40 °C ("cold application," step 6.6), where the CoF values are very close (Figure 9)

Figure 9. Comparison of the SAE J2522 – 6.6 cold application. Results for material "A."

The comparison of all the calculated average coefficient of friction for the different tests, for both material A and B, confirms that the TriboLab Brake Material Screening Tester can be employed as a reliable instrument to evaluate the friction behavior of brake materials, and as a good complement to the dynamometer tests.

The comparison in the CoF for each step is shown in Figure 10, highlighting a very good correlation as most of the CoF values are within 10% difference, and many variables are not the same, such as the inertia, or the effectiveness of the caliper. As expected, some of the more difficult-to-reproduce steps, such as the fade step (6.9), wherein the material temperature raises very rapidly, show a greater difference between the TriboLab and dynamometer results than other steps.

Figure 10. Comparison of the SAE J2522 from the full-scale dynamometer versus benchtop UMT TriboLab tests.

Conclusion

From the overall comparison of the full-scale test and benchtop test results, it has been demonstrated that the UMT TriboLab Brake Material Screening Tester is a reliable and effective tool for the screening and evaluation of friction materials used for brake applications. The advantages in size, flexibility and speed of the TriboLab make it a suitable complementary instrument to the dynamometer.

References

1. Fecher, Norbert, Jochen Thiesing, and Hermann Winner. "Caliper-independent investigation of brake pads." EuroBrake, EB2014-ST-006 (2014).

This information has been sourced, reviewed and adapted from materials provided by Bruker Nano Surfaces.

For more information on this source, please visit Bruker Nano Surfaces.