Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

BrakeTesting and
NVH Measurements
James K. Thompson
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH:
Testing and
Measurements
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Other SAE titles of interest:

Brake Technology Handbook

By Karlheinz Bill and Bert J. Breuer
(Product Code R-3375)

Brake Design and Safety, Third Edition

By Rudolf Limpert
(Product Code R-398)

Advanced Brake Technology

By Bert J. Breuer and Uwe Dausend
(Product Code: R-352)

Disc Brake Squeal

By Frank Chen, Chin An Tan, and Ronald L. Quaglia
(Product Code: R-353)

For more information or to order a book, contact SAE International at 400

Commonwealth Drive, Warrendale, PA 15096-0001, USA; phone 877-606-7323
(U.S. and Canada only) or 724-776-4970 (outside U.S. and Canada); fax 724-776-
0790; e-mail [email protected]; website http://books.sae.org.
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH:
Testing and
Measurements

James K. Thompson

Copyright © 2011 SAE International eISBN: 978-0-7680-5371-5

Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

400 Commonwealth Drive

Warrendale, PA 15096-0001 USA

E-mail: [email protected]
Phone: 877-606-7323 (inside USA and Canada)
724-776-4970 (outside USA)
Fax: 724-776-1615

Copyright © 2011 SAE International. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system, distributed,
or transmitted, in any form or by any means without the prior written permission
of SAE. For permission and licensing requests, contact SAE Permissions, 400
Commonwealth Drive, Warrendale, PA 15096-0001 USA; e-mail: [email protected];
phone: 724-772-4028; fax: 724-772-9765.

ISBN 978-0-7680-3480-6
SAE Order No. R-399

Library of Congress Cataloging-in-Publication Data

Thompson, James K., 1951–

Brake NVH : testing and measurements / James K. Thompson.
p. cm.
Includes index.
ISBN 978-0-7680-3480-6
1. Automobiles—Brakes—Testing. 2. Automobiles—Brakes—Noise. I. Title.
TL269.T48 2011
629.2’46—dc22
2010051176

Information contained in this work has been obtained by SAE International from
sources believed to be reliable. However, neither SAE International nor its authors
guarantee the accuracy or completeness of any information published herein and
neither SAE International nor its authors shall be responsible for any errors, omissions,
or damages arising out of use of this information. This work is published with the
understanding that SAE International and its authors are supplying information, but are
not attempting to render engineering or other professional services. If such services are
required, the assistance of an appropriate professional should be sought.

To purchase bulk quantities, please contact:

SAE Customer Service
E-mail: [email protected]
Phone: 877-606-7323 (inside USA and Canada)
724-776-4970 (outside USA)
Fax: 724-776-1615

Visit the SAE Bookstore at

http://store.sae.org
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

To my wife Shirley, without whom this book

and much else would not be possible.

v
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Contents

List of Figures......................................................................................................... ix

List of Tables.........................................................................................................xiii

Preface...................................................................................................................... xv

Acknowledgments.............................................................................................xvii

Chapter 1: Basic Noise and Vibration.......................................................... 1

1.1 Definition of Noise and Vibration..................................................................2
1.2 Basic Terminology............................................................................................2
1.3 Single-Degree-of-Freedom Model..................................................................6
1.4 Spectral Analysis...............................................................................................9
1.5 Acoustics..........................................................................................................13
1.6 Acoustic Environments..................................................................................19
1.7 Frequency Analysis.........................................................................................21
1.8 Types of Sound Signals...................................................................................25
1.9 Summary......................................................................................................... 26

Chapter 2: Brake Noise Characteristics.................................................. 27

2.1 Introduction.................................................................................................... 28
2.2 Why Is Brake NVH Important?.................................................................. 28
2.3 What Is Brake NVH?.....................................................................................29
2.4 Brake Squeal................................................................................................... 30
2.5 Moan and Groan.............................................................................................33
2.6 Judder................................................................................................................35
2.7 Summary..........................................................................................................37

Chapter 3: Brake NVH Measurements...................................................... 39

3.1 Introduction.................................................................................................... 40
3.2 Transducers..................................................................................................... 40
3.3 Data Acquisition and Analysis..................................................................... 43
3.4 Coherent Output Power.................................................................................47
3.5 Summary..........................................................................................................53

vii
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 4: Laboratory Testing...................................................................... 55

4.1 Introduction.................................................................................................... 56
4.2 Requirements for Brake Noise Testing in the Laboratory....................... 56
4.3 Brake NVH Dynamometers..........................................................................57
4.4 Electric Motor................................................................................................. 60
4.5 Mechanical Inertia and Inertia Simulation................................................61
4.6 Enclosures........................................................................................................63
4.7 NVH Dynamometer Special Considerations............................................ 66
4.8 Brake Noise Measurement Standards..........................................................67
4.9 Examples of Brake Dynamometers..............................................................78
4.10 Summary........................................................................................................85

Chapter 5: On-Road Testing........................................................................... 87

5.1 Introduction.................................................................................................... 88
5.2 Goals for On-Road Brake NVH Testing..................................................... 88
5.3 Vehicle Instrumentation............................................................................... 88
5.4 Vehicle Test Procedures ................................................................................91
5.5 On-Road Measurement Systems.................................................................103
5.6 Summary........................................................................................................108

Chapter 6: Pad Natural Frequency and Damping Testing.............109

6.1 Introduction...................................................................................................110
6.2 Standardization Process...............................................................................110
6.3 Summary....................................................................................................... 120

Chapter 7: Future Trends...............................................................................121

7.1 Introduction.................................................................................................. 122
7.2 Environmental Conditions......................................................................... 122
7.3 Low-Frequency Brake Noise Tests............................................................. 123
7.4 Representing Vehicle Effects....................................................................... 124
7.5 Judder Measurement.................................................................................... 124
7.6 Subjective Correlation................................................................................. 125
7.7 New Equipment............................................................................................ 126
7.8 Summary........................................................................................................127

References............................................................................................................. 129

Index......................................................................................................................... 131

About the Author................................................................................................ 135

Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

List of Figures
Chapter 1
1. Vehicle noise sources.
2. Simple mass and spring system (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
3. Relationship among acceleration, velocity, and displacement of vibration
(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).
4. Single-degree-of-freedom vibration model (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
5. The effect of mass changes on a one-degree-of-freedom system (Courtesy
Brüel & Kjær Sound & Vibration Measurements A/S).
6. One-degree-of-freedom system with damping (Courtesy Brüel & Kjær Sound
& Vibration Measurements A/S).
7. Multi-degree-of-freedom system model (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
8. Real-world vibrating system (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
9. Potential sources of vibration (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
10. Different types of sources of vibration (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
11. Sound as a vibration in air (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
12. Terminology of sound (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
13. Sound measurement parameters (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
14. Sound pressure versus sound power (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
15. The range of sound pressure amplitudes (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
16. Computing sound pressure levels (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
17. The A-weighting frequency correction.
18. Ideal sound fields (Courtesy Brüel & Kjær Sound & Vibration Measurements
A/S).
19. Moving from the time to the frequency domain (Courtesy Brüel & Kjær
Sound & Vibration Measurements A/S).

ix
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

20. Example of a real-world frequency spectrum (Courtesy Brüel & Kjær Sound
& Vibration Measurements A/S).
21. Example of a real-world time signal (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
22. A frequency spectrum analyzer (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
23. Common frequency analysis formats (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
24. Types of sound signals (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).

Chapter 2
1. The frequency range of common brake noises.
2. Example of typical brake rotor resonant frequencies.
3. An Example of a brake rotor bending mode shape.
4. Example of squeal during a test stop (Courtesy Link Engineering Company).
5. Time waveform of a moan noise (Courtesy Link Engineering Company).
6. Frequency spectrum of the example moan vibration (Courtesy Link
Engineering Company).

Chapter 3
1. Schematic of data Acquisition and analysis for brake NVH testing (Slide
4-16).
2. Brüel & Kjaer type 4189 microphone (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
3. Microphone directivity pattern (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
4. Brüel & Kjaer type 4393 accelerometer (Courtesy Brüel & Kjær Sound &
Vibration Measurements A/S).
5. The Brüel & Kjaer data acquisition and analysis system (Courtesy Brüel &
Kjær Sound & Vibration Measurements A/S).
6. Demonstration of the effects of averaging.
7. Example of a brake squeal spectrum (Courtesy Link Engineering Co.).
8. Typical city traffic summary report (Courtesy Link Engineering Co.).
9. Typical example of frequency data for brake NVH measurements.
10. Brake torque order plot.
11. Basic system diagram.
12. Coherence to example case (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).

x
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

13. Coherent power for example case (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
14. Comparison between sound pressure level spectrum and coherent output
power for example case (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).

Chapter 4
1. Brake NVH dynamometer drawing (Courtesy Link Engineering Co.).
2. Brake NVH dynamometer photo layout (Courtesy Link Engineering Co.).
3. Chassis dynamometer drawing (Courtesy Link Engineering Co.).
4. Chassis dynamometer designed for Brake NVH testing (Slide 4-8).
5. Example of the performance capabilities of a dynamometer using inertia
simulation (Courtesy Link Engineering Co.).
6. Typical brake NVH dynamometer enclosure (Courtesy Link Engineering
Co.).
7. Example test enclosure background noise levels.
8. SAE J2521 Pressure Apply Ramp Specification.
9. SAE J2521 Microphone Placement Specification.
10. Examples of different types of test fixtures (Courtesy Link Engineering Co.).
11. First page of Standard SAE J2521 report (Courtesy Link Engineering Co.).
12. Second page of Standard SAE J2521 report (Courtesy Link Engineering Co.).
13. Cumulative noise occurrence plot from SAE J2521 test (Courtesy Link
Engineering Co.).
14. Example of SAE J2521 test results: Noise after fade section (Courtesy Link
Engineering Co.).
15. Example of SAE J2521 test results: High noise occurrences (Courtesy Link
Engineering Co.).
16. Example of SAE J2521 test results: Strong noise suppressed after fade
(Courtesy Link Engineering Co.).
17. Example of SAE J2521 test results—Very noisy brake (Courtesy Link
Engineering Co.).
18. Link model 3900 and 4900 brake dynamometers (Courtesy Link Engineering
Co.).
19. Variations in test enclosures (Courtesy Link Engineering Co.).
20. Link model 4900 two-roll chassis dynamometer (Courtesy Link Engineering
Co.).
21. Link four-roll chassis dynamometer (Courtesy Link Engineering Co.).
22. Drawing of four-roll chassis dynamometer installation (Courtesy Link
Engineering Co.).

xi
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 5
1. Head and torso simulator for binaural sound measurements.
2. Two on-road brake NVH measurement systems (Courtesy Link Engineering
Co.).
3. Instrumented vehicle (Courtesy Link Engineering Co.).
4. SAE J2625 Example Results Report (Courtesy Link Engineering Co.).
5. Los Angeles City Traffic Route (Courtesy Link Engineering Co.).
6. Los Angeles City Traffic Route instrumentation (Courtesy Link Engineering
Co.)
7. Typical LACT Daily Summary Report page 1 (Courtesy Link Engineering
Co.).
8. Typical LACT Daily Summary Report page 2 (Courtesy Link Engineering
Co.).
9. Typical LACT Daily Summary Report page 3 (Courtesy Link Engineering
Co.).
10. Problem squeal testing vehicle corner instrumentation (Courtesy Link
Engineering Co.).
11. Problem squeal test results (Courtesy Link Engineering Co.).
12. Link model 3801 data acquisition system (Courtesy Link Engineering Co.).
13. Link model 3501 data acquisition and analysis system (Courtesy Link
Engineering Co.).
14. Link model 3501 configuration (Courtesy Link Engineering Co.).
15. Brüel & Kjaer NVH data acquisition and analysis module (Courtesy Brüel &
Kjær Sound & Vibration Measurements A/S).

Chapter 6
1. First round robin natural frequency results for a single brake pad.
2. First round robin damping results for a single brake pad.
3. Basic measurement arrangement for pad damping and natural frequency
test.
4. Instrumentation configuration for pad test.
5. Excerpt from SAE J2598 description of excitation and response measurement
locations.
6. Effect of foam on damping measurement.
7. Analyzer display for proper SAE J2598 measurement.
8. Improvement in natural frequency measurement with SAE J2598.
9. Improvement in damping measurement with SAE J2598.

xii
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

List of Tables
Chapter 1
1. Units of vibration amplitude (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).
2. Human perception of sound (Courtesy Brüel & Kjær Sound & Vibration
Measurements A/S).

Chapter 2
None

Chapter 3
None

Chapter 4
1. SAE J2521 Test Matrix summary.
2. Classification of brake NVH test fixtures (Courtesy Link Engineering Co.).

Chapter 5
1. Summary of SAE J2625 Test Matrix.

Chapter 6
1. SAE J2598 Gage R&R results with manual hammer.
2. SAE J2598 Gage R&R results with automated hammer.

xiii
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Preface
This book is intended to serve as an introduction to brake noise and vibration
testing and measurement for those new to the field. It should be useful to both
those familiar with vibration and acoustics and those who are just entering the
field. Those familiar with noise and vibration may want to skip or quickly scan
the first chapter to go directly into the discussion of brake-related noise and
vibration issues.

Despite many decades of work on the measurement of brake noise and vibration,
the field has only become mature in the last eight to 10 years. During this period,
standard test procedures have been established and some of the key issues to
generating consistent results have been documented. This text summarizes the
recent standards and practices that have led to successful measurements among
investigators worldwide.

In Chapter 1, a brief review of noise and vibration is provided for those not
familiar with the field. While this discussion is not a comprehensive course
in noise and vibration, it does review the basic concepts and much of the
terminology to be used throughout the text. Where possible concepts discussed
are related to brake noise and vibration issues to make the material as relevant as
possible.

A summary of common brake noise and vibration issues is provided in Chapter

2. The broad scope of brake noise issues is defined in terms of the frequency
range and conditions leading to various kinds of brake noise. Several specific
common brake noise and vibration issues are defined in detail. These detailed
descriptions also serve to highlight the approach to be used throughout the text
in defining measurements and the critical issues in consistent and reliable brake
noise and vibration tests.

The instrumentation, transducers, and other technical details of brake

measurement are described in Chapter 3. The critical considerations in selecting,
configuring, and validating measurement systems and transducers are described
in this section. In addition, background material is provide to explain why some
of the common transducer types and analyzer settings are not appropriate for
brake noise and vibration measurements.

Chapters 4 and 5 describe measurement practice in detail for laboratory and on-
road measurements, respectively. In each chapter, measurement standards and
examples of measurement results are used to illustrate best practices. Example
results are used to highlight the common characteristics and how to analyze
the results obtained. These chapters also contain extensive discussions of the
available and commonly used test equipment and measurement systems to

xv
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

illustrate the current capabilities and where different types of equipment can best
be used.

The development of a standard for the measurement of brake pad damping and
natural frequencies is described in detail in Chapter 6. This segment provides a
detailed discussion of the issues and best practices in such a measurement. The
reader is provided with a great deal of insight into the detailed issues in brake
vibration measurement and its relevance to radiated noise.

Finally, Chapter 7 describes the current trends in brake noise and vibration
measurements. This section highlights the upcoming modifications and test
requirements for brake noise and vibration in both the laboratory and on the
road. This discussion should be particularly helpful to those just beginning to
develop new laboratories or considering future updates of their laboratories.

xvi
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Acknowledgements
First and foremost, the author is indebted to the students who have attended his
brake NVH courses. Many of these students have provided invaluable feedback
on the structure and content of the course that is reflected in this text. The input
from these individuals is much appreciated.

The support and graphic materials provided by Link Engineering and Brüel &
Kjær are also much appreciated. Quite literally, this book would not be possible
without their contribution. In particular, the help and support of Roy Link and
Joe Burchill of Link Engineering are much appreciated.

The support of the SAE staff in composing and publishing this text is also much
appreciated.

xvii
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Introduction
Brake noise has been a problem since the earliest of vehicles. Whether the
mechanisms were leather against a wooden wheel, brakes acting on the wheels
of early trains, or modern passenger car disc brakes, braking a moving vehicle
requires the dissipation of large amounts of energy. In such a rotating system,
it is inevitable that a small fraction of this energy is turned into vibration and
potentially noise. With a sufficient sound radiating surface, even this small
amount of energy can produce significant sound levels. When one is dissipating
hundreds of watts of power, a few tenths of watts in vibration and noise may not
seem significant. However, even this amount of sound power can result in sound
levels of 100 dB or more.

As has been often observed, every vehicle in the movies exhibits brake squeal.
Brake noise can be seen from several perspectives: For some, the occurrence of
squeal is seen as welcome confirmation that the brakes are working. For others,
brake noise can be a minor irritation. Yet for others, brake noise may be seen as
an indication that the brakes are not working properly. In the worst case, brake
noise may lead to returns to the dealer and warranty claims.

Since the beginning of mass production of the automobile, brake noise has
been an annoyance and consequently there have been attempts to quantify and
understand the noise. While instrumentation has been available for brake noise
measurement since the 1940s, it took more than five decades to develop useful
and reliable measurement practices.

While brake noise is a common experience for those who ride in or operate
vehicles, it has proven difficult to recreate in the laboratory. Developing a test
apparatus that can reliably reproduce brake noise is a difficult undertaking.
One must recreate all the vehicle operating conditions and provide a sufficient
controlled acoustical environment to measure the resulting noise. The first
internationally recognized brake noise test standard was not published until
2001. Therefore the standards of measurement and instrumentation to be
discussed in this text will be those developed in the period from 2001 to
the present. The discussion that follows will deal with current practices and
procedures for quantifying and understanding brake noise.

xix
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 1
Basic Noise and Vibration

1
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

B
efore discussing the specifics of brake broader category and includes all sounds. This
noise and its measurement, it is necessary could be something pleasant like your favorite
to discuss the fundamentals of noise music or the birds on a nice summer day. Noise
and vibration. It is hoped that this material will on the other hand, refers only to unwanted
provide a solid foundation for the discussion to sounds. Figure 1.1 shows some of the common
follow for those who are not trained in acoustics or sources of vehicle sound or noise.
mechanical vibrations. Those who are so trained
may wish to skip this chapter or skim through it Noise could be associated with the aircraft fly
quickly on their way to the second chapter. over noise when you are trying to enjoy a nice
day outside. It should be evident that sound
and noise may overlap depending on context
1.1 Definition of Noise and Vibration and personal preferences. The music that one
The first step in defining brake NVH is to define person enjoys may be intolerably loud noise to
the term NVH. These initials stand for noise, the person next to you in traffic.
vibration, and harshness. These terms go back
to the history of automotive noise and vibration The important point is the perception of the
work. In early work to understand vehicle noise sound from a product can be important to
and vibration, it was clear that high frequency how it is valued by the customer. Just enough
items could be categorized as noise. Very sound can be seen as pleasant or an important
simply, this is what you heard. In the same vein, part of the character of the vehicle or other
structural vibrations at low frequency—those product. Just the right “exhaust note” is an
that you could only feel—were called vibration. important part of the customer’s perception of
However, there was the frequency range in a performance or sports car. Many customers
between where phenomena could be both felt have in mind a desirable amount of sound
and heard. A very common case in this category that is associated with proper operation of a
was the high frequency vibration that was felt product. However, too much sound can become
or low-frequency noise that was heard in ride noise and cause a decline in the perceived value
testing when the vehicle encounters sharp of the product. The noisy vacuum cleaner or
impacts such as pavement expansion joints. lawn mower are good examples of noise that
This category was named harshness since it felt many people have experienced.
harsh to the vehicle occupant and was heard
even when it was not recognized as noise. In the case of automobiles, the auto
This terminology remains with us today even manufacturers have done a great deal of
though modern measurements and analysis marketing to link low noise or quiet with
are able to very finely distinguish thousands of quality. Many car buyers associate a quiet ride
separate noise and vibration issues in the vehicle with a quality vehicle. The negative aspect of
development process. Today, NVH is used to this association in the even small noises can be
signify the entire range of noise and vibration seen as not only annoying, but also indicators
issues encountered in the vehicle. of poor quality or low value. In a very real sense
the right sound or the minimization of noise
can be crucial to the perceived value of a vehicle.
1.2 Basic Terminology
Speaking of NVH, the difference between noise There is a whole field of study related to
and sound should be explained. Sound is the the annoyance of sound and when a sound

2
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

Figure 1.1 Vehicle noise sources.

becomes noise to someone. For the purpose of Acceleration is the most common descriptor
this text the discussion will be limited to the because it is easily measured. There are
noise or unwanted aspects of brakes. numerous manufacturers of accelerometers
that can be easily attached to an object and
The next term to discuss is vibration. provide an electrical signal proportional to
Generally, vibration is defined with another acceleration. Accelerometers will be discussed
term: oscillation. It is important to note that in greater detail in a later section of this text.
oscillation or vibration refers to a repetitive
motion about some reference. It may be as Another measure of the magnitude of vibration
simple as a pendulum oscillating about its static is velocity. The vibration force can be computed
position. A good way to define vibration for the by multiplying the velocity and the damping
purpose of this text is repetitive mechanical constant for the object in question. The damping
motion about a reference point. Figure 1.2 constant is a mathematical simplification of
shows an example of vibration. the losses incurred during vibration in the
system. For brakes one often thinks of Coulomb
In this case, a mass and spring vibrate frictional damping as the primary loss term.
repetitively with sinusoidal motion. The damping coefficient, C, is an approximation
for the damping in the system. There are
To describe the motion of vibration, a number much more complex damping models, but this
of terms are used. The most common term to approach will be sufficient for the discussions
describe vibration is acceleration. As shown of this text [1, 2]. There are transducers for
in Figure 1.3, the force causing acceleration is measuring vibration velocity, but they are less
equal to the mass of the object multiplied by common and more difficult to use.
the acceleration.

3
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 1.2 Simple mass and spring system (Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

An important reason to consider velocity is stiffness, k. As with damping, the stiffness is

that it most directly correlates to the noise a mathematical approximation. It represents
produced when an object vibrates. As an object the stiffness of the system being studied. There
vibrates, it moves the air molecules around it certainly could be a spring involved, but it could
and this motion can become sound. A basic just as easily represent the stiffness of a beam or
measure of the sound generated by a vibrating some other structure in the system. In brakes,
surface like a brake rotor is to multiply the this could be the stiffness of the caliper bracket
surface area with the velocity. Depending on or the bending stiffness of the rotor.
the radiation efficiency of the surface, the sound
energy generated can be directly calculated in If one looks at a representation of the motion
this way. So, there is a need to sometimes look described by each of these measures, there is
beyond acceleration and consider the velocity a clear relationship between each of these as
of vibration. Velocity is also used in many shown in Figure 1.3. For most systems, the
specifications to describe allowable motion in acceleration can be described in its simplest
machines related to balance and reliability issues. terms as harmonic or sinusoidal motion. As
shown in Fig. 1.3, this means a regular pattern
Finally, displacement can also be used as a of repetition with the motion described by a
measure of vibration amplitude. In this case, sine function. The acceleration magnitude is
the force causing the vibration can be computed A. However, there is another term required
by multiplying the displacement and the spring to describe this motion. It is ω, the rotational

4
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

frequency of the motion. It can be defined as two ω. Also, due to the integration there is a phase
π multiplied by the frequency of oscillation, f. shift of 90o and the velocity is expressed in
terms of a cosine function. The displacement is
This is a critical point. To describe vibration and then the velocity divided by ω or acceleration
noise, two parameters are required to represent divided by ω2 with the appropriate phase shift.
amplitude and frequency. The amplitude alone
is of little value unless one knows at what As always in engineering, there are more than
frequency or how often it occurs. This will be one set of units to describe vibration amplitude.
discussed in greater detail in the next section. Table 1.1 shows the major units used.

Another important implication that can be Acceleration is commonly described in g’s,

drawn from Figure 1.3 is that there is a direct which represent multiples of the acceleration
relationship between each of these amplitude of gravity, 9.806 m/s2 (32.174 ft/s2). The most
parameters. Velocity is the time integral of common alternative is m/s2. This measure
acceleration. Displacement is the integral of is particularly useful for low amplitudes of
velocity and the double integral of acceleration. motion where one would be using fractions
In addition, since we defined acceleration as of a g. Of course, English units are also used
a sinusoidal quantity, this relationship can to describe these quantities, but SI units are
be expressed more simply. Velocity is the the most common. One unique unit that often
acceleration divided by the rotational frequency, causes confusion is a displacement specification

Figure 1.3 Relationship among acceleration, velocity, and displacement of vibration

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

5
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Table 1.1
units of vibration amplitude
(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S)

Measurement Metric units english units

Acceleration 1 m/s2 0.102g = 39.4 in/s2

Velocity 1 m/s 3.6 km/h = 39.4 in/s

Displacement 1m t mm = 39.4 in

in mils. In this case, mils refers to thousandths This model is a convenient mathematical
of an inch. Some reliability specifications still simplification of physical systems, but it provides
use this terminology. good insight into how physical systems perform.
If the mass is perturbed from its static potion on
the spring, sinusoidal motion results as shown
1.3 Single-Degree-of-Freedom Model in the figure. This motion can only be in one
There are various models used to describe direction or degree of freedom in this model.
vibrating systems. The most common is the
single-degree-of-freedom model as shown in There are several important concepts in
Figure 1.4. vibration illustrated in this model. Note

Figure 1.4 Single-degree-of-freedom vibration model

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

6
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

that the vibration is at a constant rotational its natural frequencies. Some common examples
frequency, ω, or phrased in another way, it include the guitar string, a tuning fork, the tires
has a constant period, T. It can be shown on a car, or a brake rotor. If you tap a brake rotor,
that T is the inverse of the frequency, f. The what happens? It vibrates and makes noise at a
period typically has the units of seconds. The distinct set of frequencies. These are the natural
frequency is then in terms of cycles per second frequencies of the rotor. This concept is so
or the international unit Hertz (Hz). It is also important that nearly 50% of the measurements
apparent from this figure that the frequency and analyses done for vibration are devoted to
of oscillation can be defined in terms of determining the system’s natural frequencies.
the physical parameters of the system. The The natural frequencies of an object or system
square root of the ratio of stiff ness to mass is a are an inherent property of the system. Please
fundamental quantity in vibration and is used note that some will use the term eigenfrequency
in understanding much more complex systems. in place of natural frequency. This term simply
comes from the early use of the German
Perhaps the most important point in this simple language in the early descriptions of this
model is that the system has a natural frequency. phenomenon that is still used by some today. It is
That is a frequency at which it will naturally the same as natural frequency.
vibrate. This is an inherent quality of all objects.
Every object or group of objects has one or more Examining Figure 1.5, it illustrates how an
natural frequencies. If the object or system is understanding of the one-degree-of-freedom
perturbed or acted on by a force or another system supports making changes to alter the
object, it will tend to vibrate at one or more of vibration.

Figure 1.5 The effect of mass changes on a one-degree-of-freedom system

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

7
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

In this case, the mass is increased. Based on the amplitude to decrease over time. The larger
the previous case, it is clear that the addition of the damping, the more quickly the motion
mass should lower the natural frequency, which decreases in amplitude. Eventually, if there is
is clearly demonstrated in the diagram. What no other forcing function the motion will stop.
would happen if the stiffness was increased
by using a different spring? Could you change If one looks closely at Figure 1.6, there are two
the spring and mass in such a way to keep the motion plots: one for only the C1 damping
natural frequency constant? and the other for the combination of C1 and
C2. The addition of damping has slightly
Based on this simple model it would seem that lowered the natural frequency. The damped
once a system starts vibrating it continues to do natural frequency is slightly lower than the
so forever. From common experience, it is clear (undamped) natural frequency.
that this is not the case. The plucked violin sting
or the tapped brake rotor vibration gradually The simple one-degree-of-freedom model
diminishes. There is clearly a loss term that has provided a good basis to understand the
prevents the motion from continuing forever. fundamentals of vibration. However, there is
As was noted above, this loss term is called seldom a one-degree-of-freedom system in
damping. Figure 1.6 provides an example of a the physical world. In fact, nearly everything
one-degree-of-freedom system with damping. is a multiple degree of systems with motions
in multiple directions and at multiple natural
The added damping in terms of the two frequencies occurring simultaneously as shown
damping coefficients C1 and C2 have caused in Figure 1.7.

Figure 1.6 One-degree-of-freedom system with damping

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

8
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

The multiple-degree-of-freedom model is still a spectral line at the single natural frequency. In
simplification of the physical object or system, the multiple-degree-of-freedom case, there are
but it better characterizes the vibration than the multiple peaks in the spectrum representing
one-degree-of-freedom model. It is important all the natural frequencies of the system. For
to note the multiple-degree-of-freedom model a more comprehensive treatment of vibrating
is simply a group of interconnected one-degree- systems, see Reference 1.
of-freedom models. The principles discussed
previously still apply. Adding mass will tend to
reduce one or more of the natural frequencies. 1.4 Spectral Analysis
If the stiffness is increased, the corresponding The next step in moving toward real physical
natural frequency will increase. systems is to consider an actual shaft vibration
case. As shown in Figure 1.8, this system
Below each of these models, a frequency involves a number of inputs and responses.
spectrum is shown. These graphs or spectra
show the amplitude or magnitude of vibration The total response of the system is a
versus the frequency of oscillation. This is a combination of input forces, the effect of
commonly used display for describing systems the response or mobility of the system,
and will be used throughout this text. Note that and the resulting vibration. This concept
for the single degree of freedom this is a single is fundamental to all work in noise and

Figure 1.7 Multi-degree-of-freedom system model

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

9
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 1.8 Real-world vibrating system (Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

vibration. This progression through the system vibration velocity by the input force. Thus, it
is normally called source, path, and receiver provides a measure of what will be the resulting
in acoustics. The input forces or source may vibration velocity for a given force input. With
be many different things. In a case like this proper analysis, one can go back to the mass,
shaft in a bearing, it may be as simple as just spring, damper models discussed previously to
an imbalance in the shaft creating a once-per- understand the response of the system in terms
shaft rotation forcing function. If there are of these basic parameters.
other pieces of equipment attached to the shaft,
the input forcing function can become quite The resulting vibration spectrum for such a
complex. For even this simple case, we can shaft is shown in Figure 1.9.
consider rotating unbalance, friction forces,
acoustic excitation, aerodynamic forces, etc. In this particular case, the results of the
bearing that may lead to response at less than
The response of the system can be measured the rotational frequency can be seen. The
and quantified in a number of ways. These unbalance in the shaft is shown to give rise to
measurements will be discussed in greater responses at the first several harmonics of the
detail in later segments of this text. The rotational speed—first through the fift h. In
mobility referred to here is a well-defined addition, a fan is attached to this shaft. With
quantity that is obtained by dividing the six blades, it naturally leads to response at six

10
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

times the rotational speed. There are also at what the sources of the spectral components
least two structures attached to the shaft that are. In fact, this is what much of vibration
have resonances between seven and 12 times analysis is about: trying to deduce the sources
the rotational speed. Finally, there is a gearbox of the frequency components seen in the
attached to the shaft. Due to the gear meshing response of a system. It is easy to imagine that
frequency, some slight misalignment, and this can be difficult to do in a complex system
backlash in the gears, a series of responses with many interconnected components.
is seen just above and just below the twelft h
harmonic of the rotational speed. Finally, as in Another important distinction that can be
many systems, there are frequency components made looking at this system is to divide the
from AC line frequency. Depending on where vibration into forced and resonance cases.
this system is located in the world, these Forced vibration can be at any frequency
spectral lines will be at multiples of 50 or 60 matching that of the forcing function. On the
Hz. In some cases, these spectral components other hand, the resonant vibration can only
may be simply electrical noise and in other occur at the actual system resonances. One
cases such spectral lines may represent input important characteristic is that the frequency
from devices like electric motors. of the forced vibrations that are being forced
by phenomena related to rotational speed will
With this spectrum and an understanding of change with rotational speed. The components
the components attached to the shaft, it is clear related to the fan, shaft unbalance, and the

Figure 1.9 Potential sources of vibration (Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

11
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

gear meshing are all examples of these. The rotating rotor, a bearing, the vanes in the
resonant vibration components will not change rotor are similar to a fan, there is a transaxle
frequency when the shaft rotational speed is or differential attached to the brake with
changed. meshing gears, and there are many components
like the caliper, pad, and suspension that
However, Figure 1.10 plots amplitude versus have resonances that can be seen in the
order of multiples of the rotational speeds. vibration. The forced response items seen in
the case of the brake include random forcing
In this case, the forced vibration components across a wide range of frequencies due to the
related to shaft rotational speed do not change friction interface. The tires, rotor imbalance,
order when the shaft speed is changed. The rotor vanes, and other components provide
resonant items will change relative order frequency functions related to rotational speed.
location with shaft speed changes since their There may also be speed related forcing inputs
frequency remains constant. Therefore, it from the transaxle of differential that can
is apparent that using order and absolute be seen in the brake components. There will
frequency analysis is one way to differentiate be resonant responses seen from the rotors,
resonant and forced inputs. caliper, anchor bracket, brake pad, wheels,
and suspension components. Clearly this is
Before leaving this section, it is important to a complex system, but the same techniques
relate this discussion to a vehicle brake system. discussed for the simple shaft are applicable
In the case of disc brake, there is clearly a with the brake system.

Figure 1.10 Different types of sources of vibration

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

12
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

Figure 1.11 Sound as a vibration in air (Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

1.5 Acoustics There are a lot of different terms used to

The basic concepts to understanding sound describe sound. Figure 1.12 shows some of
are very similar to those of vibration. The only these.
major differences are sound is vibration of the
air and one hears as opposed to feels it. Figure Typically measurements are related to sound
1.11 shows a good illustration of sound and its pressure, pascals, the variation in pressure
relationship to vibration. on top of barometric pressure, or sound
power, watts, the amount of power creating
The tuning fork vibrates primarily at its first the sound. Because sound covers a large
natural frequency. This vibration disturbs dynamic range it is often represented on
the particles of air. The vibration in the air a logarithmic scale using levels, which are
creates a wave of compressions and rarefactions designated by L with a subscript depending
that propagate and one hears as sound. The on the measure.
amplitude of this pressure wave is very small.
It is a perturbation on barometric pressure that It is important to understand the difference
requires very specialized pressure transducers between sound power and sound pressure.
and microphones to measure. Figure 1.13 provides a good illustration.

13
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 1.12 Terminology of sound (Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

Figure 1.13 Sound measurement parameters

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

14
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

Sound pressure is the magnitude of the pressure where:

variation. The magnitude of this pressure is
dependent on where it is measured. In general Lw is the sound power level, dB
it is higher near the source and decreases at
W is the sound power, W
greater distances. The sound pressure level is
used to quantify the pressure as shown below: Wref is the reference sound intensity, 1pW

Note that sound power is an inherent property

Lp = 10log10 (1.1) of the source and is not dependent on distance
or other external factors.
where:
It is important to make a distinction between
Lp is the sound pressure level, sound power and sound pressure. Figure 1.14
in decibels or dB provides a good illustration of this difference.
p is measured pressure in pascals Sound power is like the electrical power of
the heater. The heater generates the same
pref is the reference pressure,
amount of heat as long as the electrical power
20 micropascals (µPa)
is the same. In a small room it may raise the
The reference pressure is chosen to roughly temperature to a very high value, but in a very
correspond to the threshold of hearing, which large room it may not raise the temperature at
is the lowest level pressure variation that can all. In this analogy sound pressure is like the
be heard. The propagation of sound energy is temperature. The sound pressure is dependent
described as intensity or the power per unit on where it is measured and even the room in
area. The sound intensity level is defined as: which it is measured. Near the source the sound
pressure will be highest. At large distances
away, it may be much lower. The sound pressure
is also dependent on the room or acoustic
Li = 10log10 (1.2) environment. If the walls of the room are highly
where: reflective a large amount of sound energy can
build up in the room and the measured sound
Li is the sound intensity level, dB pressure levels will be high. If the walls of the
room are highly absorptive, the walls dissipate
I is the sound intensity, W/m2
the sound energy and then energy will not
Iref is the reference sound intensity, 1pW/m2 accumulate and the sound pressure level will be
lower. Note that in both of the previous cases
The sound power is then the amount of power the sound power is not changed.
from the source creating sound. The sound
power level is defined as: The range of pressure variations amplitudes
that are experienced as sound are very large.
Figure 1.15 provides a good illustration of this
Lw = 10log10 (1.3) range.

15
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 1.14 Sound pressure versus sound power

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

As noted previously, the 0 dB level represents value was chosen as a convenient pressure
the threshold of hearing. At roughly 130 dB or that was reasonable in magnitude and easily
20 Pa, the sound wave goes beyond hearing and standardized. The value of 124 dB is obtained
reaches the threshold of pain. Looking at the from a pressure of 31.7 Pa, which is not such
left-hand scale, the reason for the decibel scale a convenient number. However, in the early
is evident. Simply keeping track of the right days of acoustics it was felt to represent a
number of zeros after the decimal point can be reasonable maximum sound pressure that
difficult. The dB scale provides a much more would be measured in normal circumstances.
convenient way to quantify noise. Using this value supported the need with early
equipment to assure linearity over the potential
Using equation (1.1) it is possible to explain measurement range.
some of the common values seen in acoustics.
Nearly every microphone calibrator ever One of the difficulties of using the decibel
made uses one of two values for amplitude scale is that addition and subtraction is on a
calibration, 94 or 124 dB. The reason for this is logarithmic scale. One plus one does not equal
illustrated in Figure 1.16. two. Note: In the paragraph above, the sound
pressure was increased by 30 times and the
The level of 94 dB is equivalent to a sound level increased by 30 dB. This seems reasonably
pressure of 1 Pa. Clearly, this calibration close. On the other hand, if one is to add two

16
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

Figure 1.15 The range of sound pressure amplitudes

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

Figure 1.16 Computing sound pressure levels

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

17
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

signals with the level of 94 dB, the result is not 104 dB + 94 dB = 104.4 dB (1.5)
188 dB, but 97 dB.
or
94 dB + 94 dB = 97 dB (1.4) 10 Pa2 + 1 Pa2 = 11 Pa2

This result is simple to see using equation (1.1). In the field of acoustics it is sometimes
In this case p2 is 1+1 or 2 Pa2. Substituting 2 convenient to apply frequency weighting to
for p2, one obtains 97 dB. Also, since this is a sound measurements. These weighting factors
logarithmic calculation, it can be broken down or corrections can be applied for a number
to the sound pressure level for 1 Pa added to the of reasons. The most common correction is
sound pressure level for 2 (two times the squared A-weighting. Therefore, one may see a sound
amplitude). Noting that ten times the log of 2 is pressure level expressed as 87 dB or 87 dB(A).
3.01 dB is a convenient relationship to remember. These two values are not the same. Figure 1.17
shows the A-weighting curve and the additions
So, all decibel additions and subtractions must or subtractions to the actual values to be applied.
be made logarithmically. One convenient result
of this mathematics is that a level 10 dB less A sound level of 87 dB becomes 88.1 dB(A) at
than another is relatively inconsequential to the 3000 Hz. The reason A-weighting is used is to
total level. For instance, approximate the response of the human ear.

Figure 1.17 The A-weighting frequency correction

18
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

This does not mean it corresponds to how one An increase of 3 dB, a doubling of the sound
perceives sound. It is only a rough correction pressure squared, or the sound power, will
to match the response of the human ear. The be just perceptible. A change of 5 dB will be
A-weighting curve increases the sound level necessary to be noticeable and a change of 10
around 3000 Hz because this is the range of the dB will be required for a perception of twice or
first acoustic resonance in the ear canal between one-half as loud.
the outer ear and the ear drum. The important
concept to understand is that A-weighting It must be remembered these rules are only
is simply a correction factor for human ear rough approximations. In the case of brake
response. It is often used in engineering squeal where there may be one strong discrete
measurements due to the mistaken belief that it tone, smaller changes than 3 dB may be
better correlates to human response. perceptible. Changes of greater than 3 dB may
be required in low-frequency groan noise for
However, the important question to everyone the difference to be perceptible.
working in NVH is how are these level
changes perceived? Does a doubling of the
squared sound pressure, equivalent to a 3 dB 1.6 Acoustic Environments
increase, sound twice as loud? The answer is An important aspect of all sound
no. There is not a direct correlation between measurements and perception is the
the sound pressure level in decibels and the acoustic environment. Typically the acoustic
perceived magnitude of the noise. It turns out environment is compared to two extremes. The
this relationship is very complex and depends first extreme is the free-field environment. In
on sound pressure level, frequency range, this case, all sound from a source is dissipated.
frequency content, and many other factors. There are no reflections from walls or other
There are a few rules of thumb that are useful surfaces. Such an environment can be achieved
in estimating how changes in level will be in the outdoors with a source far away from any
perceived. Table 1.2 provides a summary of reflecting surface. An airplane flying at altitude
these rules of thumb. is a good example of such a case.

Table 1.2
Human perception of sound
(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S)

Change in Sound Level, dB Change in Perceived Loudness

3 Just perceptible

5 Noticeable difference

10 Twice (or 1/2) as loud

15 Large change

20 Four times (or 1/4) as loud

19
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

The other extreme is an environment Note that the free field is approximated in an
where there are strong reflections from anechoic (no echo) room (as pictured on the left
all the walls in a room. This is called a of Figure 1.18). Such rooms have wedges on all
diff use-field environment. While such a their surfaces that absorb the incident sound.
field is approximated in some real-world Such rooms must be carefully designed and are
environments, it is only really achieved in a at best a close approximation to the outdoor
specially constructed reverberation chamber. environment. They do have the advantage
The key attribute of such a sound field is that, that by their design they eliminate all other
because there are so many strong reflections, extraneous noise sources and they eliminate
the sound pressure level is uniform throughout the complications of weather and wind when
the space. Thus the reflected sound energy is precise measurements are desired.
equal to the energy coming from the source.
Clearly, this is an ideal that can never be For brake noise measurements, what acoustic
achieved since there will always be some small environment does one encounter? In a
energy loss with each reflection. However, well- general sense, the vehicle is outdoors, so the
designed, reverberant rooms come very close to environment is not a diff use field. However, the
this ideal and a diff use sound field assumption road surface is a good reflector of sound energy
works very well until one gets very close to the over many of the frequencies associated with
source. brake noise. This environment is called a free
field over a reflecting plane. Of course, there are
Figure 1.18 provides a good illustration of both reflections from the subject vehicle and other
of these types of rooms. vehicles on the road in a traffic situation. The

Figure 1.18 Ideal sound fields (Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

20
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

free field over a reflecting plane provides a good the time domain. The variation in amplitude
approximation for the majority of brake noise of the sound or vibration is recorded over
cases and has become the standard approach in time. As with the case of the single-degree-of-
the industry. The best way to achieve this sort freedom system (see Figures 1.4 and 1.5), when
of environment is a hemianechoic room, which the motion is relatively simple the time domain
has a concrete floor and absorptive wedges on data can provide a lot of information. However,
all other surfaces. This type of room and its with multiple-degree-of-freedom systems,
design will be discussed in greater detail in the time data are very complex and difficult
later chapters. to understand. Figure 1.19 provides a good
example of this difficulty.

1.7 Frequency Analysis Looking at the time waveform of vibration for

In the previous sections frequency spectra this complex machine does not reveal much
(Figures 1.7 through 1.9) were used to about what is occurring. The spectrum shown
describe both sound and vibration. It is in the figure illustrates the value of the data
important to understand a little more about in the frequency domain. In this case, the
frequency analysis and how it can be used in contribution by the various gears becomes
understanding noise and vibration. When one much clearer.
measures a dynamic signal one is measuring in

Figure 1.19 Moving from the time to the frequency domain

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

21
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

The same sort of clarification was achieved in accelerometer is used. The electrical signal from
the previous discussion of the rotating shaft the accelerometer must be conditioned and
and the various components contributing to its amplified to permit it to be processed by other
spectrum. Figure 1.20 shows this example. electronic systems.

Looking at the time waveform for this case This processing can provide a large number of
in Figure 1.21, it is very difficult to see the possible results. In this case, in the upper right
contributions from the different components. of the figure a simple average or maximum
amplitude is shown. In the lower right a
Clearly frequency domain analysis is important frequency spectrum is obtained. There are
to understanding noise and vibration. number ways of obtaining such a spectrum. In
the past a whole series of filters would be used
So, how does one convert from the time to to filter the spectrum and obtain the amplitude
the frequency domain? Figure 1.22 illustrates of acceleration over a small frequency range.
a measurement system that incorporates When the results from hundreds of these filters
frequency analysis. are combined one has a spectrum. This took a
lot of time and a lot of electronics.
The system begins with a transducer that
converts vibration into an electrical signal that With the Fast Fourier Transform (FFT)
can be recorded and analyzed. In this case, an algorithm implemented on a computer, the

Figure 1.20 Example of a real-world frequency spectrum

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

22
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

Figure 1.21 Example of a real-world time signal

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

Figure 1.22 A frequency spectrum analyzer

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

23
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

processing time can be extremely fast with looking at discrete frequency signals like brake
very fine resolution. A detailed discussion of squeal.
the FFT algorithm is beyond the scope of this
text. For more information on this topic, see The constant percentage bandwidth (CPB)
Reference 3. In very basic terms, a Fourier plot is very commonly used in noise control
series is used to approximate the magnitude engineering. It is usually seen as octave or 1/3
and phase of each of the frequency components octave formats where the frequency spectrum
in the signal. When properly applied, the FFT is divided into segments as defined by
can provide a very accurate computation of the international standards. This approach can be
frequency components in a complex signal. very effective in understanding complex signals
and how they may be perceived. The bands are
Some of the common formats used for too broad to show brake squeal, and therefore,
frequency analysis are shown in Figure 1.23. this type of plot is not used for brake analysis.

The most common form used in brake analysis The use of order tracking plots was shown in
is the FFT or discrete frequency plot shown at the previous discussion. By representing the
the top of the figure. This is the best format for frequency spectrum in terms of multiples of

Figure 1.23 Common frequency analysis formats

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

24
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Basic Noise and Vibration

the shaft rotational speed, this type of plot been developed to identify the type of signal
can be very helpful in understanding some and in many cases to help understand how it
noise and vibration issues. Such plots are used can be best be analyzed. Figure 1.24 provides
extensively in study brake judder issues. They a good illustration of the common types of
can clearly show the effects of manufacturing signals.
discrepancies, run-out, DTV, and other
problems. One of the first steps is to distinguish between
stationary and non-stationary signals.
The final plot in Figure 1.23—while useful in Stationary signals are those signals that are
some noise and vibration investigations—is statistically consistent. If they change, they
not commonly used in brake noise problems. do so in a repetitive pattern. These types
In the past, it was sometimes used when older of signals can be analyzed using the FFT
instruments had trouble representing the algorithm discussed previously. Non-stationary
full dynamic ranges of wide-frequency brake signals are not statistically consistent. Special
measurements. With modern measurement techniques are required for their analysis.
systems, this type of analysis is no longer Stationary signals can again be divided into
necessary. deterministic signals and random signals, and
non-stationary signals into continuous and
transient signals.
1.8 Types of Sound Signals
Because of the complexity of sound and Stationary deterministic signals are made up
vibration signals, various classifications have entirely of sinusoidal components at discrete

Figure 1.24 Types of sound signals (Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

25
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

frequencies. Such signals rarely occur in the following basic concepts: it should be clear what
real world. Random signals are characterized NVH is and that it is an important part of how
as signals where the instantaneous value customers perceive a product; the fundamental
cannot be predicted, but where the values quantities and terms describing noise and
can be characterized by a certain probability vibration should be well understood; and it
density function, i.e., one can measure its should be clear that adding tow signals at 80 dB
average value. Random signals have a frequency does not give 160 dB.
spectrum that is continuously distributed with
frequency. Brake squeal is best approximated The important distinction between forced and
as one or more discrete frequency tones in a resonant vibration should also be clear. This
stationary random signal. is a critical concept in understanding and
measuring brake noise. Since brake noise is
The continuous non-stationary signal has some most often reported as a sound pressure level,
similarities with both transient and stationary it is important to understand this quantity
signals. During analysis, continuous non- and the effects of the ways in which it is
stationary signals should normally be treated as measured. As an example, understanding that
random signals or separated into the individual sound pressure levels measured at different
transients and treated as transients. Transient distances may be different even if the brakes
signals are defined as signals that commence being studied have the same sound power can
and finish at a constant level—normally zero— be critical to properly understanding such
within the analysis time. A good example measurements.
of transient signals is the creep groan noise
generated during low-speed vehicle creep. Perhaps the most important concept of this
chapter is the source, path, and receiver model
for understanding sound and vibration issues.
1.9 Summary Understanding and being able to use this model
This discussion has covered a broad range is crucial to making any noise measurement
of topics related to noise and vibration. The and interpreting the results. In subsequent
reader should have a good understanding of the chapters this concept will be used repeatedly.

26
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 2
Brake Noise Characteristics

27
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

2.1 Introduction normal when brakes were working effectively.

In this chapter the focus will shift from general Many older drivers today are surprised when
aspects of noise to a specific study of brake someone points out that they have brake squeal.
noise. Several common brake noise issues will They tend to disregard it as normal operation.
be defined and described with examples. While
there are too many different brake noises to However, as vehicles have gotten quieter in
describe, all the important and most common general, the majority of car buyers have come to
types will be addressed. expect quiet operation. Brake noise issues that
would have been considered insignificant two
decades ago, are now major problems that must
2.2 Why is Brake NVH Important? be corrected. Today’s vehicle owner considers
This is a fundamental question that needs to an unusual or intermittent noise to be a sign
be addressed. Clearly brake noise, like many of a problem or vehicle defect. If the owner is
other vehicle noises, can be annoying, but what able to associate this noise or vibration with
makes brake noise important? The fundamental brake operation, then he or she is even more
justification for the time and expense devoted concerned since the braking is a critical safety
to brake noise is the warranty cost. It is hard to item in the vehicle. These combined premises—
obtain specific numbers, but estimates are that noise and vibration indicate poor quality or a
the cost of brake NVH warranty issues for the defect, and brakes are a critical safety system—
U.S. automakers is in excess of $100 million. lead to strong reactions on the part of the
This is an excellent financial incentive to drive consumer.
noise control and research activities to reduce
or eliminate brake noise. A good case in point is an automobile company
that had a significant noise problem with the
Another driving force for the importance of rear drum brakes during low-speed stops.
brake noise is the J. D. Power surveys. For a The problem was a pronounced noise, but the
number of years, these customer surveys have brakes worked fine with no loss in performance
shown NVH to be the number one brake- or functionality. Every vehicle of this particular
related problem cited by customers. The relative model made over a series of months had this
rankings between various NVH components problem. The result was that one in ten buyers
have changed, but these issues have been brought the vehicle back to the dealer within
consistently the highest and typically much hours of taking receipt of the vehicle saying
higher than any other issue. they would not drive the vehicle because the
brakes were obviously defective. In the final
The important question related to both of these analysis, nearly one-third of these vehicles had
drivers is why is brake noise so important to been brought in for service or had a warranty
the customer? The definitive answer is very claim filed before the issue was resolved.
complex, but there are a few important issues
in considering vehicle owners’ and occupants’ In the previous chapter, it was noted that the
responses to brake noise. The first issue is that right amount or character of noise could be
the automakers have done a lot of marketing instrumental in the perception of a product.
to support the idea that noise is closely related Brake noise and vibration are excellent
to quality and proper operation. Two decades examples of where sound quality combined
ago, brake noise was annoying, but accepted as with concerns about safety lead to strong

28
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake Noise Characteristics

reactions on the part of the customer. This is description of numerous brake noises can
why brake NVH is important. be found in Reference 4. Other brake NVH
issues will have many similarities to the items
discussed in this text and many of the same
2.3 What is Brake NVH? measurement approaches will still apply.
The most common noise associated with
brake NVH is squeal. However, brake NVH One way to look at brake NVH is to consider
is actually a wide range of noise and vibration the frequency range it covers. Figure 2.1 is a
issues that occur over a broad spectrum of good overview of the spectrum of common
frequencies. It is the only vehicle NVH issue brake noise and vibration issues.
that spans the range from below the frequencies
of hearing to the ultrasonic region. In the ideal case people can hear noises ranging
from 20 to 20,000 Hz. It is clear that brake
Brake NVH can occur over a wide range of noise covers this entire range. The moan and
vehicle operating conditions from fractional groan noises occur at the low end of this range
speeds of less than a km/h (mph) to highways and the various types of squeal cover the rest of
speeds. Some brake noise or vibration the range to the upper limits of human hearing.
issues only occur when a combination of Disc brake squeal is a common issue. However,
conditions are present, such as temperature, drum brakes also make noise with howl being a
speed, deceleration rate, and environmental type of noise similar to squeal that is unique to
conditions. The result is a large number of drum brakes.
issues that cannot all be treated in detail in
this text. Instead the most prevalent issues Most adults have difficulty hearing noises
will be discussed to ensure that the most likely above 18,000 Hz, so noises above this range
issues are properly considered. An excellent are not given much consideration even though

Figure 2.1 The frequency range of common brake noises.

29
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

they are known to be present. At the opposite temperature, speed, clamping pressure, and
end of the spectrum there are some brake braking history.
NVH issues that are below human hearing.
Brake judder is such a case. It can be felt, but Going back to the source, path, and receiver
it is too low in frequency to be heard. It should model discussed previously, one can
be evident from Figure 2.1 that brake NVH characterize a squeal in quite clear terms. The
is not one issue, but numerous types of noise source is the brake rotor. While there may be
and vibration that need to be considered and a particular squeal where the noise is radiated
measured carefully. from some other part of the brake, the vast
majority of squeals originate from vibration
by the rotor in a disc brake and the drum in a
2.4 Brake Squeal drum brake. For significant squeals, the rotor
The most commonly considered brake NVH will be vibrating at one or more of its natural
issue is brake squeal. Every vehicle that stops frequencies. As was discussed before, vibration
in a television show or a movie makes brake at a natural frequency will produce the
squeal. This is part of the “realism” of the show, greatest amplitudes, or in this case, the highest
even if it represents an unreal case. In general, amplitude squeals.
brake squeal is defined as one or more pure
tones occurring between 1,000 and 18,000 For squeals above 2000 Hz, 100% of the sound
Hz. The term pure tone is used to indicate energy follows an airborne path to the vehicle
a sharp spectral peak over a very narrow interior. That is, the sound is radiated from the
frequency range, as opposed to a broad rise in brake, travels through the air, and enters the
the spectrum covering hundreds of hertz. A vehicle interior passing through seals, window
significant squeal can range in amplitude from glass, or other paths.
60 to 120 dB when measured within a meter of
the brake. A squeal results from the vibration The receiver for all of the cases of brake noise
of the brake components. Loud squeals will discussed in this text will be the vehicle
be caused by large amplitude vibrations. It is occupant. There are certainly cases where the
not uncommon to see brake caliper or pad noise radiated to the environment by brake
vibration amplitudes as high as 20 g’s during a operation is a concern. Anyone who has stood
significant squeal. on street corner or been in traffic with a bus or
truck stopping next to them can attest to the
The reason that squeal occurs is still a noise radiated to the environment. While this
matter of some conjecture by researchers. environmental noise is an issue, the focus of
The most commonly accepted explanation this text will be the interior noise.
is that the instability at the friction creates
large forcing functions. In addition, at the It is important to take note of the fact that
right conditions due to instability in the one of the reasons that brake squeal is such
system there is a coalescing of the modes of a noise problem is that it is very annoying
the individual system components including to the vehicle occupant. It is a pure tone at
the rotor, pad, and caliper that leads to high frequencies. This sort of tone is not
large vibration responses. The necessary common in any other situation and therefore
conditions for these phenomena can include stands out from other vehicle noises. It can
ambient temperature and humidity, brake be distinguished even when it is no higher in

30
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake Noise Characteristics

amplitude than other sounds in the vehicle. Clearly avoiding one of the rotor resonances
This unnatural sound only occurs during part is no easy matter. As a rule of thumb the
of a stop. resonances of the rotor are roughly 1000 Hz
apart. Figure 2.2 is a perfect example of how it
Considering each of the regimes noted above is nearly impossible avoiding some combination
will allow a better understanding of brake of temperature, load, etc that does not lead to a
squeal. It would be simple to prevent brake match with one of the rotor resonances.
squeal. One just has to shift the frequency of
excitation coming from the friction interface Why are the rotor resonances so closely spaced
away from the resonances of the rotor. Figures and consistent? This is the result of the regular
2.2 and 2.3 provide examples of the resonant geometry of the rotor. Such regularity leads
frequencies and the modes of vibration for a to uniformly spaced modes. The brake rotor
typical brake rotor. or drum can be thought of as a bell. Their

Figure 2.2 Example of typical brake rotor resonant frequencies.

31
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 2.3 An example of a brake rotor bending mode shape.

regular structure and low damping means in frequency. By its nature, it is short lived.
they are easy to excite and vibrate quite well at A number of short-event squeals are called
resonances with a long decay time. Like a bell, squeaks. Finally, there is wire brush noise,
they have large flat surfaces that are excellent which is several simultaneous squeals usually
sound radiators. If one set out to make the best occurring at higher frequencies typically 10,000
possible noise radiator, one could not do much Hz or higher.
better than the brake rotor.
Figure 2.4 shows an example of the data
As has been noted, there are a number of collected during a stop in which there was a
different categories of squeal. The following squeal.
is a description of just a few. The apply squeal
occurs with the initial brake application; it is In this case, a waterfall plot is used to show the
usually of short duration. The chirp is another entire duration of the stop. The horizontal scale
short-duration squeal. It often repeats with is frequency, the vertical scale is sound pressure
the wheel rotation and is usually at a fi xed level in dB, and the scale into the page is time
frequency. Pinch-out squeal occurs at the in seconds from the beginning of the brake
end of a stop. It usually occurs just before the apply. In this case, the occurrence of squeal is
vehicle reaches zero velocity and may vary quite easy to see. It begins about four seconds

32
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake Noise Characteristics

into the stop and ceases after two seconds. area must be moving to produce loud sounds.
This illustrates how squeals stand out from the Consider that for the same acceleration, the
background noise and that they do not occur velocity at 200 Hz will be 75 times smaller than
uniformly through a stop. that at 15,000 Hz. This is due to the fact that
velocity is acceleration divided by the rotational
frequency as discussed in Chapter 1. This
2.5 Moan and Groan means that there must be participation by the
At the lower end of the frequency range, well suspension and possibly vehicle structure for
below the squeal frequencies, moan and groan moan and groan noise to become significant.
occur. Because moan and groan are at such
low frequencies they involve more than just The mechanism for generating this type of
the vibration of the brake parts. To generate noise is also different than for squeal. It is
significant noise at low frequencies where believed that this type of noise is the result of a
velocities are relatively low, significant surface stick-slip type motion at the friction interface.

Figure 2.4 Example of squeal during a test stop (Courtesy Link Engineering Co.).

33
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

A good example of this sort of motion occurs from squeal. In this case, the vehicle structure
with a vehicle on a hill if the brake pressure is or suspension is the source. While the vibration
lessened until the vehicle just starts to move. at the friction interface starts the process, the
There is often a little motion and the brakes brake by itself cannot generate enough noise
grab. Under the right conditions, this process to be an issue. One or more parts of the vehicle
will repeat itself and be accompanied by low- structure or suspension must respond to this
frequency noise resulting from the stick-slip at input from the brake vibration for there to
the friction interface. In this classic case, the be significant noise. In general, this means
vehicle speed is less than 1 km/h (0.62 mph). there must be a resonance in one or more
components at the frequency at which input
There are different types of moan and groan is being provided by the brake vibration. This
that can occur. Some people use the term hum could be a suspension component, an axle, a
to refer to a highway speed moan or groan. This body panel, or any number of sources.
can be due to light brake drag that can set up
the stick-slip process, although at much higher The path is purely structural. The brake
speeds. mechanical energy travels into the suspension.
It travels from the suspension into the vehicle
In all these cases the resulting noise is in the structure, the steering system, and other
100 to 400 Hz range. For the most, part moan component to the vehicle interior panels. It is
and groan tend to happen at low speeds, the vibration of these interior panels that the
generally less than 20 km/h (12.4 mph). vehicle occupant—the receiver—hears as noise.

Applying the standard source, path, and receive An example of a low-speed moan occurrence is
model to this case, there are some differences shown in Figure 2.5.

Figure 2.5 Time waveform of a moan noise (Courtesy Link Engineering Co.).

34
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake Noise Characteristics

While the noise or vibration clearly repeats over hear judder because it is too low in frequency.
time, it is not perfectly consistent. Looking at It is sensed as a vibration. As will be discussed,
this time waveform, it is difficult to understand judder is not a large amplitude vibration;
much about the vibration. the manner in which judder manifests itself
leads to even small vibrations being seen as a
A much more useful look at this phenomenon significant problem.
is shown in Figure 2.6.
Before going any further, it is important to note
In this case, frequency analysis has been that there are two types of judder: hot and cold
performed that shows, despite the time judder. Hot judder occurs due to heavy usage of
variation, the vibration is centered at a the brake that results in uneven heating or hot
particular frequency just above 300 Hz. This spots that cause the distortion of the rotor of a
frequency plot represents only one of the bursts disc brake. This distortion of the rotor causes
of vibration from the previous figure, but it is brake torque variation (BTV) that leads to
typical. Clearly to get such a strong peak with vibration sensed by the vehicle driver. This is a
no broadening of the vibration and no other much more common problem in Europe where
nearby peaks, there is a strong resonance at this higher driving speeds and more aggressive
frequency. friction materials contribute to this problem.

Cold judder can be caused by a number of

2.6 Judder mechanisms, the most common being that off-
As the last example case, judder is the most brake drag can led to uneven wear of the rotor.
extreme case. It is not a noise. One does not Uneven wear leads to disc thickness variation

Figure 2.6 Frequency spectrum of the example moan vibration (Courtesy Link Engineering Co.).

35
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

(DTV) as the wheel rotates. The thickness or not working properly. It is easy to imagine
variation leads again to torque variation that is that this may be disconcerting to the driver. At
sensed as vibration by the driver. This is a much the time of judder, the other vehicle occupants
more common source of judder in the U.S. and may have no sensation of this pulsation or
Japan. Similarly, DTV and BTV can be the significant vibration.
result of corrosion on the rotor. The mechanism
for this corrosion can be simply due to a The second major path for the mechanical
vehicle sitting parked for a long period of time energy of judder is through the steering
in conditions that lead to corrosion to much mechanism and into the steering wheel. When
more complex circumstances involving vehicle judder occurs, the driver may feel a small
operation where snow clearing chemicals, oscillation in the steering wheel. In problem
heating cycles, and brake usage can lead to the cases the driver may feel the oscillation and see
accumulation of DTV. the steering wheel oscillating back and forth.
This too can be a source of concern for the
For the purposes of this discussion, the root driver. In extreme cases, the driver may feel
cause of the DTV or BTV is not critical. For that he is losing control of the vehicle.
any of these mechanisms, the end result
is a vibration in the 10-50 Hz range that is The final path of concern is through the
dependent on vehicle speed. Since the variation suspension, into the structure, and to the
is related to the rotor rotational speed, order driver’s seat track. In this instance, the driver
analysis is sometimes used in the study of feels the vibration in his seat. Depending
judder. on the severity of the judder, this may range
from being hardly noticed to a very upsetting
From this discussion, it should be clear that experience.
the source in the source, path, and receiver
analysis is brake torque variation. Clearly this If one considers the common scenario, it is
variation drives all that follows. The path is easy to understand how a small amount of
clearly structural. At such low frequencies vibration can become a major warranty issue.
acoustic radiation is very inefficient. As in all The best example is a driver making a trip on
other cases, the receiver is the vehicle occupant. the interstate highway system in the U.S. He or
However, in this case it is often only the driver she may drive for hours with little or no brake
who experiences a sensation of judder. This is usage. All of this time DTV can be building up
not typical of the other noise sources and is on brake rotors. After a considerable period like
related to the critical paths for judder. this, the driver has to apply the brake to achieve
significant deceleration for whatever reason.
Typically there are three paths that As he or she does this, there is a pulsation in
are considered the most important in the brake pedal. The driver reduces pressure
understanding judder. One path is through on the pedal because something does not
the brake components and suspension up to feel right. The driver feels the vehicle is not
the brake pedal. In many instances, the driver stopping properly and applies more pressure to
“feels” judder as a pulsation in the brake pedal. the brake. This time the driver notices that the
As noted above, this is not a large amplitude steering wheel seems to be pulling left and right
vibration, but to the driver it may seem like the and there is actual movement. As the driver
brake is pushing back against his or her foot pushes harder on the brake, the seat is vibrating

36
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake Noise Characteristics

and it may feel like he or she cannot properly generate major reactions from a customer. This
apply the brake and hold the steering wheel is a problem of high priority for the automakers
with the seat shaking. for which considerable money and effort has
and will continue to be devoted.
It is not hard to imagine that such an incident
could upset the driver. If the braking maneuver It has been shown that brake NVH is not one
is completed successfully, the problem may be problem, but a large number of noise and
forgotten or put off until the next time that it vibration issues. Brake NVH covers a wider
happens. In the worst case scenario, the driver frequency range than any other vehicle NVH
may be afraid to fully apply the brakes and issue. Brake squeal is the most common brake
have an accident or a near miss. In the end, it noise and the most commonly thought of when
is clear that this could lead to a warranty issue someone mentions brake noise. However, there
and a complaint that the vehicle is not safe. are many other issues for concern. Some of the
Studies have shown that small oscillations of issues do not even produce noise, but result in
the steering wheel and almost imperceptible small vibrations that are very disconcerting to
pulsation in the pedal generate major concerns the driver.
for the driver in such situations.
Just a few years ago, there was no question
An important characteristic of judder is that the number one brake NVH issue was
that it is highly vehicle dependent. Since squeal. It was the highest ranking problem
the mechanical energy travels through the in the J. D. Power surveys and the number
structure and interacts with the steering, pedal, one brake warranty issue. This situation has
and seats, there can be major differences in changed somewhat. The occurrence of squeal
how the same amount of DTV or BTV excites has been diminished. Because of this reduction
these systems. A level of DTV that is perfectly in squeal, some lower frequency issues in the
acceptable in one vehicle may lead to a major category of roughness have gained increasing
problem on another. Judder can be a very importance. This has resulted in the need
difficult NVH issue. for new tests and different approaches in
evaluating brake NVH.

2.7 Summary Considering how to reproduce and then

It is clear the brake NVH is a major issue in measure these noises and vibrations will be the
terms of cost and vehicle perception. Noise and subject of the next chapter.
vibration related to the braking system can

37
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 3
Brake NVH Measurements

39
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

3.1 Introduction to calibrated time or frequency domain data

Whether measurements are to be done in a for storage and further analysis. This chapter
laboratory or on the road, the key parts of any will look at each of these components and
NVH measurement system are the transducers, provide sufficient background for the reader to
the data acquisition, and data analysis systems. understand how they are used in a brake NVH
The transducers are the first element in the measurement system.
measurement chain. As shown in Figure 3.1,
the collection of NVH signals in brake testing
3.2 Transducers
is done with a microphone for sound and an
accelerometer for acceleration. The data acquisition system begins with the
transducers used to measure the noise and
Each of these transducers is very effective and vibration. There are a large number of suppliers
highly specialized. Both the accelerometer of accelerometers and microphones in the
and microphone need a preamplifier or a market place. A typical microphone is the B&K
signal conditioner to bring their output signals type 4189 with the type 2671 preamplifier as
to sufficient voltage to be used as input to a shown in Figure 3.2.
frequency analyzer. These transducers convert
sound or vibration into electrical signals that This microphone provides accurate
can be acquired and processed in a manner to measurements from 6.3 to 20,000 Hz with a
understand the physical phenomena. The data dynamic range from 14.6 to 146 dB, which
acquisition and analysis systems convert the is more than sufficient for brake noise
raw time based signals from the transducers measurements.

Figure 3.1 Schematic of data acquisition and analysis for brake NVH testing.

40
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH Measurements

The key points for selecting a microphone for

use for brake noise testing are the frequency
range, the appropriate sound field, and
accuracy. Since frequencies of up to 18,000
Hz are of interest if one is measuring squeal,
the microphone should meet or exceed this
range. A minimum requirement is a 20 Hz to
18,000 Hz frequency range. Microphones are
typically designed for optimum performance
in a particular sound field. It is, therefore,
important to select a microphone for use
in free-field environments. The brake on
vehicles operating outdoors is in a free-field
environment and hence the requirement for
measurement microphones. The most basic
parameter for microphone accuracy is its
classification according to the IEC standards
[5]. For accurate measurements out to squeal
frequencies, a type 1 microphone should be
Figure 3.2 Brüel & Kjaer type 4189 microphone
used. The B&K type 4189 meets all of these
(Courtesy Brüel & Kjær Sound & Vibration
criteria.
Measurements A/S).
However, the choice of microphone does
present some issues of which the user should
patterns depending on the modes of vibration
be aware. This free-field microphone does not
generating noise. It is therefore very difficult to
have equal sensitivity in all directions above
predict an angle of arrival.
approximately 8000 Hz. In the range of many
brake squeal issues, there can be differences of
Until microphones are available that avoid
as much as 10 dB in the microphone response
this problem, one can only be aware of the
depending on the direction of the sound
limitation and consider that higher frequency
source. Figure 3.3 shows a good example of the
sounds may be underestimated by a free-field
directional sensitivity of such a microphone.
microphone. Commercial microphones are just
beginning to be available that address this issue
At the higher frequencies where brake squeal
and reduce the sensitivity to angle of arrival at
is found, the free-field microphone is very
high frequencies.
sensitive to direction from which sound
arrives. It is important to keep in mind that
A typical accelerometer used for brake NVH
the direction of arrival of sound signals can
testing is the B&K type 4393 as shown in
be complex. The measurement configuration
Figure 3.4.
is one important factor. There could be
significant differences for a microphone used
This unit provides a minimum usable
for vehicle interior noise measurements and
frequency range of 0.1 to 16,500 Hz with a
one used for measurements near the brake.
maximum amplitude of 5000 g’s.
Also, each brake has different sound radiation

41
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 3.3 Microphone directivity pattern

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

The most important issue in selecting an not. Therefore, using accelerometers with an
accelerometer for brake measurements is the appropriate frequency range is very important.
frequency range of the unit. A significant
portion of brake NVH measurements are Another important consideration for
made with accelerometers that do not have accelerometers is the type to be used. There
sufficient frequency range. It is not uncommon are two common types in use. The oldest type
to find brake squeal measurements made with is the charge accelerometer. This type requires
an accelerometer with a maximum usable special impedance-matching amplifiers to get
frequency of only 9000 to 12,000 Hz. This
means that many higher frequency squeal
vibrations will be inaccurately measured. One
of the unique characteristics of accelerometers
is that at frequencies higher than the usable
range amplitudes are reported as higher than
they actually are. This means that many people
do not realize that they are getting erroneous
data in such circumstances. So, even though
an accelerometer cannot accurately measure
accelerations above 10,000 Hz, it may show
peaks in the frequency spectrum at 12,000
or even 16,000 Hz. The accelerations at these Figure 3.4 Brüel & Kjaer type 4393 accelerometer
frequencies may be over estimated by more (Courtesy Brüel & Kjær Sound & Vibration
than 100%. The tester may be deceived into Measurements A/S).
thinking that a squeal exists where it does

42
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH Measurements

an accurate signal proportional to acceleration. Typically, they will have to sample at 2.56 times
The more modern accelerometers use a small the highest analysis frequency, which would
amplifier integrated into the accelerometer and be roughly 50,000 Hz. Also, it is important the
are referred to as integrated electronics piezo analyzer be able to perform the data acquisition
electric (IEPE) types. For brake testing it is and analysis in real time. This means that the
best to use the charge-type accelerometers. The analyzer can acquire and process the data
reason for this is that the IEPE accelerometer before the next sample is completed. This
amplifiers have a typical maximum prevents any data from being lost. In older
temperature of 125oC (257 oF). Since brake analyzers, and some of the less expensive
temperatures, even for noise testing, can exceed current models, it is not possible to perform
600oC (1112oF), these types of accelerometers this analysis in real time. When some data
can be damaged or rendered inoperative at such is lost, by Murphy’s Law it is always the data
extreme temperatures. In practice, a typical sample containing the squeal of interest.
approach is to use the type 4393 with the type
2647 in-line amplifier that can be located Also, a Hanning window function is applied to
up to several feet from the brake to avoid the sampled data. This window is designed for
temperature issues. random noise type applications and it provides
good accuracy when analyzing random noise
The B&K type 4393 is an excellent choice type signals. With modern analyzers, this is
for brake NVH testing. This charge-type simply a selection made in the set-up process.
accelerometer has a useable temperature range All the necessary calculations are done in
from -74 to 250oC (-101.2 to 482o F). It has a software for the user. For a more complete
wide frequency range from 0.1 to 16,500 Hz. description of windowing functions and why
The temperature range is typical of charge-type they are used see Bendat and Piersol [6].
accelerometers from several vendors. However,
the frequency range is quite variable between Generally to give even higher confidence in
vendors and models. acquiring squeal data overlap sampling is used.
Typically a 50% overlap is used. That is, each
sample is a combination of 50% of the previous
3.3 Data Acquisition and Analysis sample and 50% of the next sample. While
There are many commercially available some may argue for large overlaps, experience
analyzers that have the basic capability to shows that 50% is a good compromise between
perform the fundamental data acquisition data quality and the ability to catch sharp rises
and analysis tasks required for brake testing. and falls in squeal tones. Figure 3.5 shows an
However, there are some unique requirements example of an analyzer that can provide the
for brake noise analysis that need to be required capabilities.
considered.
This is the Brüel & Kjær (B&K) LAN-XI
One of the difficulties with brake squeal is that module. It provides the required sampling
it occurs over such a wide frequency range. It speed, overlap processing, and other essential
is important to use an analyzer that can cover functions. Additionally it provides the ability
the range effectively. To cover up to 18,000 to do multiple analyses on the given channel.
Hz, the analyzer must be able to sample at This capability is often very convenient in brake
frequencies in excess of 36,000 Hz or more. noise measurement where one may want to

43
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 3.5 The Brüel & Kjaer Data Acquisition and Analysis System
(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

study averaged data, peak hold data, and time the squeal exceeds the noise levels before and
data from the same signal. The LAN-XI system after the event.
can do all of these processes at the same time
and in real time. The typical approach to do this is to collect
two samples of data. In one sample the sound
Signal averaging is something that has not pressure level spectra is averaged over some
yet been discussed. Sampling at high rates, period of time. Typical time periods will
greater than 50,000 Hz, means that a lot of range from a few tenths of a second to a few
data is collected quickly. When one is studying seconds. Figure 3.6 shows a comparison
brake squeal it is sometimes important to get between the instantaneous and averaged
both the average or typical sound during a spectra.
stop and the maximum or peak sound when a
squeal occurs. To evaluate how a squeal may The average spectrum is a collection of the
be perceived, it is useful to know by how much average magnitudes for each spectral line.

44
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH Measurements

With sufficient averaging time, this is a good It is clear that peak hold type averaging was
description of the sound level spectrum outside used to show the squeal tones so clearly above
the squeal event. the background noise.

To best describe the sound pressure magnitude, Another important aspect of NVH testing is
a peak hold average is used. In this case the being able to report data in a useful way. This is
resulting spectrum represents the maximum true for both laboratory and on-road data, but
level for each spectral line. As such, it provides the different types of reports and the need for
an accurate level for the maximum level of rapid turnaround make this a greater need for
the squeal tone or tones. With both of these vehicle testing.
spectra, one can see the magnitude of the
squeal in comparison to the normal spectrum. Figure 3.8 shows an example of a daily report
The ability to calculate both spectra in real time for a vehicle city traffic type test.
with the LAN-XI allows one to instantaneously
see the impact of a squeal event. While this report looks relatively
straightforward and the major graphs and
Having discussed the FFT and averaging, a tables have been discussed above the data
spectrum such as shown in Figure 3.7 should be analysis, generating this report is a significant
quite familiar to the reader. task. Once the raw data are obtained, sound

Figure 3.6 Demonstration of the effects of averaging.

45
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 3.7 Example of a brake squeal spectrum (Courtesy Link Engineering Co.).

spectra must be calculated for each stop an automated fashion special attention must be
using Fast Fourier Analysis. In addition, to paid to thresholds for peak detection, tolerance
determine what corner position the sound in the frequency match, and the background
came from, there is a complex analysis of noise that must be exceeded for the sound to be
several channels of data. The basic approach heard. With the background level varying with
for squeal analysis is to go through each corner speed and different sound treatments for each
position (accelerometer data in this case) and vehicle, this is a complex tuning process.
find significant pure tones or nearly pure tones
in the spectra. There are a variety of ways to Note that the end result seems quite simple,
do this, including running averages, slope but the computation is complex. In addition,
analysis, and more. Once significant peaks it is important to get this calculation correct
are found above 900 Hz and some threshold since product development or vehicle sourcing
above a computed noise floor, the individual decisions may be made based on this data.
corner spectra are compared to the interior
microphone spectrum to see if there are peaks Underlying this report are the frequency
at the same frequency. If so, then the corner spectra for each location and for each stop.
position source for the interior noise is known. Depending on the test sequence, this can
To be successful at this sort of comparison in represent thousands of spectra that must be

46
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH Measurements

Figure 3.8 Typical city traffic summary report (Courtesy Link Engineering Co.).

analyzed. Figure 3.9 shows an example brake Such an analysis can be very effective in
noise frequency spectrum. highlighting manufacturing issues that
manifest themselves in rotational orders.
There are many ways to analyze the noise
and vibration data to better understand the In this particular case, there are strong
characteristics of the noise problem. There is first through third order responses. These
still much work to be done developing more are typical results and represent the source
advanced analyses and reports to understand mechanism for judder.
the underlying causes of brake noise and
vibration. Some of the potential approaches and
promising new techniques will be discussed 3.4 Coherent Output Power (COP)
in the following segment. Figure 3.10 shows Perhaps the most important technique for the
an order plot of brake torque used to see the future of brake squeal analysis is the coherent
relationship to the rotational speed of the output power analysis. This technique can
brake. compute the relationship between measured

47
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 3.9 Typical example of frequency data for brake NVH measurements.

noise and the brake vibration. Coherent output (3.1)

power (COP) will be defined in detail to
provide a basis for discussing its use for brake where:
noise analysis.
= the complex conjugate
To begin this discussion, it is necessary to of the Fourier transform of the signal x(t)
define the signal analysis building blocks on
which coherence analysis is built. = the Fourier transform
of the signal y(t)
3.4.1 Cross Spectrum The best way of looking at the cross spectrum is
The cross spectrum is the function describing as the product of the two signals x(t) and y(t) in
the product of two signals x and y. It can be the frequency domain. This function also serves
written as: as a rough indicator of the relationship between

48
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH Measurements

Figure 3.10 Brake torque order plot.

x(t) and y(t). There are specialized applications where:

where the cross spectrum represents a physical
signal. One example is the measurement of = Coherence (0-100%)
sound intensity (not sound pressure) using two
microphones. In this case, the imaginary part
of the cross spectrum is the radiating intensity. = the magnitude of the cross
The real part is the reactive (non-radiating) spectrum between signal x and y, units of
energy. For the most part, the cross spectrum is amplitude squared
a building block for more complex analyses.
= the auto spectrum
of signal x to itself, units of amplitude squared
3.4.2 Coherence
The calculation of coherence is shown below: = the cross spectrum of signal x to
signal y, units of amplitude squared
The vertical scale is usually expressed as a
(3.2) percentage between 0 and 100. The meaning of

49
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

coherence is how much one signal is linearly to enter the system different definitions are used.
related to another. The simplest way of looking The most basic definitions are shown below:
at coherence is to consider a black box as shown
in Figure 3.11.

Note that the term n(t) represents noise in the (3.3)
system that is incorporated in the input. For
or
a linear system the output y(t) will be linearly
related to the input x(t) if no noise is present.
In this case, the coherence is 100%. However, (3.4)
this is seldom the case. Beyond the simplest where:
systems coherence above 90% is rare unless at a
resonance or where noise is simply absent.
H(f) = Frequency response function (FRF)
dimensionless
It is important to point out that coherence
is the correlation function in the frequency When expressed in the form of a magnitude
domain. It is an excellent indicator of the and a phase, this function describes how the
amount of noise in a system and of how one input is related to the output of a system similar
signal is related to another. The magnitude to that shown in Figure 3.11. The magnitude
of coherence is used to indicate how well one can be seen as a multiplier at particular
signal is related to another. Identical signals or frequency to bring x(t) to the same amplitude
linearly related signals will have a coherence of as y(t). The phase angle is the phase difference
100%. Coherence magnitude is often plotted between x(t) and y(t).
with a frequency response function as an
indication of the validity of the function.
and (3.5)
3.4.3 Frequency Response Function where:
The frequency response function (FRF) has
several definitions. Depending on the nature of = magnitude of the
the signals involved and where noise is thought frequency response function (FRF)

n(t)

x(t) System y(t)

Figure 3.11 Basic system.

50
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH Measurements

the phase angle between road, wind, and powertrain inputs. Coherent
the input and output signals output power can eliminate much of this noise.

3.4.4 Coherent Output Power The value of coherent output power is that

it virtually eliminates all extraneous noise
To determine the amount of the output that is
sources. For on-road testing it is far better
coherent with the input, the coherent output
than trying to align spectral peaks from
function is used. It is defined as:
corner acceleration measurements and noise
measurements in the interior. The potential of
coherent output power is that it can say that X%
(3.6) of the squeal noise or interior noise comes from
The value of coherent output power is that it a particular brake.
eliminates the noise that often obscures the
important signal content. Where input may be An example coherence calculation for an actual
from multiple sources, coherence can also be brake test is shown in Figure 3.12.
used to estimate the contribution from each
source. This sort of estimate has to be used with In this particular case, it is clear that there
caution since it is possible that there may be are three strong spikes in coherence at 5500
coherence between the inputs. Hz, 11,000 Hz, and 16,500 Hz; the highest
coherence being 80% and the lowest peak
In the case of vehicle interior noise, coherent having a coherence value of 30%.
output power may be used to distinguish what
part of the interior noise is due to a particular The product of the coherence and the
brake. Also, since there are other noise sources autospectrum of microphone signal are shown
that reach the interior, coherent output power in Figure 3.13.
can be used to eliminate much of the non-brake
related noise measured by the microphone. In The strong peaks are well defined and represent
typical vehicle measurements there is a great the squeal components in the microphone
deal of low- to mid-frequency noise due to the signal.

Figure 3.12 Coherence for example case

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

51
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 3.13 Coherent power for example case

(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

A comparison between the measured sound signal to noise ratio between the peaks and the
pressure levels and the coherent output powers backgrounds is increased from 15 dB to 45 dB
is shown in Figure 3.14. for the first peak.

The red line is the normal sound pressure The value of COP is that it reduces the
measurement inside the vehicle. It is clear that interference from extraneous noise sources and
this is a squeal peak at approximately 5700 provides a clear picture of the cause and effect
Hz and a just visible peak at 11,500 Hz. The relationship for brake noise. It can be helpful
blue line represents the COP result. The first for both dynamometer and vehicle on-road
two peaks are evident. However, there is also measurements. In the dynamometer it can
a third peak at approximately 17,000 Hz. It is suppress the effects of tire noise, HVAC noise,
easy to see why many people regard COP as a and other contaminants. On the road, it will
means to increase the signal to noise ratio in a reduce the effects of noise from other vehicles
measurement. The “noise” between the peaks and sources and allow one to accurately
is greatly reduced in amplitude by COP. The

Figure 3.14 Comparison between sound pressure level spectrum and coherent output power for
example case (Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

52
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH Measurements

attribute a noise to a specific corner of the data collected. The use of high-frequency,
vehicle. charge-type accelerometers is critical to getting
accurate data for brake vibration.
As brake design improves and brake noise
becomes more difficult to find in the The use of COP has been shown to be an
background noise in the vehicle interior or in excellent means to highlight brake squeal.
a brake NVH dynamometer, COP will be a This technique provides a valid mathematical
necessary part of all brake noise measurements. foundation for isolating brake squeal and
verifying the presence of squeal. For vehicle
testing in particular, this analytical technique
3.5 Summary provides major advantages over the more
From the transducers to the methods of data traditional approaches of peak picking and
analysis, care is needed to properly represent setting amplitude thresholds. It is strongly
brake noise. The high-frequency nature of recommended that this analytical approach
brake noise makes it particularly important be used when processing brake noise and
that one carefully select the accelerometers to vibration data.
be used for squeal measurements. The high
temperatures generated by the braking system Now that the basic tools and techniques for
make the choice of the proper accelerometer collecting and analyzing brake NVH data have
even more difficult. In many instances, brake been defined, the next step is to move to the
vibration data is currently being collected laboratory to discuss measurement techniques
with inadequate transducers and users are and standards.
unaware of the resulting inaccuracies in the

53
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 4
Laboratory Testing

55
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

4.1 Introduction of the lower frequency noises, this may even

The goal of this chapter is to provide a complete extend to whole axles and vehicle structures.
discussion of the equipment and methods As has already been discussed, the resonances
used in laboratory dynamometer testing. The of suspension components, wheels, and even
discussion will begin with the dynamometer structural members are important to recreating
used for such testing. Their major components some noises. In addition, the transmission
and performance requirements will be paths through suspension and vehicle structure
described in some detail. An explanation will may be necessary to reproduce a noise issue.
be provided for the configuration of a brake The key for accurate noise testing is matching
noise dynamometer. the test system to the types of noise to be
reproduced. This will be a topic to be discussed
Next, the existing test standards and throughout this chapter.
measurement procedures will be described. The
only internationally accepted brake noise test It is also necessary to match the acoustic
standard will be reviewed in detail. characteristics of the vehicle in operation.
This does not mean that the whole vehicle
Finally, current dynamometers and their body structure must be used. However, the
capabilities will be discussed. The goal will be overall acoustic conditions must be matched
to provide insight as to which dynamometer is as well as possible. As was noted previously,
best for specific noise tests or issues. the vehicle operates in a hemianechoic
environment. The road serves as a reflecting
plane and to a good approximation the rest
4.2 Requirements for Brake Noise of the environment is free of reflections or
Testing in the Laboratory is anechoic. Achieving a true hemianechoic
The most fundamental requirement of a environment in a laboratory test can be done,
laboratory brake noise test is that the vehicle but at significant cost with a need for a large
operating conditions must be recreated. The test space. This extreme is often not necessary
speed, brake temperature, vehicle inertia, for a good brake noise measurement. Another
and corner dynamics are some of the more important need in recreating the acoustic
obvious requirements. Clearly the speed, environment is the elimination of other noises
inertia, and temperature requirements are that are not from the brake or typical of the
necessary to ensure that the brake is absorbing vehicle test. For instance, if an electric motor
the same amount of energy as the brake on is used to drive the brake it is imperative that
the actual vehicle. If this is not the case, it may the motor noise be suppressed to the point
be impossible to generate the same noise. The that it does not contaminate the brake noise
brake temperature is important since it affects measurement. Achieving the desired acoustic
friction characteristics and the resonances of environment is a significant challenge that
many of the components such as the brake pad. is once again related to the frequency of the
noise of interest.
The only item that may be unfamiliar to the
reader is need to match the corner dynamics. More recently the need to control temperature
For any sound measurements, other than the and humidity in the test environment has
highest frequency squeals, the dynamics of the received a great deal of attention. As improved
entire brake corner are important. For some testing and understanding of brake noise

56
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

generation has led to the reduction of some Until recently, the only way to test for this
of the more common brake noise issues, less sort of noise was in cold chambers with whole
common NVH issues related to extremes of vehicle or geographic locations where cold and
temperature and humidity have become more humid conditions could be found consistently.
important. A good example is the phenomena These limitations were a major hindrance
called “morning sickness.” The classic case to effective testing. Now, many brake noise
is the vehicle parked outside overnight in test facilities provide the ability to control
moderate climates around the world. During temperature and humidity to replicate these
certain seasons the temperature will drop to and other extreme conditions. Newer test
near freezing overnight. At the same time procedures require these sorts of controls and
the relative humidity may rise to high levels therefore, they have become necessary parts of
of 80% or more. The next morning when the a brake NVH test system.
driver reverses from his driveway or pulls up
to the first stop sign there is a loud squeal.
This may happen only for the first stop or the 4.3 Brake NVH Dynamometers
first few stops until there is enough energy A brake NVH dynamometer is typically what
in the brakes to heat them up and drive is used to do brake noise testing. Figure 4.1 and
out moisture. This is a common warranty 4.2 show an example of such a dynamometer.
complaint. The owner may not experience
any other noisy stops, but this happening Figure 4.1 provides a diagram of the layout and
nearly every morning can be a major source of Figure 4.2 shows a photograph of the major
concern. sections. To understand the components and

Figure 4.1 Brake NVH dynamometer drawing (Courtesy Link Engineering Co.).

57
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 4.2 Brake NVH dynamometer photo layout (Courtesy Link Engineering Co.).

their role in brake NVH testing, each will be at moderate to high torques. The brakes will
described briefly here and in more detail in the be applied and the pressure decreased just
following sections. enough to allow slipping to begin in such a
test. The main electric motor has difficulty with
The driving force for the entire system is the maintaining such low speeds with varying
electric motor second from the left hand side of loads. So, for such tests the low-speed drive is
Figure 4.1 and the far right hand side of Figure engaged and it takes control of the operation.
4.2. The motor is used to spin the brake up to
the necessary speed before the brake is applied. Moving to the right of the main motor is
For drag tests, which are common in noise the inertia section in Figure 4.1 and in the
testing, it holds speed constant as the brake is center of Figure 4.2. In the operation of the
applied or dragged. In addition, the electric dynamometer these inertia discs are used
motor can be used as part of the control system to simulate the linear inertia of the vehicle.
to simulate the inertia of the vehicle. In most modern dynamometers using a
combination of simulation with the motor and
To the far left of Figure 4.1 and not shown in these discs allows one to accurately simulate
Figure 4.2, is the low-speed and high-torque the vehicle. The required number of discs
drive. This is used in place of the main electric has been reduced from 10 or more to three as
motor when tests are to be done for phenomena shown here or even as low as one by the use of
such as moan. The requirement for such cases electric motor inertia simulation.
is to run at low speeds, maybe fractional rpm,

58
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

It is important to note that in this entire The suspension would be supported from the
section the low-speed drive, the main motor, tailstock and the brake would be driven by the
and the inertia discs are inside an enclosure. dynamometer shaft. The suspension attachment
This enclosure serves two purposes: it is a can be seen in Figure 4.2. The test section is
safety measure to protect personnel from the enclosed in a separate chamber to provide sound
rotating components and it serves to contain isolation and to provide the necessary acoustic
much of the noise from these components. As environment. The acoustic treatment of the
was discussed above, it is important that any walls can be seen in Figure 4.2. The important
extraneous noise not seen in normal vehicle characteristics of the acoustic enclosure will be
operation is suppressed. This enclosure is part discussed in greater detail below.
of that suppression system.
The shaft type dynamometer shown in Figure 4.1
Moving further to the right in Figure 4.1, and and 4.2 is the most common for laboratory brake
on the far left in Figure 4.2, the test section noise testing. It is the most compact and efficient
is seen. This is where the brake is mounted. way of performing such tests in the majority of
Note that the typical dynamometer tests only these cases. However, another approach is to
one brake at a time. The brake is attached to place the vehicle under study in a hemianechoic
the dynamometer shaft and anchored on the chamber with a dynamometer for the vehicle
tailstock. For many tests it is necessary to tires to run on—such a chassis dynamometer is
use the complete vehicle suspension corner. shown in Figure 4.3 and Figure 4.4.

Figure 4.3 Chassis dynamometer drawing (Courtesy Link Engineering Co.).

59
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Clearly such a test facility,

if properly designed, can
provide a more realistic
test environment in a lot
of ways. The required
space and cost for such
a facility are important
considerations. An
additional consideration
is the fact that a full
vehicle is needed for
testing. In the early stages
of brake development a
representative vehicle may
not be available.

4.4 Electric motor

As noted above, the
heart of the brake NVH
dynamometer is the
electric motor. This is true
for the both the shaft and
chassis dynamometers.
The electric motor has
to accelerate the shaft,
inertia discs, and braking
components to the desired Figure 4.4 Chassis dynamometer designed for brake NVH testing.
speeds before the actual
braking can begin. While
this acceleration does not
have to be instantaneous, Both direct current (DC) and alternating
it does represent wasted test time. In the current (AC) motors are used for NVH
course of a typical brake noise test with 2000 dynamometers. Considering the motor and
or more brake applies, even a few seconds for the other components required to drive it,
acceleration makes a significant difference in the DC motor is more economical in the
total test time. sizes and speeds commonly used for NVH
dynamometers. On the other hand, the AC
The motor also has to maintain constant speed motor is a little more versatile and has better
during the drag tests that are part of most control characteristics at low speeds. The AC
NVH test sequences. At low speeds with high motor can introduce more electrical noise in
apply pressures, this can be a challenging task the measurement system if the components are
for an electric motor with insufficient torque not properly shielded. The decision between an
capacity, slow response, or poor controls. AC and DC motor is often based on the cost.

60
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

However, there may be specific applications that water-cooled motors are larger and more
where one type of motor may be better suited. expensive.

Typical motor sizes for shaft NVH

dynamometers are roughly 185 kW (250 hp). 4.5 Mechanical Inertia and Inertia
The motors for chassis dynamometers may be Simulation
larger or smaller depending on whether one or It was noted previously that, in a dynamometer,
two wheels are being driven by each motor. The rotational inertia is used to simulate the
maximum speeds for a motor used on a NVH inertia of the vehicle. Without the proper
shaft dynamometer do not need to be higher inertia, the brake would not provide the right
than 2000 rpm since the maximum speeds amount of energy dissipation and the noise
required are roughly 80 km/h (50 mph). The performance may be different since noise
result is a fairly basic motor specification in a is highly dependent on speed, temperature,
common size. There are many manufacturers deceleration, and clamping force. A decade ago
who can provide motors that meet these most dynamometers had to rely on mechanical
requirements. inertia. Therefore, they employed large
numbers of discs, as many as ten or more, to
A major consideration in the application of the provide the necessary increments of inertia for
motor is preventing the noise and vibration simulating different vehicles. It was necessary
from the motor contaminating the brake noise to select the right combination of discs to
measurement. Depending on the design and replicate the vehicle’s inertia. At the same time,
construction of the motor, the noise radiated the applied inertia could only be as accurate as
from the motor casing can be substantial. In the smallest disc on the dynamometer. When
some designs, the structure can serve as a going from one vehicle test to another, setting
good radiator for the vibration created by the up a dynamometer was a time-consuming
magnetic flux in the motor. However, the major process and at best a good approximation.
problem is the cooling air system for the motor. There are still many legacy dynamometers in
Generally, this is the noisiest component. These use today employing this approach.
noise concerns are typically addressed by
placing the motor in an insulated enclosure as With computerization of the controls of the
shown in Figure 4.1 and 4.2 or by locating the dynamometer came the opportunity to control
motor in a separate room from the test section. the motor in such a way as to simulate inertia.
Often it is necessary to place inlet and outlet With high-speed controls, the motor can add
silencers on the motor cooling air ducts to energy or absorb energy to match the effects of
further reduce the noise. vehicle inertia. The amount of inertia that can
be added or subtracted from the system by a
In the case of very large motors or in given motor is shown in the equation below [7].
installations where there is great concern about
the noise from the drive system, water-cooled
motors may be used. The water-cooling casing Ia = (Pm*RR*9549.3)/(RPM*a*g) (4.1)
on the outside of the motor greatly reduces the
noise radiated from the motor. In addition, where:
the cooling air system and its related noise
are completely eliminated. The drawback is Ia = available inertia, kg-m2

61
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Pm = maximum motor power, typically 150% calibrated, inertia simulation is much more
or rated power, kW accurate than using discs alone in matching
the vehicle inertia. When mechanical inertia
RPM = rotational speed, rev/m is used, even if the exactly right amount of
inertia can be applied, there are errors due to
g = acceleration of gravity, 9.81 m/s2
bearing losses and windage. Since these losses
a = vehicle deceleration, m/s2 change with speed, and in some cases load,
only inertial simulation can correct for them
RR = tire loaded rolling radius, m properly throughout the speed range.

9549.3 = a constant to correct for units Inertia simulation provides a number of

benefits. It extends the range of inertia that can
For a 185 kW (248 hp) motor running at 1000 be simulated on a dynamometer. Depending on
rev/m with a tire rolling radius of 0.28 m (0.92 the inertia present and the power of the motor,
ft) and deceleration of 0.24 m/s2 (0.79 ft/s2), the this extension can range from simply adjusting
available inertia is 21.5 kg-m-s2 (155.5 slug-ft2) between the minimum steps in the physical
or 212.6 kg-m2. This is a large amount of inertia inertia system to extending the minimum and
that would be equal to that for a full vehicle at maximum inertia that can be provided. Figure
moderate speeds. If one uses 150% of the rated 4.5 shows an example of the extension in the
motor power, this value climbs to 32.2 kg-m-s2 range of inertia that can be achieved.
(232.9 slug-ft2) or 315.3 kg-m2. The value of
150% of the rated motor power is typically used The green shaded area represents the
in this calculation since most motors are able capabilities with inertia simulation. In this case,
to generate 150% or more of their rated power the calculation is done for a rolling radius of
for several seconds. For inertia simulation such 0.35 m (1.15 ft), 1099 rev/m, and a 186 kW (248
short bursts are all that is needed and thus this hp) motor. The maximum deceleration rate
extra capability can be used with no harm to considered is one g or 9.86 m/s2 (32.35 ft/s2).
the motor.
In this particular case the mechanical inertia
The key to effectively simulating inertia is to in the system was approximately 150 kg-m2 to
have an accurate measurement of speed and the provide perspective in how far the range has
change in speed and to be able to then calculate been extended. At small decelerations the effect
and effect motor current changes quickly. of inertia simulation is quite large. Also, inertia
A typical sampling and update speed is 100 simulation allows testing to very low inertias to
times each second. At such speeds, the changes cover an even wider range of vehicles.
needed for inertia corrections are small, and
highly accurate control can be maintained. One word of caution, while it may be possible
to design a dynamometer with minimal to
One of the important benefits of the inertia no physical inertia, this is not wise for an
simulation system is that it negates any effects NVH shaft dynamometer. Due to the nature
of bearing friction or windage in the rotating of electric motors and the low speeds of 3
discs when it is applied. Since these effects km/h (1.9 mph) required in noise tests, a small
are occurring as the inertia calculations are amount of physical inertia in the system is
made, they are automatically corrected in the desirable. At low speeds motors have a tendency
process. Tests have shown that when properly

62
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

Figure 4.5 Example of the performance capabilities of a dynamometer using inertia simulation.
(Courtesy Link Engineering Co.)

to produce small regular speed surges. Having outside of the enclosure, the noise or vibration
a small amount of inertia in the system tends to does not ruin the ongoing noise test in the
smooth these surges and improves the ability to chamber. The enclosure also prevents high
replicate vehicle operation. noise or vibration levels in the chamber from
causing personnel discomfort or disturbing
other nearby noise tests. It attenuates the brake
4.6 Enclosures noise such that what escapes through the
The next part the dynamometer to be chamber walls is insignificant. The enclosure
considered is the test chamber or enclosure. also maintains the environmental conditions
The enclosure serves a number of functions. in the test space, and through insulation and
It isolates the test environment from exterior sealing it reduces thermal losses and prevents
noise and vibration. If someone drops a part leaks. Finally, the enclosure has to provide the

63
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

desired acoustic conditions. As was discussed of temperature and humidity and prevents
before, these conditions are highly absorptive sweating on the outside of the enclosure when
walls and ceiling and a reflective floor to match low-temperature tests are being run in the
vehicle operation on a roadway in an open summer months.
environment. The enclosure is one of the most
important and most mundane looking parts of Another very important part of the enclosure
the NVH dynamometer. is the seals. The ideal test chamber would
have no doors or windows. However, this
A typical enclosure is shown in Figure 4.6. would make it very difficult to move parts in
and out for testing. A small leak past a seal
In this case and in many such enclosures, can significantly degrade the performance
several layers of materials are used to provide of the enclosure both acoustically and
the required performance. There is an outer environmentally. For an enclosure designed
wall of heavy-gauge steel, a layer of thermal for 50 dB of transmission loss, a 0.1% area leak
insulation, another wall of stainless steel, and will degrade the performance to 30 dB. If a
an inner lining of sound absorption material. typical brake enclosure measures 3 m (9.8 ft) on
These layers work as a system to absorb and a side and is 2.5-m (8.2-ft) high, this amounts
attenuate noise and serve as a barrier for to an open area of 0.039 m2 (0.128 ft2) or 7 mm
sound traveling into or out of the enclosure. (0.275 in) opening around a regular door. The
The thermal insulation helps in the control traditional rule of thumb with such enclosures

Figure 4.6 Typical brake NVH dynamometer enclosure (Courtesy Link Engineering Co.).

64
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

is that they need to be water tight to work well small leak can completely upset the stability of
acoustically. the environmental control system.

As was noted, the effects of leak on the Noting that every door, window, and cable
environmental conditioning system are also pass through represents a leakage path is bad
critical. Consider a test where the interior of enough. In addition to this, the dynamometer’s
the chamber is being maintained at 10oC (50oF) rotating shaft, 10 cm (3.9 in) or more in
and a relative humidity of 20%. If the outside diameter, also has to pass through the
air is at 25oC (77oF) and 50% relative humidity, enclosure wall. Sealing the noise enclosure is a
the amount of moisture in the outside air is major design and maintenance concern.
nine times that in the enclosure. A small leak of
outside air is not only 15oC (59oF) warmer, but The performance of a well-designed brake test
it also contains nine times the moisture of the enclosure is shown in Figure 4.7.
air in the enclosure. It is not hard to see how a

Figure 4.7 Example of test enclosure background noise levels.

65
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Two test conditions are represented here. The fan blade passage frequency near 2000 Hz can
blue line is the sound pressure level inside the be easily identified. Installing silencers would
enclosure with the dynamometer all-related significantly reduce the noise in this problem
equipment shut down. This is the minimum frequency range.
ambient noise in the chamber. The red line
represents a measurement with all equipment
running. The dynamometer is running at 80 4.7 NVH Dynamometer Special
km/h (50 mph), and the cooling air system is Considerations
running at maximum speed. The preceding discussion has highlighted
many of the components that make up a
Since even the best of enclosures are less dynamometer. However, there a few other
effective at low frequencies, it is not surprising considerations that need to be discussed. One
to see higher noise levels at lowest frequencies. of the unique characteristics of brake NVH test
This particular enclosure was optimized for matrices is the use of low-speed drags. These
brake squeal testing. Therefore, the major are extended periods, often 10 seconds, where
performance concern was for frequencies above the brake rotational speed is held constant with
900 Hz. no or small variations in the apply pressure.
Such drags have been used in noise tests for
The international standard for brake NVH some time and are believed to be effective in
testing specifies a background level of 60 eliciting brake squeal. These drags can occur at
dB(A) from 900 to 19,000. Going back to the speeds as low as 3 km/h (1.9 mph).
early discussion of decibel addition, the use
of 60 dB(A) is predicated on the concept that To implement such drags on a dynamometer,
the background level should be 10 dB lower good motor speed control is required. In many
than the lowest level to be measured. Thus, of the more modern tests, the apply pressure
a 60 dB(A) background means that effective is varied as much as 5 bar (73 lb/in2) during
measurements can be made to 70 dB(A) the course of the drag to try to induce drag.
without the need for background corrections. The dynamometer pressure controls and
Looking at the red line, this chamber appears motors controls must be capable of making
to be in compliance with this requirement these adjustments and holding speed constant
throughout the spectrum. However, many throughout the drag. This test requirement
people wish to make brake noise measurements is one of the drivers for larger motors with
at even lower levels and they have chosen to use sufficient torque capacity to provide minimal
a background level of 50 dB(A). In this case, the speed variations during these pressure
red line exceeds 50 dB(A) around 2000 Hz. The profiles.
weakness is not the enclosure itself. In order
to meet the low noise level requirements, it is Another special consideration is the need to
necessary to fit silencers on the environmental do creep-groan type tests. This is perhaps
system inlet and outlet ducts before they reach the most demanding requirement for a noise
the enclosure. This measure ensures that noise dynamometer. To perform such a test, a
levels from the fans in the cooling air system to series of requirements must be met. First the
do not create a noise problem. In this enclosure dynamometer must be able to do a breakaway
installation, the owner decided that he would type test where the brake is held stationary and
not use silencers to save money. Therefore, the with a torque held by the clamped rotor. Then,

66
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

through adjustment of the line pressure and 4.8 Brake Noise Measurement
motor torque, the brake must be released and Standards
allowed to enter a stick-slip motion and speed There is only one internationally recognized
held to be very slow. Many dynamometers and laboratory brake noise measurement
motor controllers find this to be a very difficult standard. This is SAE J2521 Disc and Drum
control regime. Since the electric motor Brake Dynamometer Squeal Noise Matrix.
produced maximum torque at zero speed, it This standard was originally published in
tends to speed up very quickly when the brake 2001 and was updated in 2005 [8]. This
slips. This is not at all like what happens with test procedure is based on the use of a
a vehicle on a hill. It has been suggested that dynamometer as described previously for
such an operation would be better done with a noise testing. It can be applied to both a shaft-
hydraulic motor, but such a dynamometer has and a chassis-type dynamometer. The SAE
not been built. J2521 standard has been credited as one of the
reasons for the decline of brake squeal issues
As an alternative approach, many in the last decade. This standard has provided
dynamometers provide a slow-speed, high- a common “yardstick” that has fostered
torque drive that can be used to run at very consistent and meaningful evaluations that
slow rotational speeds with pressure applied have been critical to avoiding brake squeal
to the caliper. It is possible, in many cases, problems.
to generate a stick-slip motion that is a good
approximation for creep groan with such a It is very important to be aware that this
system. However, it is not exactly the same as standard is for a brake noise screening
the actual creep groan situation that has been a procedure. It does not attempt to simulate
limitation for some investigators. typical vehicle operation or attempt to
reproduce noise common to a particular type
However, there is an advantage to the low-speed of operation. Instead, it is designed to elicit
drive. Since it can operate for long periods of every possible squeal noise in routine operating
time at fractional rpm and hold speed quite conditions. As such, it may generate squeals
constant, it allows one to establish the groan that the average customer may never hear. The
sound and hold it for an extended period of intent is to provide the worst-case noise as a
time. This allows for much better measurement screening tool.
and understanding of the problems than can be
obtained with the intermittent occurrence on As a first step, each of the test sequence steps
an actual vehicle. will be described. Table 4.1 provides a detailed
listing of each of the major segments of the test
It is important to note that, since creep groan procedure.
is a low-frequency sound, at least the full
vehicle suspension corner must be present The first segment of the test procedure is the
for this test to provide representative noise. Initial Conditioning section. The goal of this
Some investigators have found it necessary to segment is to produce a defined transfer layer
include part of the vehicle structure to match using a series of 80 to 30 km/h (50 to 19 mph)
the response seen for the actual vehicle. At the snubs. Only 62 snubs are executed to condition
very minimum, the full suspension corner is the brake the minimum necessary. One of the
necessary. important concepts of SAE J2521 is to examine

67
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Table 4.1
SAE J2521 Test Matrix summary
Test Segment Section Description Number of Speed, Pressure, Initial Notes
No. Applications km/h bar Temp., o C

Initial 1 Break-in 30 80-30 30 100

Conditioning

2 Bedding 32 80-30 15-51 100

3 Friction 6 80-30 30 100

Characteristic
Value after
Break-In

Evaluation 4, 9, & Drag Module 266 3 & 10 0-30 50-200

14

Repeat 4-7, 9-12 5, 10, & Intermediate 24 50-0 5-30 100 & 150
& 14-17 3 times 15 Conditioning &
Warm-Up

6, 11, & Backward & 50 -3 & 3 0-20 150-50

16 Forward

7, 12, & Deceleration 108 50-0 5-30 50-250

17 Module

18 Friction 6 80-30 30 100

Characterization

Optional Cold 19,22 Cold Matrix A 77 5 0-12 0-50-0

Section

20, 23 Cold Matrix B 55 30 0-20 0-50-0

21, 24 Cold Backward/ 70 -5, 5 0-12 0-50

Forward

29 Backward & 50 -3 & 3 0-20 150-50

Forward

30 Deceleration 108 50-0 5-30 50-250

Module

31 Friction 6 80-30 30 100

Characterization

all aspects of brake squeal. The developers delivery of the vehicle. In many instances, this
of the procedure were cognizant that “green noise subsides after brake conditioning, but it
brake” noise was a concern for many vehicle may lead to recalls and warranty complaints.
manufacturers. Green brake noise refers to Therefore, the conditioning step was kept short
noise seen in the first few brake applications at to allow the possibility of measuring this sort of
the factory or by the customer when first taking noise in SAE J2521.

68
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

It is important to note that noise is measured in of the test matrix. It is repeated for a total of
every segment of the J2521 matrix. Even if the three runs to be sure that every noise event
section is not specifically designed to induce is properly captured. It is not unusual for the
noise, it is imperative the noise be measured. noise performance of the brake to change over
the course of the three runs of this segment
The first Friction Characterization Value test since in some cases the brake is being broken in
occurs next. The original goal of this section during the course of the test.
was to measure the friction characteristics of
the brake to determine if they could be related The first section of this module is the
to changes in noise performance. Although no evaluation when low-speed drag tests are run.
such relationship has been developed, these There are 266 brake applications with drags
segments do sometimes generate noise and are at 3 and 10 km/h (1.8 to 6 mph). The drags are
therefore, an important part of the test. Six 10 seconds in length and include initial brake
snubs are made from 80 to 30 km/h (50 to 19 temperatures from 50 to 200oC (122 to 392oF).
mph) with 30 bar (435 lb/in2) pressures and The nominal pressures are between 0 and 30
initial brake temperatures of 100oC (212oF). bar (0 and 435 lb/in2). In this section, pressure
These are the same parameters used for all profiles are applied to be sure that a large
friction characterization segments. range of pressures are experienced and that
any tendency to generate noise at a particular
Next, the test matrix moves to the noise pressure is captured. The specified pressure
evaluation module. This is a core segment profile is shown in Figure 4.8.

Figure 4.8 SAE J2521 Pressure Apply Ramp Specification.

69
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

For all cases except the 0 bar (0 lb/in2) case, the entire Evaluation section is then repeated two
pressure ranges from 5 bar (73 lb/in2) below the more times to complete this portion of the test
nominal pressure to 5 bar (73 lb/in2) above the matrix.
nominal pressure and back down. The profile is
timed to cover the entire duration of the drag. All of the previously discussed segments are
For the case when the nominal pressure is 0 bar mandatory portions of the test. They must be
(0 lb/in2), the minimum is 0 bar (0 lb/in2) and run every time SAE J2521 is executed. However,
the maximum is 2.5 bar (36 lb/in2). there are some optional sections that can be
run at the discretion of the user. The first of
The next step in the procedure is the these is the Cold Temperature segment. This
Intermediate Conditioning and Warm-Up. This section provides three sets of tests in an attempt
segment is appropriately labeled. It is simply a to induce noise at low temperatures. The Cold
step to take the brake through an intermediate Matrix A segment incorporates brake drags
temperature range using full stops. The speed at 5 km/h (3.1 mph) with pressures ranging
range is from 50 to 0 km/h (31 to 0 mph) with from 0 to 12 bar (0 to 174 lb/in2). The initial
pressures from 5 to 30 bar (73 to 435 lb/in2). brake temperature is 0oC (32oF), cycles to 50oC
The initial brake temperatures are 100 and (122oF), and returns to 0oC (32oF). In each case
150oC (212 and 302oF). where the temperature begins at 0oC (32oF),
the segment begins with a 45-minute soak of
Following this is the backward and forward the brake parts to be sure they are uniformly
section. Some brakes have a tendency to make at 0oC (32oF). For this purpose, -10oC (14oF)
more noise when rotating in one direction than cooling air is directed on the parts. This process
the other. The intent of this section is to make is repeated for each section starting at this
sure this is accounted for in the procedure. temperature.
Alternating forward and reverse drags are
made at 3 km/h (1.9 mph) with pressures The next cold sequence is Cold Matrix B. In this
ranging from 0 to 20 bar (0 to 290 lb/in2) and sequence a higher speed drag is performed at
temperatures from 150 to 50oC (302 to 122 oF). 30 km/h (19 mph). The goal for this sequence is
Note that the temperature will be coming down to replicate the noise issues seen by one vehicle
from the previous section. These low-speed manufacturer on highway off ramps. The
drags do not put a lot of energy into the brakes, pressure range for this section is 0 to 20 bar (0
so the temperature will be decreasing. to 290 lb/in2) with an initial brake temperature
of 0oC (32oF), going to 50oC (122oF), and back to
The next section in the Evaluation segment is 0oC (32oF).
the Deceleration Module. In this case, braking
is applied to go to a full stop from 50 km/h Finally, there is the Cold Backward/Forward
(31 mph). The applied pressures range from 5 segment. This segment is meant to replicate
to 30 bar (73 to 435 lb/in2) with initial brake the morning sickness problem discussed
temperatures of 50 to 250oC (122 to 482oF). previously. Speeds of -5 and 5 km/h (-3.1 and
3.1 mph) are used with pressures ranging from
The final item of the Evaluation section is 0 to 12 bar (0 to 174 lb/in2). The initial brake
the Friction Characterization test, which is temperatures are stepped between 0 and 50oC
identical to the segment described above. The (32 and 122oF).

70
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

This completes the Optional Cold Section. It segment range from 100 to 500oC (212 to
is important to note that this complete section 932oF). In some instances, testers have taken
must be executed if it is run. In reporting the exception to the maximum temperature range
test results, the testers are required to note that based on the capabilities of the brakes being
this section was run and report the results in tested. Any such exceptions need to be noted
the standard data formats provided. in the test results.

The other optional section is the Fade and The Recovery segment consists of a series of
Recovery Section. Th is test segment is well snubs from 80 to 30 km/h (50 to 19 mph) with
described by its name. The goal is to see if the 30 bar (435 lb/in2) of apply pressure and an
brake fade and recovery—taking the brakes initial brake temperature of 100oC (212 oF). This
to very high temperature and letting them is followed by the same sequence of segments as
recover—has an effect on noise performance. in the Evaluation section. The only difference
Experience has shown that all possible being that they are run only once.
outcomes can occur. The brake squeal may
get worse, not change, or lessen after the fade. Additionally, SAE J2521 specifies the location
The fade test is a series of stops from 100 to 0 of the measurement microphone as shown in
km/h (62 to 0 mph) using a 0.4 g deceleration Figure 4.9.
criteria. The initial temperatures for this

Figure 4.9 SAE J2521 Microphone Placement Specification.

71
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

This location was chosen to approximate the dB(A) and corresponding frequency at fixed
location at the edge of a vehicle wheel well intervals. Accelerometers may be used for the
where microphones were traditionally located validation of noise events.
in vehicle evaluations. For higher frequencies
where the sound wavelength is much smaller An important part of this specification is the
than the dimensions of the rotor, a directional fixture used for mounting the brake system.
radiation pattern may be anticipated and a It is noted that brake applications with a peak
single microphone may not be sufficient to sound level ≥ 70 dB(A) are determined to
fully characterize the radiated noise. However, exceed the noise data recording peak threshold.
the need for a standard microphone location The frequency ranges for this threshold are:
outweighed such considerations. Experience
has indicated that this choice has worked • 0.9 to 17 kHz for disc brake tests utilizing a
well with comparable results from different full suspension corner assembly.
dynamometers. • 2 to 17 kHz for disc brake tests NOT
utilizing a full suspension corner assembly.
The measurement procedure also specifies
• 0.5 to 17 kHz for drum brake tests.
many aspects of how the test should be run
and the necessary measurement requirements.
Note that the minimum 2 kHz frequency limit
These are too lengthy and numerous to go
is used with the non-suspension fixture because
through in this text. However, a few highlights
structural transmission may occur below this
will be noted. The maximum permissible
frequency. Experience has shown that vehicle
background sound pressure level of 60 dB(A)
performance below 2000 Hz can only be
at frequencies above 0.9 kHz is specified
simulated with a full suspension fixture.
for within the test chamber. It is noted that
a frequency analyzer and/or digital data
The recommended fixture is with a complete
acquisition system shall be used to acquire
suspension corner to enable evaluation over the
the sound measurements. The system must
complete 0.9 to 17 kHz range. Four different
compute and record narrowband sound
levels of fixtures are described in Table 4.2.
pressure spectra using the following parameters
and capabilities:
The minimum is rigidly connecting the brake
parts to the dynamometer shaft. The most
• 500 to 20 kHz frequency bandwidth with a
complete is the soft mounting where the full
25-Hz frequency resolution.
suspension, with the inclusion of connecting
• Peak-hold averaging with 50% overlap bushings, is used.
processing and Hanning windowing
applied to the time records. Some examples of these types of suspension
• Anti-aliasing filtering. fixtures are shown in Figure 4.10.
• A-weighting.
It is important to note that it is necessary
Noise measurement duration is the same as to load the suspension to actual operational
the duration of the brake application for all geometry. Because of nonlinearities in the
sections. Microphone information may be suspension, the proper loading and geometry
saved as peak-hold spectra, time recordings, are required to replicate the dynamic
peak-hold history at fixed intervals, or peak characteristics of the system at the frequencies

72
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

Table 4.2
classification of brake NVH test fixtures
(Courtesy Link Engineering Co.).

step 0 1 2 3
Type Rigid Knuckle Suspension Soft

Brake rotor X X X X

Brake lining X X X X

Brake caliper X X X X

knuckle X X X

Hub and Bearings assembly X X X

drive shaft (driven axle) X X

control and upper arms X X

spring and shock absorber X X

stabilizer Bar X X

sub frame X

Figure 4.10 Examples of different types of test fixtures (Courtesy Link Engineering Co.).

73
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 4.11 First page of Standard SAE J2521 report (Courtesy Link Engineering Co.).

in question. This means that the suspension are much more sophisticated forms of analysis,
must see the proper loading and be positioned but these provide a great place to begin.
as it would be on the operating vehicle.
The first page of the standard report is shown
The final and one of the most important in Figure 4.11.
chapters of the SAE J2521 standard is the
description of the required standard report At the top of the figure is a summary of the
formats. Although not all standards provide results of some of the key sections of the test
required report format, it was done with matrix. The requirement is to plot the forward
SAE J2521 to provide a common means of and reverse low-speed test segments. In this
quantifying performance that was lacking case, the tester has chosen to go further and
in the brake community at the time. These plot the drag and deceleration sections as well.
standards reports are a good starting point for Note the plot shows the spectrum of sound
analyzing the data from the test process. There pressure level amplitude versus frequency

74
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

with each symbol representing the peak level is a slight elevation of the percentage of
measured for a brake apply. occurrences in the second and third drags of
the Evaluation section. However, there are no
In this particular example, there are strong dramatic trends.
brake noise occurrences at approximately 1200
Hz and 14,500 Hz. The dominant contributors The second page of the standard report format
at these frequencies are the drag and is shown in Figure 4.12.
deceleration test sequences. The number and
level of the occurrences indicate a significant The purpose of this page is to examine trends in
noise issue with this brake. As often happens the noise that may be useful in understanding
with brake squeal, the squeals consistently the causes of the noise. In the upper left plot,
occur at particular frequencies. This supports the cumulative percentage of noise occurrences
the concept that loud tones occur at rotor is shown. Some have used this type of plot as an
or other component resonances. Since these indicator of the severity or potential customer
resonances are a property of the components, reaction to noise.
one would expect the noise to occur at the
same frequency each time unless the system is The upper right plot examines the relationship
changed. between noise occurrence and line pressure. In
this particular case, there does not appear to be
The middle plot on this page shows the a strong correlation. There is a slightly higher
temperature of the brake and the noise cumulative occurrence at 10 bar (145 lb/in2),
occurrences for each stop number. Note the but this is only marginally higher value than
left scale is temperature and the right scale is for other pressures.
the frequency of the noise occurrences. The
temperature plot indicates that the optional The lower left graph shows the relationship of
fade and recovery segment was run as shown noise occurrence to temperature. Again, there
in stops 1400 and higher. The optional low- is no strong trend shown. There is a slightly
temperature segment was not run. This plot higher percentage for the lower temperatures,
is a good “first look” to see if the brake noise but this is not a strong factor.
occurs predominantly in a particular test
section or in a particular frequency range. In Finally, the lower right shows the plot of
this particular case, the lower frequency noise cumulative occurrence versus speed. As is
occurs throughout the test matrix with no typically the case, there is a larger percentage of
strong tendency with respect to test segment occurrences at the lower speeds. Note that there
or temperature. On the other hand, the higher are very few occurrences at -3 km/h (-1.9 mph).
frequency noise does not begin until the second Since squeal tends to be a low-speed issue, these
evaluation cycle and does not occur after the results are not a surprise.
fade section.
One form of data reporting not required, but
Finally, at the bottom of Figure 4.11 is often used with SAE J2521 is shown in Figure
a summary of the percentage of noise 4.13.
occurrences for each test sequence. Because the
fade section has so few stops, the percentage This is a tabulation of the percentage of noise
occurrence is very high in this section. There occurrences in different frequency ranges. In

75
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 4.12 Second page of Standard SAE J2521 report (Courtesy Link Engineering Co.).

this particular case, the highest percentages are occurs after the fade. Experience has shown
in the 1000 to 2000 Hz range and the 14,000 to that fade may have no effect, may increase or
18,000 Hz range. Percentages of over 3% would generate noise for the first time, or cause noise
be considered a noise issue for most OEM to be reduced.
applications.
A much different case is illustrated in Figure
These examples of the results from an SAE 4.15.
J2521 test provide a good illustration of the
format for the plots and how they might be In this case, noise occurs at many different
used. However, it is useful to look at some frequencies and the fade and recovery section
specific extremes in the results to understand do not appear to have any impact on the
what are significant and meaningful trends. occurrence of noise. There is a strong tendency
Figure 4.14 shows an example that contradicts for noise to occur near the highest brake
what was seen in the previous example. temperatures at multiple frequencies ranging
from 100 to 18,000 Hz. Clearly, this is a very
In this case, there was no significant noise noisy brake with sound levels over 100 dB(A)
before the fade section and suddenly noise occurring at multiple frequencies.

76
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

Figure 4.13 Cumulative noise occurrence plot from SAE J2521 test
(Courtesy Link Engineering Co.).

In contrast, Figure 4.16 shows a case with a few frequently after the fade and recovery section.
strong frequencies of noise with some of the An important point to note in examining this
noise occurrences disappearing after the fade figure is the sound level of the squeals. The
and recovery. squeals above 10,000 Hz reach levels of 110
dB, which is quite intense and would be very
An interesting item in this test is that there disturbing to a vehicle occupant.
is green brake noise at roughly 7000 Hz that
is seen at the beginning of the test and then Finally, Figure 4.17 represents every brake
disappears as the brake is broken in. The noise development engineer’s nightmare.
at just over 10,000 Hz is quite strong until the
fade and recovery section and then disappears This is an extremely noisy brake with high-
afterward. The squeal at 2000 Hz is present amplitude squeals at multiple frequencies that
throughout the test, but appears to occur more occur throughout the test. Hopefully, such

77
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 4.14 Example of SAE J2521 test results: Noise after fade section
(Courtesy Link Engineering Co.).

performance is not a common sight, and clearly There are several other manufacturers making
this brake needs significant improvement similar dynamometers. These are only used as
before it would be commercially viable. examples.

The focus will be on two models from Link: the

4.9 Examples of Brake Dynamometers model 3900 shaft dynamometer and the model
The next step is to examine commercially 4900 chassis dynamometer. Drawings of both
available dynamometers and their capabilities are shown in Figure 4.18.
and characteristics. For the purpose of this
text, Link Engineering brake dynamometers The most significant difference between shaft
will be used to illustrate the different types dynamometers beyond motor power and
of dynamometers available. This is not a inertial ranges are the enclosures used. Figure
commercial endorsement of this manufacturer. 4.19 shows the extremes of enclosure design.

78
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

Figure 4.15 Example of SAE J2521 test results: High noise occurrences
(Courtesy Link Engineering Co.).

In the upper left is a very small enclosure One advantage of this enclosure is that
mounted on the dynamometer’s frame. Such there is a single door and no windows. This
a small enclosure is possible because this user will reduce environmental and acoustical
does not run tests with vehicle suspension, and sealing issues and probably lead to good
therefore, is not concerned about frequencies performance.
below 2000 Hz. By mounting the enclosure on
the dynamometer frame, there can be more The more common and one of the largest
vibration and higher background sound levels shaft dynamometer enclosures is shown at the
in the enclosure. This user was able to live with bottom right of Figure 4.19.
these compromises to control the size and cost
of the dynamometer. This would not be the This enclosure provides the room to install
dynamometer to look for a 1500 Hz squeal at 72 suspension corners and, in the case of solid
dB(A). axles, whole axles for test. The platform shown

79
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 4.16 Example of SAE J2521 test results: Strong noise suppressed after fade
(Courtesy Link Engineering Co.).

has a T-slotted surface to permit the attachment This unit is typically used with one- or two-
of fixtures and axles for proper location of corner assemblies mounted and acoustic
components. The double doors present some enclosures placed over each corner section.
seal issues, but they are required to install large The floor level is at the surface on which the
axles into the unit. tire is running. While some excavation may be
necessary, this is minor compared to a four-
In the case of chassis dynamometers, there are roll chassis dynamometer. At later stages in
a lot of different approaches taken to provide the development program when a vehicle is
the needed capabilities. A few examples will available, the corner sections can be removed
be considered here. Figure 4.20 shows a design and one end of the vehicle placed on the rolls
to achieve the ability to run a brake corner on for brake noise testing. This is a compromise
a road wheel, but avoid some of the cost of the that falls in between the shaft and four-wheel
four-wheel dynamometer with large rolls and chassis dynamometers in terms of cost and
drive mechanisms. capabilities.

80
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

Figure 4.17 Example of SAE J2521 test results—Very noisy brake (Courtesy Link Engineering Co.).

Figure 4.18 Link models 3900 and 4900 brake dynamometers (Courtesy Link Engineering Co.).

81
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 4.19 Variations in test enclosures (Courtesy Link Engineering Co.).

A full four-roll chassis dynamometer is shown above the rolls when the rolls are running at
in Figure 4.21. 100 km/h (62 mph).

The photograph on the left shows a vehicle Figure 4.22 shows a drawing of the
running on the chassis dynamometer with configuration of the chassis dynamometer as
a semianechoic room installed above the installed.
dynamometer. Note that when running such
a test the vehicle must be restrained and the There are many steps taken in this design
exhaust gases ducted outside of the room for to ensure low background sound levels and
safety. On the right side of the figure is the accurate measurements. To minimize the
chassis dynamometer before it was installed sound in the test chamber, it is constructed
in the pit below the test room and before as a room within a room. This provides high
the motor drives were connected. Note that transmission loss and better insulates the test
the rolls are fully enclosed except for the cell from outside noise. The magnitude of the
portion that is exposed for the tire interface. space consumed by the acoustic wedges for
This is done to reduce noise from the rolls sound absorption on the ceiling and walls is
to meet background noise requirements in evident. Note that the foundation for the test
the test chamber. In this particular case, the cell is isolated from the building foundation. In
specification was for 50 dB(A) at 1 m (3.3 ft) addition, the road wheels and motors are placed

82
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

Figure 4.20 Link model 4900 two-roll chassis dynamometer (Courtesy Link Engineering Co.).

Figure 4.21 Link four-roll chassis dynamometer (Courtesy Link Engineering Co.).

83
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 4.22 Drawing of four-roll chassis dynamometer installation (Courtesy Link Engineering Co.).

on large concrete masses that are isolated from is that there should be two wavelengths
the rest of the building structure. This is done between the test object and the nearest wall
to reduce the transmission of vibration and the surface for a proper test environment.
resulting noise from one structure to another.
The drive motors are located in a separate room Since both shaft and chassis dynamometers
with seals around the moving shaft passing can be used for brake NVH testing, the
between the rooms to minimize the noise question is when is the proper application of
transmitted into the pit below the test chamber. each? The chassis dynamometer provides the
most accurate simulation of actual vehicle
Do not be confused by the three rolls shown operation and dynamics. However, with the
in the side view. Since one role is moveable, vehicle tires running on the rolls surfaces,
this drawing reflects the amount of movement the background level is much higher than
possible to accommodate vehicles of different that in the shaft dynamometer. To run tests
wheelbase. It should also be noted that there on the chassis dynamometer, at least a full
is a significant amount of clear space around vehicle corner is required as shown in Figure
the vehicle. This is important for free field 4.20. However, for low-frequency tests and
performance by the room. If the room walls are vehicle interior noise measurements the
too close to the vehicle under test, a free-field chassis dynamometer is the best approach.
simulation will not be possible. The calculation For understanding the interior noise impact
of the required free space is dependent on the of brake noise and vibration, there is no
frequency of interest, but a good approximation substitute for the chassis dynamometer.

84
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Laboratory Testing

As has been shown previously, one can choose evaluations, the shaft dynamometer is an
between a two-wheel and a four-wheel chassis efficient and useful tool.
dynamometer. The two-wheel model is roughly
the same price as the shaft dynamometer. It In the end, the decision as to which type of
provides a good measurement platform for dynamometer to use depends on budget, the
a wide range of brake studies. However, it stage in the product development at which the
cannot replace the four-wheel dynamometer test is to be run, and the frequency range or
for simulating on-road operation of the vehicle. type of noise issue to be evaluated. The chassis
The two-roll dynamometer can be installed in dynamometer is as or more expensive than the
most facilities without major modifications. shaft dynamometer. For low-frequency noise
The four-roll dynamometer requires a custom- and especially judder, the chassis dynamometer
designed facility or major modification is best. Early in the product development
to an existing facility. Not counting the cycle, the shaft dynamometer may be the
building modifications, the four-roll chassis only solution since a vehicle or a complete
dynamometer costs run 1.5 to 3 times the suspension corner may not be available. Late
amount for a shaft dynamometer. With the in the development process, testing can be
required building modifications and the cost, done on either type of dynamometer. These
the decision to build a chassis dynamometer is are decisions that each investigator must make
not an easy one. There are less than 10 four-roll for his- or her self. There is no right answer.
chassis dynamometers in the world dedicated However, as will be discussed in a later chapter,
to brake noise testing. there is a trend to greater usage of the chassis
dynamometer driven by a trend of more lower
The shaft dynamometer has been used for frequency noise issues and the need to make
brake NVH testing for some time. In the last fine distinctions based on vehicle interior noise
decade, the ability and sophistication of such measurements.
dynamometers has increased significantly.
Today, a fully loaded shaft-brake NVH
dynamometer costs more than $1 million. 4.10 Summary
There are two primary advantages of the This chapter has shown brake NVH
shaft dynamometer: it is very flexible being dynamometers to be sophisticated test
able to test a range of configurations from machines that continue to grow in complexity
only the brake parts to a full-vehicle corner to meet changing test standards and the
and it can be installed and operated in any challenges of recreating ever more difficult
laboratory with minimal site preparation. The noise issues. There is an internationally
major deficiency of the shaft dynamometer accepted noise measurement standard, SAE
is that it does not do a good job reproducing J2521, that has played a key role in reducing the
low-frequency brake NVH issues. With the number of brake squeal issues. This standard
use of a passive road wheel, it may be able will continue to evolve and be supplemented by
to overcome this limitation in many cases. other brake NVH measurements standards.
The major exception is judder where the
whole vehicle with the steering mechanism When it comes to running laboratory brake
is required to completely reproduce the tests, there are many issues to be considered.
phenomenon. For high frequency squeal Not the least of which is how to fixture the
testing and lining or corner components brake corner including the suspension. It has

85
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

been shown that it is important to consider how Finally, the different types of dynamometers
complete the suspension corner should be for were discussed. Both a chassis and a shaft
testing and that one’s results can be limited by dynamometer can be used from brake NVH
this choice. testing. In some instances either can be used to
generate the same sort of results.

86
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 5
On-Road Testing

87
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

5.1 Introduction It is critical to concentrate on tests that identify

The goal of this chapter is to provide a complete interior noise due to brakes and that eliminate
discussion of the equipment and methods used in or prevent interference from noise due to other
on-road brake NVH testing. The discussion will sources. Even in modern vehicles in which
begin with the instrumentation used for such much has been done to reduce interior noise,
tests. Their major components and performance there are many noises due to many sources.
requirements will be described in some detail. A Later in this chapter, the task of minimizing the
good question to consider is: Why can one not effects of non-brake noise will be discussed.
use a simple meter for vehicle sound tests?
Before going further, it is important to note
Next, the existing test standards and some that one of the most important goals in on-
common tests will be described. This will road measurement is to measure the critical
include international standards, industry operating conditions when a noise event
standard practice, and problem investigations. occurs. While it is important to measure the
There is a wide range of on-road tests run for interior noise to understand the annoyance
brake noise evaluation. These include screening and potential customer reaction, it is just as
tests to evaluate the potential for brake noise. important to know the speed, line pressure,
Such tests often include severe or specialized temperature, and other parameters that
operating conditions designed to elicit brake describe the brake operational conditions.
noise. At the other end of the range of tests are With these additional measurements it may be
the problem resolution studies where tests are possible to diagnose the problem, determine
conducted to understand or evaluate solutions if the operating conditions are common ones
for a particular problem. Such tests can be that a customer will encounter frequently, and
performed in as short as a few hours. However, possibly to provide the necessary information
screening tests can take as long as weeks to to allow one to recreate the conditions using a
complete. This range illustrates the variation in dynamometer to allow laboratory studies.
the nature of on-road NVH testing. Examples
of each of these will be discussed.
5.3 Vehicle Instrumentation
Finally, current measurement systems and their Depending on the goal of the particular
capabilities will be discussed. The goal will be test, the required instrumentation can be
to provide insight on how to prepare for and quite varied. As a baseline, the minimum
conduct on-road brake NVH measurements. requirements will include:

• Accelerometers at each corner;

5.2 Goals for On-Road Brake NVH • At least one thermocouple (or other
Testing temperature measurement transducer) at
The focus this text and the discussion of on- each corner;
road vehicle testing is occupant annoyance. • A brake switch indicator to provide a signal
While brake squeal in urban environments can when the brake is applied;
be an issue, this will not be discussed here. For • A speed measurement; and
this treatment, exterior noise is only important
as to how it relates to interior noise. • An interior microphone.

88
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

The accelerometers at each corner provide a applied and stopping data acquisition when the
record of vibration that can be compared to the brake is released. This approach assures that
interior noise to assist in identifying whether data is acquired when it is needed and avoids
measured noise is due to the brakes or some the collection of vast amounts of unwanted
other source. There are some variations to data.
the practice of using accelerometers mounted
on the brake. Some testers insist on using Clearly, collecting vehicle speed is important
microphones in the vehicle wheel well instead for understanding and recreating noise events
of accelerometers. The reasoning for this in the laboratory. In some cases, testers collect
approach is that for squeal the dominant path the speed of the vehicle and the rotational
for energy transmission from the brake to the speed of each wheel. This allows them to know
interior is airborne. Therefore, by measuring when slip occurs. For noise testing, vehicle
exterior noise one has better correlation to speed alone is generally sufficient. Today, this
the interior noise. Those using accelerometers data is most often acquired from the vehicle’s
cite practical reasons for their choice on-board data system, a GPS device, or a
instrumentation. The first is that microphones tachometer on one of the vehicle’s driveline
are expensive and somewhat fragile rotating members. At some test facilities,
transducers. Placing them in a wheel well where a fifth-wheel will be used to acquire speed
they will be exposed to water, rocks, mud, and information.
other road debris is likely to lead to inaccurate
measurements or failure. In addition, the Finally, the most important measurement is the
exterior microphone cannot distinguish interior noise. There are a variety of means used
whether the noise measured is coming from the to collect this important information. For the
brakes of the test vehicle or the bus that is in majority of tests a microphone is placed next
the adjacent lane with loud brake squeal. Since to the driver’s ear. This can be hung from the
the accelerometer is measuring brake vibration vehicle headliner or suspended from the seat.
directly, it is not affected by other noise sources The most critical issue is the placement and
on the test vehicle or from other vehicles on resilient support for the microphone so that it
the road. The vast majority of investigators use does not see excessive vibration. The selection
accelerometers to provide a reference signal of the microphone has already been covered in
from each corner. The use of microphones is the previous section on instrumentation.
predominantly limited to studies in Europe.
The second most common approach is the use
It is important to obtain information on the of a binaural head with microphones placed in
brake temperature. The use of a thermocouple each ear. Such a unit is shown in Figure 5.1.
typically embedded in one or both brake pads
is very common. This provides a reliable record Usually placed in the front passenger seat,
of temperature, which can be critical in the this unit provides a much better simulation
coalescing of modes to create noise and in of what is heard by the vehicle occupant and
recreating the noise in the laboratory. offers the benefit of binaural signals to help in
locating sources. While more expensive and
The brake switch is most often used as a trigger more elaborate to configure, this can provide
for data acquisition. A simple signal can be much greater insight into noise issues that are
used to trigger sampling when the brake is encountered.

89
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 5.1 Head and torso simulator for binaural sound measurements.

Figure 5.2 Two on-road brake NVH measurement systems (Courtesy Link Engineering Co.).

90
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

The other major part of the measurement vehicle was to be used for other tests requiring
apparatus is the data acquisition system. this capability.
Two examples of such systems are shown in
Figure 5.2.
5.4 Vehicle Test Procedures
These are Link systems. The model 3501 is With this introduction, the next step is to
shown on the left, and the model 3801 is shown look at specific test procedures to better
on the right. These will be discussed in greater understand how such tests are conducted and
detail in a later section of the chapter. the instrumentation used. Three test procedures
will be discussed. The first is the SAE J2625
Figure 5.3 shows a full instrumented vehicle [9] standard, which is the only internationally
with the instrumentation discussed noted. recognized vehicle brake noise test procedure.
Next, the Los Angeles City Traffic procedure
In this case, an optical sensor is used to will be discussed. This is a widely used screening
measure speed. This is not the typical case for procedure. Finally, an example of a problem
noise testing. It was done in this case since the resolution test program will be provided.

Figure 5.3 Instrumented vehicle (Courtesy Link Engineering Co.).

91
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

5.4.1 SAE J2625 Instead, the specification of less than 10oC (50

The SAE J2625—Automotive Vehicle Brake
o
F) is used. Low-pressures brake applies are
Squeal Test Recommended Practice was used to elicit noise, but not rapidly heat up
published in 2008 as the first brake NVH test the brakes. Braking is performed at 10 km/h
standard for vehicle testing. The concept for (6.2 mph) in both the forward and reverse
this procedure is to provide a core vehicle directions. It is important to note that the test
squeal evaluation module. It can stand on requires a soak of four hours at less than 10oC
its own or be part of another procedure. It (50 oF) to be sure all the brake parts are at a
is focused on low-speed tests with a range uniform temperature.
of temperatures and pressures as part of the
evaluation. It includes a module for evaluating The first required test section is the Parking Lot
“morning sickness” at low temperatures. The Maneuver – Low Temperature Sequence. In this
goal was to provide a relative quick evaluation section, the initial brake temperature is kept
that could be conducted in a parking lot if below 50oC (122oF) with forward and reverse
necessary for a quick instrumented assessment direction brake appliers at 5 and 10 bar (73 to
of the propensity of a vehicle to generate squeal. 145 lb/in2). The goal is to determine whether
The test typically takes less than four hours, so brake squeal is generated by low-speed and low-
it can be an efficient quick screening procedure. pressure applies that are commonly a source of
squeal. Depending on the state of the vehicle
Table 5.1 provides a summary of the test before the start of the test, it may take care to
matrix. cool the brakes to the necessary temperature.
Section 9 describes the required cool-down
The first step is preconditioning of the friction procedure to be run as part of this test to get
interface. A test sequence is not specified, the brake to the desired temperatures. On
so that the user is free to use his/her own warm days, there may be several cool-downs
procedure. If the user has no condition required to complete this sequence.
procedure, the FMVSS 135 burnish procedure
is recommended. It is anticipated that in some The next section is an Initial Brake
cases this procedure may be inserted within Temperature (IBT) matrix. The temperature
an existing vehicle evaluation process and that range is from 50 to 250oC (122 to 482oF) at
preconditioning may already have occurred. speeds of 20 and 40 km/h (12.4 and 25 mph).
This section is different from the previous ones
The next section is the optional Winter Parking in that deceleration rates of 0.1, 0.2, and 0.3 g’s
Lot Maneuver. It is designed to provide a are defined as opposed to line pressures. Since
quick check for “morning sickness” or cold experience has shown that these decelerations
weather noise. Obviously, one is limited by were effective in generating squeal, it was
the environmental conditions. Ambient important to match the accelerations knowing
temperatures will determine whether this test the actual line pressure required would vary
section can be accomplished. An alternative among vehicles.
used by some is to soak the vehicle in a cold
chamber and then run this low-temperature This section is followed by a Parking Lot
section quickly before the brake parts have Maneuver—High Temperature Sequence.
heated up. Because of the above limitations, Employing a speed of 10 km/h (6.2 mph) in
a specific brake temperature is not specified. both the forward and reverse directions noise

92
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

Table 5.1
Summary of SAE J2625 Test Matrix
Initial
Section Number of Speed, Pressure,
Description Temperature, Notes
Number Applications km/h bar o
C

Open to user, US FMVSS

1 Preconditioning * * * *
135 burnish is suggested

Optional Winter Test begins with a 4 hour

2 4 10, -10 5 & 10 < 10
Parking Lot Maneuver soak at < 10o C

Parking Lot Maneuver

3 – Low Temperature 12 10, -10 5 & 10 < 50
Sequence

Initial Brake
Decelerations defined,
Temperature (IBT) 20 &
4 30 * 50-250 not pressures, 0.1, 0.2,
Matrix—Increasing 40
& 0.3gs
Temperature

Parking Lot
Maneuver—High
5 12 10, -10 5 & 10 > 250
Temperature
Sequence

IBT Matrix— Decelerations defined, not

20 &
6 Decreasing 36 * 50-300 pressures, of 0.1, 0.2, &
40
Temperature 0.3gs

Parking Lot
7 Maneuver—Moderate 12 10, -10 5 & 10 > 50
Temperature

0.3 g snubs until desired

Warm-Up—when temperature is reached
8 70-35 *
required

Cool-Down—when Constant speed operation

9 70 0
required without brake applies

measurements are made for initial brake to warm-up the brake to assure comparable
temperatures greater than 250oC (482oF). results to others running the same test matrix.
Pressures of 5 and 10 bar (73 to 145 lb/in2) are
used for this test. The difficulty with this test Next is the IBT Matrix—Decreasing
is sometimes in getting the brake temperature Temperature section. Coming from the
to the required value in a small parking lot. previous section it is assumed the brakes will be
Section 8 specifies the warm-up procedure warm and this procedure then conducts a series
that is to be used to get the brake to the proper of tests with decreasing temperature. Once
temperature. This procedure must be used again, the deceleration required is specified

93
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

as opposed to the line pressure. The tests are procedures that must be followed as part of
conducted at 20 and 40 km/h (12.4 and 25 this test standard. This should be used for any
mph). necessary temperature adjustment.

The final section is the Parking Lot Maneuver— Figure 5.4 shows an example of a test report
Moderate Temperature. Low-speed tests are made using SAE J2625.
done for temperatures above 50oC (122oF).
While this temperature range may have been In this particular case, only the front brakes
covered by previous tests, these were done at were instrumented. The upper left hand
higher speeds. This section is run at 10 km/h tables in the figure summarize the operating
(6.2 mph) in both the forward and reverse condition ranges for the brake and the vehicle
directions with light pressures of 5 and 10 bar parameters. The top graph shows the trends in
(73 to 145 lb/in2). noise occurrences with brake line pressure and
brake temperature. It appears that there is a
As has been previously noted Sections 8 cluster of noise occurrences centered around 20
and 9 specify the warm-up and cool-down bar (290 lb/in2) pressure and 170oC (338 oF). The

Figure 5.4 SAE J2625 example results report (Courtesy Link Engineering Co.).

94
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

graph just below shows the noise occurrences as accelerometer data to the interior sound
a function of deceleration and speed. As would pressure data. When frequency peaks from a
be expected, the majority of the noise clusters at corner align with noise in the interior, the noise
just below 10 km/h (6.2 mph) and decelerations is attributed to that corner. In this case, there
between 0.2 and 0.3 g’s. were roughly 40% more occurrences on the
left side than the right. As has been noted, this
The plot on the lower left of the figure show the is an unusual case and should be investigated
traditional frequency spectrum. It is interesting further.
to see that there is a distinction between the
noise from the left and right positions on
5.4.2 City Traffic Tests
the vehicle. The left brake shows a cluster of
occurrences at 3000 and 7000 Hz. The right One of the commonly used brake squeal
position shows a cluster of noise occurrences vehicle screening tests is the city traffic test.
at 11,000 Hz. This is an interesting result The general approach is to run a vehicle for an
that should prompt an investigation as to the extended period of time in urban traffic with
difference between two nominally identical routes typically selected to provide operating
brake corners. This could be the result of conditions that tend to induce squeal. In the
manufacturing variation in the rotors, brake United States, there are several such test routes
pads, or calipers. In this case, SAE J2625 in Los Angeles, Detroit, Phoenix, and other
provided some useful information. urban centers. In Europe, the predominant
route is Mojacar in Spain. There are many other
An important component in this spectrum city traffic routes around the world including a
is the black line, which is the daily average of relatively new Shanghai city route.
sound levels in the vehicle interior. Early in
this text, the concepts of averaging and peak In the United States, the Los Angeles City
hold in the sound spectrum were used. In Traffic (LACT) test is the most widely used
this instance, the average is used to show the and accepted. Most auto manufacturers
nominal sound level versus frequency in the selling vehicles in the U.S. employ this test
vehicle. This average line forms an important as a final sign-off procedure on brake noise
basis of comparison to assess the magnitude of and sometimes wear. The duration of this test
squeals. Unlike the case of NVH dynamometer varies but generally runs from 16,000 to 32,000
where an absolute threshold can be set, in a km (10,000 to 20,000 miles) running on Los
vehicle the sound level changes with speed Angeles metropolitan area roads. With roughly
and each model. The best reference point is 320 to 400 km (200 to 250 miles) covered per
the sound level when there is no squeal. The day, this is a multiple week test. The average
average shown here represents this threshold. number of brake applies in the route per day
The squeal occurrences above this line are most has risen significantly over the last several
likely audible by vehicle occupants and are years due to traffic increases to over 1000 on
likely to lead to noise concerns by the customer. average. As was noted above, the primary focus
is to monitor wear and noise. In most cases,
The table at the lower right of the figure reports are generated each day to track trends
provides information on the number of in noise performance in particular sections
squeal occurrences identified for each corner. of road, under particular conditions, or to see
This distinction is made by comparing variations over multiple days. Many testers use
the percentage of occurrences of noise through

95
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

the course of the test as a primary indicator of will be examined in some detail here. The
whether a brake system is acceptable. LACT route is shown in Figure 5.5.

As an example of the whole category of city This map illustrates the variety of roads, traffic,
traffic tests, the Los Angeles City Traffic test and environments covered in test. The route

Figure 5.5 Los Angeles City Traffic Route (Courtesy Link Engineering Co.).

96
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

includes highways to suburban roads with for 16 thermocouple channels, 16 strain-

high to moderate speeds and moderate brake gage channels such as pressure, 12 speeds,
applications. It also encompasses cannon and six NVH channels. At the right is the
roads with steep descents and heavy brake instrumentation applied to one corner. The
applications. There is a long stretch of driving accelerometer and the signal wire for the
in the city with the typical stop and go traffic. thermocouple can be clearly seen. Note that
The route includes significant changes in both the accelerometer and the thermocouple
altitude and near desert to seaside locations for are attached to the outside brake pad. In a
significant temperature and humidity changes. typical test this same instrumentation is
The route is effective in exposing the braking repeated at each corner. In addition, a vehicle
system to a wide range of conditions with the speed signal is obtained from one or more
exception of extremely cold temperatures. sources, and there is an interior microphone
and a switch to indicate when the brake is
A typical configuration for the LACT is shown applied. In some cases, there will also be a
in Figure 5.6. button or some other device available for the
driver to indicate when he hears brake squeal.
On the left is the data acquisition system
with laptop for control and data storage. An example of a typical daily summary from a
Th is Link model 3501 system is configured LACT test is shown in Figure 5.7.

Figure 5.6 Los Angeles City Traffic Route Instrumentation (Courtesy Link Engineering Co.).

97
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 5.7 Typical LACT Daily Summary Report page 1 (Courtesy Link Engineering Co.).

At the top of this figure is a plot of the interior average, the correlation with the driver’s subject
sound pressure level versus frequency. In this evaluation diminishes. It is not unusual for
particular case, the right rear brake seems there to be a difference between the number
to be the noisiest with a strong tendency for of events noted by the driver and the objective
noise at 7200 Hz. With the daily average sound total. There are several possible causes: the
level at less than 30 dB(A) at this frequency, driver may have been distracted or busy and
the squeals at over 60 dB(A) are clearly unable to indicate an event, the driver may not
audible by vehicle occupants and represent have heard the event due to other loud noises
a substantial noise issue. The yellow dots in in the environment, the driver may have gotten
this plot represent the times where the driver fatigued by the numerous events, etc.
cited a noise event. There is a good correlation
between the subjective and objective data The graph at the lower left shows the correlation
for the higher levels at 7200 Hz. It appears between speed and deceleration to noise
that as the sound levels get closer to the daily occurrences. Again, the right rear position

98
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

dominates with the majority of the noise measured objectively. The important point is
occurring below 60 km/h (37 mph) and 0.25 g’s that 13 events in one day are significant, and
of acceleration. 53 events only indicate a more severe noise
problem. This is clearly a brake squeal issue in
In the middle right of Figure 5.7, the noise need of attention.
events for each corner position are shown. This
is a very unusual case in which 50% of the noise The second sheet of this daily summary is
events are attributable to the right rear position. shown in Figure 5.8.
This bias to the right rear could indicate a
problem with the right rear and perhaps the The top graph shows the temperatures for all
right rear brake should be inspected before four corners versus distance along the route.
continuing the test. Another important piece There are a wide range of temperatures from
of data in this table is the number of subjective a little above 0 to over 200oC (32 to 392oF).
noise events recorded by the driver. There There are some gaps in this data where no
were only 12 while the data indicates 53 events temperatures were recorded. This is not an

Figure 5.8 Typical LACT Daily Summary Report page 2 (Courtesy Link Engineering Co.).

99
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

error, but instead indicates stretches of road A better perspective on the effects of operational
where the brakes were not applied. This parameters on noise can be seen in Figure 5.9.
represents perfectly normal operation. It is
easy to see that the driver’s lunch break was at The frequency of noise is compared to brake
around the 140 mile (225 km) point. pressure, temperature, speed, and deceleration.
In each case, there is a significant range of the
The lower graph in this figure illustrates where parameter at which the squeal occurs. The most
the noise events occurred in relationship to important result from this particular plot is
temperature and mileage. Most of the noise that there is not a critical set of parameters that
events occurred in the 100 to 150oC (212 to have to be achieved for the noise to occur.
302oF) temperature range. There were two
clusters of noise events after 50 miles (80 km) Projecting to a period of several weeks where
and after 140 miles (225 km). one would have 10 or more of these daily

Figure 5.9 Typical LACT Daily Summary Report page 3 (Courtesy Link Engineering Co.).

100
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

reports, it is easy to see that the result would the problem. Five NVH measurements
be a good picture of brake noise performance. were made along with several operational
There would be sufficient data to see trends measurements including speed, temperature,
in terms of operational conditions that and line pressure. The five NVH channels
lead to noise, environmental effects, and were: an interior microphone, a microphone
the magnitude of any noise issues. It is also in the wheel well, and accelerometers on the
evident that this test generates a lot of data, outer pad, the caliper bridge, and the anchor
and it would be difficult to go through each bracket. The exterior microphone was included
of the spectra from each of the five NVH because the customer insisted that this data be
measurement channels and perform detailed collected.
analysis. Therefore, many testers rely on
the total percentage of noisy stops as a first The instrumented corner is shown in Figure
indicator to see if further analysis is necessary. 5.10.

The accelerometers mounted on the caliper

5.4.3 Problem Squeal Resolution
bridge, outer pad, and anchor bracket are
The final category of on-road evaluations that shown in this photograph. Note, that in
will be discussed is a program to understand addition to covering three major components
and resolve a squeal-noise issue. In this of the brake assembly, these locations allowed
particular example, a customer was supplying the quantification of vibration in the three
brake hardware to a vehicle manufacturer who principle directions of the brake system.
was a few weeks short of the start of production
of a new vehicle when a brake squeal problem Typical data from the measurements made on
was discovered. The supplier had a very short this vehicle are shown in Figure 5.11.
time to resolve the problem and came to the
author for assistance. Fortunately, the supplier It is clear there is more than one strong noise
was able to provide a lot of information about peak in the spectra. The top left spectrum
the nature of the problem. This allowed testing for the interior microphone illustrates that
to quickly replicate the problem and move to the problem noise is the strong peak at
trying to resolve the issue. This information approximately 7700 Hz. Having amplitude
about the occurrence of the noise meant that of approximately 70 dB(A), this was a high
a great of deal of experimentation to attempt level sound that was very disturbing to vehicle
to isolate when the noise occurs was not occupants. The analysis of the problem in
necessary. In many problem resolution studies, this case was greatly assisted by the strong
this first phase of finding the noise can be correlation between the interior noise and
the most time consuming part of the entire the vibration measurements. Looking at each
program. of these plots it is clear that the same strong
spectral peak is seen in each measurement.
In this case, it was known the noise occurred
only on the front corner and at moderate In the case of the exterior microphone, the peak
temperatures and decelerations. Therefore, reached an amplitude of 90 dB(A). This was a
one front corner of the vehicle was fully loud squeal at a very well defined frequency.
instrumented to acquire the information The anchor bracket showed a strong peak
necessary to quantify and understand with 30 g’s of amplitude. This is a very high

101
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 5.10 Problem squeal testing vehicle corner instrumentation (Courtesy Link Engineering Co.).

amplitude and represents major vibration performance. The interior noise level was
amplitude at this frequency. There were also decreased by 7 dB at the problem frequency. The
very high levels of acceleration at the other subjective rating of the vehicle was improved
measurement points. Normal to the surface of by two to three points. This was a major
the pad the acceleration was 22 g’s. improvement and allowed the customers parts
to be used on the production vehicle. This is an
This project to resolve the problem squeal used example of an excellent test program in which
both dynamometer and vehicle measurements. the necessary information and parts were
Initial investigations showed that there was available. As a result, a dramatic improvement
good correlation between the dynamometer and was accomplished in a very short period of time.
the vehicle as would be expected in this higher
frequency range with a strong discrete peak. In summarizing this section, there are a
number of different on-road tests that need to
Within a few weeks, a solution was found that be performed. While there are some common
made a dramatic improvement in the noise requirements, there is significant variation in

102
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

the details. In some instances, like the problem endorsement of the Link systems. They are
resolution, results are needed at the end of simply used as examples for discussion.
each stop and there cannot be too many NVH
measurement channels. In other cases, such The first system is the model 3801 as shown in
as a city traffic test, data at the end of the day Figure 5.12.
or week may be sufficient. With these varying
requirements in mind, the discussion now This unit is intended for use in city traffic
shifts to the measurement systems themselves. type tests. This unit is compact measuring
317.5 mm (12.5 in) by 228.6 mm (9 in) by 95.3
mm (3.75 in) high. It can easily be placed in
5.5 On-Road Measurement Systems multiple positions in the vehicle. The standard
There are a wide variety of measurement unit shown here has six NVH channels, eight
systems that can be used for on-road testing. thermocouple channels, and eight high-level
For the purpose of this discussion, models channels that can be used for general input or
made by Link Engineering will be used as they can provide conditions for strain-gage
examples. This is not meant to be a commercial channels such as pressure transducer and six-

Figure 5.11 Problem squeal test results (Courtesy Link Engineering Co.).

103
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

pulse channels for speeds signals. There also of stop analysis or data display. The computer
digital signal inputs to permit input signals for inside the units manages the data acquisition
brake pedal switches and test-driver noise event and stores the data for later retrieval and
designation. analysis.

This array of channels is sufficient for most For problem resolution tests or other test where
city traffic type tests of passenger vehicles. end of stop data analysis or driver feedback is
It allows for one NVH measurement at necessary, Link provides the model 3501 system
each corner and two measurements in the as shown in Figure 5.13.
vehicle interior. The system also permits two
temperature measurements at each corner, As illustrated in this figure, this unit can
speed measurements from the four corners, acquire data from a large array of transducers
two additional gross vehicle speed signals, and and uses a laptop to provide data analysis and
up to eight line pressure or other signals. For reports in the vehicle.
commercial vehicles or other applications with
more wheels, there are options to add channels. The model 3501 is a modular unit that can be
configured for particular applications or classes
Since this unit is intended for city traffic type of tests. The most common modules for vehicle
tests, there is a very limited capability for end brake testing are shown in Figure 5.14.

Figure 5.12. Link model 3801 data acquisition system (Courtesy Link Engineering Co.).

104
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

Figure 5.13 Link model 3501 data acquisition and analysis system (Courtesy Link Engineering Co.).

The top module is a general purpose data with two line pressure measurements per
acquisition module. It provides the ability vehicle being typical. It is also used to collect
to acquire eight channels of signals from a data on brake pedal apply forces when this
wide range of transducers providing voltage type of data is required.
outputs of to 10 volts. This unit also provides
the control in synchronizing functions for the Just below this unit is the thermocouple module
module stack. that provides conditioning for 14 temperature
measurements. This permits one or more
Next is the encoder module. This unit provides temperature measurements at each corner with
the ability to convert a pulse train from extra capacity to handle vehicles with more
an encoder or other speeds sensing device than two axles.
and provide a signal proportional to speed.
There are eight channels that can be used for For the purposes of this text, the next module
measuring wheel or vehicle speed using this in the stack is the most important: the NVH
unit. module. For this unit, a Brüel & Kjaer PULSE
system is used. This unit provides six or more
The next unit down in the stack is the NVH measurement channels with the signal
strain gage module, which provides the condition for most commercial accelerometers
ability to condition any sort of strain gage and microphones. If necessary, this unit can be
or bridge type transducer. Th is module is expanded to provide more than 1000 channels.
most commonly used for pressure and load
transducers applied to the vehicle. The unit is Below the NVH module is the vehicle bus
most frequently used for brake line pressure interface module. This unit allows one to

105
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

extract data from the vehicle bus. Signals This unit provides a minimum of six input
such as vehicles speed, engine speed, and channels. It can be expanded to 1000 channels
other parameters can be obtained in this way. or more if required. As with laboratory testing,
This unit is unique in that, instead of simply this unit is typically configured to do peak-
providing digital signals from the bus, it hold sampling when doing squeal testing. The
converts the digital signals to analog signals other sampling parameters are just like for a
that are acquired and stored along with the true laboratory test: 50% overlap, Hanning windows
analog signals to provide full synchronization with sampling rates at roughly 50,000 Hz.
for all the data channels.
The one unique aspect of vehicle testing is
The final module is the Link power module. that stop length is a complete unknown. On
This unit provides battery backup to power the a dynamometer, one knows that the stops or
stack should the unit be disconnected or should drags will be no more than several seconds
voltage from the vehicle electrical system drop and data storage or processing large amounts
too low during starting or other situations. of data are not the problem. In vehicle testing,
one often encounters two extremes. The first
The NVH module or B&K PULSE system is is a rapid succession of short stops. This often
shown in Figure 5.15. occurs in urban traffic situations and frequently

Figure 5.14 Link model 3501 configuration (Courtesy Link Engineering Co.).

106
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

On-Road Testing

Figure 5.15 Brüel & Kjaer NVH data acquisition and analysis module
(Courtesy Brüel & Kjær Sound & Vibration Measurements A/S).

can be a situation where noise occurs. The does one establish a criterion to confirm that
challenge for the NVH system is to be able a noise event has occurred? Since it is critical
to acquire the noise, process the data, and to have an automated process to handle the
be ready to acquire new data before the next vast amounts of data that are generated by on-
stop occurs. This can sometimes be tenths of a road or dynamometer tests, the selection of
second. The B&K analyzer has shown that it is criteria becomes very important. The simplest
capable of doing this. of criteria is the constant threshold used across
the entire squeal frequency range as done in
The other extreme case is the long stop. On SAE J2521. While the value of 70 dB(A) may
mountain descents, the driver may apply the not be universal, this is the most common form
brakes for long periods of time. One of the of squeal detection in the industry. However,
commercially available systems stops collecting this approach only applies to laboratory
NVH data after 15 seconds due to a data storage dynamometers.
problem. While stops of over 15 seconds are not
the norm, due to the high heat buildup these can For on-road testing, a threshold is generally
be noise-producing stops. The Link system with established using some representation of the
the B&K analyzer can collect data for as long as background sound levels in the vehicle. The two
is necessary. The author has done test cases with most common approaches are to use a unique
stops lasting 30 minutes with no difficulties. test to establish one or more background
spectra or to use a data processing technique to
The final topic for discussion in this section is establish an average threshold for a period of
the general topic of noise-event criteria. How testing such as a day.

107
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Modifications to this approach use different 5.6 Summary

amplitudes for different frequency ranges. On-road brake noise measurement has
Many of these thresholds use higher values advanced significantly in a few years. Five
at the highest frequencies to diminish the years ago there were no well-recognized
effects of squeals near the lower limits of measurement standards. SAE J2625 has added
hearing. There is no general pattern to such an international standard that can serve as
criteria and in some cases the criteria used by an excellent starting point for more elaborate
one tester seem to contradict those used by standards.
another.
There are numerous city traffic tests done
In some instances, testers use multiple passes around the world and they are growing with
through the data to build different sets of new entries like the Shanghai City Traffic Test.
results to assess the possible impact of the The value of these tests and their continued
squeal. One approach is to compare the use is debatable. In the past when there were
averaged spectrum to the peak-hold spectrum no measurement standards, they were the only
to select the truly outstanding peaks. accepted tool. It is anticipated that this will
change.
A number of investigators are beginning to use
more sophisticated criteria such as the duration Whatever on-road test is run, it is clear that
of the squeal during a stop. Some have defined there must be a minimum of five NVH
a minimum duration for a squeal to be detected measurements to provide useful data. There
and others give added weight to a squeal of must be a measurement at each corner and
greater duration. One of the latest approaches is the consensus is that this is by use of an
to use the coherence as a rating by establishing accelerometer. There must be at least one
a minimum level of coherence and peaks interior microphone. In addition to these
above this threshold are considered significant NVH measurements, it is vitally important
squeals. to quantify the operating conditions at which
the noise occurs. To this end, it is necessary
The best assessment is that criteria for to have measurements of vehicle speed, brake
determining a significant squeal are in a temperature, line pressure, and some indication
state of transition. With better measurement of brake applications.
standards and techniques, there is the
opportunity to refine criteria and this is an The identification of brake noise with squeal
ongoing process. An important consideration being the most important is still a difficult
is correlating the criteria to vehicle occupant thing to do on the road. The process is evolving
subjective response. There have been some with more sophisticated techniques being
attempts to do this with limited success. Much incorporated as the test standards begin to be
of the ongoing criteria refinement will be in implemented.
this direction. Some have suggested including
sound quality measures to build statistical One of the coming challenges for brake noise
relationships between the character of a squeal testing is statistical significance: how does one
and the subjective response of the customer. assure the differences seen are significant or
This will most likely be the new direction for that parameters are controlled well enough to
many in this field. assure consistent tests?

108
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 6
Pad Natural Frequency
and Damping Testing

109
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

6.1 Introduction process, should not take an expert to run,

For years, the “easy” solution to all brake noise and would not be as sophisticated or as time
problems has been simple: “Just change the consuming as a full modal analysis. It was
brake pads.” Clearly the friction interface is foreseen that this test might be used as part
important to brake noise generation and in of production quality control and as part of a
many cases changing the brake pads may solve process for checking on part consistency for
the current problem. However, such simple laboratory evaluation. The goal was to develop
solutions merely solve one problem and may a test procedure that would provide natural
introduce another noise problem at a different frequency and damping values that would be
frequency or at a different set of operating accurate enough to allow the effect of these
conditions. parameters on brake noise to be analyzed for
the first time.
One of the first scientific steps was taken in
understanding the important characteristics 6.2.1 Round Robin Test
of brake pads in 2005 by the SAE Brake NVH
Since many brake component companies
Standards Committee with the development
had been measuring natural frequencies and
of the SAE J2598 Automotive Disc Brake Pad
damping for brake pads for some time, the first
Natural Frequency and Damping Test. This
step was to assess the accuracy and consistency
procedure standardizes the measurement of
of current measurement practices. A round
the first three natural frequencies and the
robin test was constructed using nine brake
corresponding damping for brake pads. While
pads from three different manufacturers. These
this may seem an obvious way to quantify
pads were as diverse as possible. They consisted
important NVH characteristics for brake pads,
of both front and read pads, large and small
it had not been standardized before. Wildly
pads, with and without shims, and all possible
different values were measured and used by
shapes. In the end, 12 companies agreed to
different people. Because of this, there was
participate in the test.
no correlation between these fundamental
parameters and noise performance.
The instructions provided to each participant
were to simply measure and document the
6.2 Standardization Process first three bending natural frequencies and
the corresponding damping values for each of
The process of standardization began with
the pads. The results obtained from this test
the U.S. Working Group on Brake NVH
program were surprising and very educational.
that became the SAE Brake NVH Standards
An example of the results for one of the brake
Committee. This group concluded that natural
pads is shown in Figure 6.1.
frequency and damping were basic parameters
that could be useful in understanding noise
The first through third natural frequencies
performance. However, it was noted that
reported seem consistent until the results from
they had to be measured correctly and in a
the fourth participant. It appears that this
consistent fashion for them to be properly
company missed the second mode and reported
utilized.
the third as the second and then reported a
much higher frequency mode as the third.
The concept was to develop a measurement
After some investigation, it became clear that
standard that could be part of a quality control
many of the testers were unable to distinguish

110
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Pad Natural Frequency and Damping Testing

Figure 6.1 First round robin natural frequency results for a single brake pad.

between bending and torsional modes, which structure. The simple structure and common
lead to a great deal of confusion in the results. materials of a brake pad suggest that typical
In the case of this example pad, errors of as values for the first few modes should be more
much as 250% in the third natural frequency like 1% or less of critical damping. It was
were seen. surprising to find that those reporting such
high levels defended these extreme values as
The damping results for the same pad are correct. In the end, errors of over 350% were
shown in Figure 6.2. not unusual in quantifying damping.

It is quite clear that there was no consistent From this round robin test, it was obvious that
measure of damping for this pad. Combining there were a number of issues to be addressed
with the confusion between torsional and with a standard procedure. It was clear that for
bending modes, there was some confusion in many companies the expected norm for pad
the measurement or calculation of damping damping was much too high. The measurement
that led to extreme variations in the results. process needed to be defined in such a way that
The reported results of over 3.5% of critical bending and torsional modes were excited and
damping were grossly too high for such a understood. In many instances, the outliers

111
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 6.2 First round robin damping results for a single brake pad.

in the data were obtained from those testers Test [10] was developed. The procedures as
using bungee cords or other resilient cords specified are applicable to brake pad modes
to suspend the pads. Unfortunately, due to between 500 and 16,000 Hz. The parameters
the geometry of brake pads these cords are measured are defined as the first three natural
commonly attached at antinodes of pad modes, frequencies and the corresponding loss
and therefore, a poor measurement of certain factors. For the purposes of this discussion,
modes is obtained. The measurement procedure the percentage of critical damping will be used
would have to specify testing on a foam pad instead of the loss factor. These two measures of
to avoid this problem. There was also concern damping are easily calculated from each other.
that some of the testers used large and heavy
accelerometers that would distort the results. The specified measurement process is shown in
Figure 6.3.
6.2.2 sae J2598
A frequency response function (FRF) is
Based on these results and the investigation of measured with an impact hammer providing
the variations seen, SAE J2598 Automotive Disc the forcing function and the input to the
Brake Pad Natural Frequency and Damping FRF. The response, which is also used in

112
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Pad Natural Frequency and Damping Testing

the FRF calculation, can be measured using The excitation defined in the procedure is to
an accelerometer or a microphone. Using be provided by a metal-tipped hammer with a
modern signal analyzers, this is an efficient force transducer. This is the preferred method.
measurement and calculation from which the Other forms of excitation may be used with
natural frequency and damping can be easily proper care and documentation that they
calculated. provide results equivalent to the preferred
method.
The procedure is divided into two parts. Part
A defines the natural frequency measurement As noted above, the response can be measured
and Part B defines the damping measurement. with either an accelerometer or a microphone.
This separation was done for two reasons. In The microphone is suggested as best in quiet
many instances, the natural frequency alone is environments since it does not distort results
a sufficient result for production plant quality with mass loading and added damping as
control. Also, the anticipated variation for these an accelerometer may. Response can also be
two measures was quite different. Different measured with a single-point laser vibrometer
tolerances on repeatability are reported for each if the user has demonstrated equivalence. If an
section. accelerometer is used, it must have less than

Figure 6.3 Basic measurement arrangement for pad damping and natural frequency test.

113
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

1 gram (0.035 ounce) mass and frequency torsional modes are excited. An example of the
response from 500 to 16,000 Hz within ±3 dB. photographs used in the procedure to define
the required locations is shown in Figure 6.5.
The brake pad to be tested must be supported on
a foam pad. The foam to be used is specified to It is critical that response and excitation are at
prevent errors in damping measurements. The the extreme ends of the pad. For bending mode
final test arrangement is shown in Figure 6.4. excitation, the response and excitation should
be aligned with the centerline of the pad. For
Note that a hammer is used for excitation and excitation and measurement of torsional modes
an accelerometer (shown) is being used for the excitation and response points are at the
response. However, the test apparatus is shown lower and upper extremes of the width at the
in an enclosure suggesting that a microphone ends of the pads.
could be used for the response measurement.
The brake pad is shown with the friction In developing the test procedure, a number of
material down on the foam pad and the impact investigations were conduced to understand
and response measured on the backing plate. the effects of test arrangement and procedures.
This is the orientation recommended in the One area of concern was the foam pad used
procedure. to isolate the brake pad. The foam originally
selected appeared to work fine, but when a
J2598 also requires at least three combinations foam with a much lower loss factor was found,
of excitation and response measurement it was evaluated. Figure 6.6 shows the results
locations to be sure that both bending and from the evaluation done.

Figure 6.4 Instrumentation configuration for pad test.

114
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Pad Natural Frequency and Damping Testing

Figure 6.5 Excerpt from SAE J2598 description of excitation and response measurement locations.

The red line represents the results from a series analyzer is to be configured. As noted
of tests using the originally selected foam. previously, the frequency range of interest
Through a series of test repetitions, the percent is from 500 to 16,000 Hz. It is important to
critical damping for a bare steel plate hovered set the pretrigger delay such that the entire
around 0.15%. input force pulse is captured. Depending on
the analyzer, this may involve adjusting the
The blue line represents a series of tests timing or a combination of timing and voltage
with the alternative foam. It is clear that the slope or level. On the response channel,
damping dropped significantly. The alternative the full response must be captured. There
foam clearly contributes less to the measured should be no windowing on the response
damping. As a check, a few tests were repeated channel. Some investigators have traditionally
with the original foam. These are shown used an exponential decay window on the
with the green line. Clearly the difference in response channel. While this may be useful
damping was due to the foam. This alternative in some studies, for this measurement
foam is what is specified in SAE J2598. such windowing has been shown to lead to
errors and should not be used. The required
An important component of the test maximum frequency resolution is 5 Hz. While
procedure is the description of how the fi ner resolution can be used, studies done in

115
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 6.6 Effect of foam on damping measurement.

developing the procedure indicated that 5 Hz The implementation of SAE J2598 has made
works quite well. significant improvements in the accuracy and
repeatability of brake pad natural frequency
It is highly recommended that the analyzer and damping measurements. It has provided
be configured to show both the FRF and the a common standard that allows the results
coherence between the input and response from different vendors to be compared reliably.
channels on the analyzer screen. This allows Figure 6.8 shows a comparison of the natural
one to see that adequate resonances have frequency measurements for one mode of all
been excited and that there is signal quality nine pads from the first round robin to a test
where the resonances occur. Figure 6.7 shows according to SAE J2598.
an example of the analyzer screen for such a
measurement. The mean values measured do not vary much
between the two measurements. However,
In this case, a resonance at 2500 Hz has been the variation as indicated by the thin vertical
highlighted with a damping of 0.5% of critical lines has decreased markedly with the
damping. The coherence at this resonance is implementation of J2598. In many cases the
100% indicating a high-quality signal. variation is hard to see with the new procedure.

116
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Pad Natural Frequency and Damping Testing

Figure 6.7 Analyzer display for proper SAE J2598 measurement.

Figure 6.9 shows the same comparison for procedure was followed carefully in conducting
measures of damping. the test.

In this case, SAE J2598 caused a significant Table 6.1 shows the results of the study using a
reduction in the measured damping and a manual hammer for excitation.
major reduction in variation. In at least one
instance, the measured damping using J2598 is The Gage R&R for the natural frequency
less than half that measured in the first round is good for both the microphone and
robin. Clearly major improvements have been accelerometer response measurements at less
made. than 10%. On the other hand, the damping
Gage R&R is 89.7 % for the accelerometer
The final step in the development of this measurement and 89.1% for the microphone
standard was to quantify the repeatability measurement. While the percentage seems
of measurements using this new standard. very high, the fact is that this represents a
Therefore, a Gage R&R study was performed magnitude of variation only 0.07% and 0.05%
using 10 nominally identical brake pads of critical damping. These are extremely small
with three testers. In this test, both a values and within the generally accepted
microphone and an accelerometer were used measurement error.
for the response measurement. The SAE J2598

117
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

Figure 6.8 Improvement in natural frequency measurement with SAE J2598.

Table 6.1
sae J2598 gage r&r results with manual hammer.
Transducer Measurement gage r&r, % gage r&r Value
accelerometer Natural Frequency 7.7 10.2 Hz.

Damping 89.7 0.07%

Microphone Natural Frequency 8.4 11.2 Hz.

Damping 89.1 0.05%

118
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Pad Natural Frequency and Damping Testing

Figure 6.9 Improvement in damping measurement with SAE J2598.

Table 6.2 shows the same data using an results and more consistent damping results.
automated hammer for excitation. The variation of 17.4% in natural frequency is a
real concern. Experience with a few automated
For production line quality control testing, hammers indicated that they had difficulty
many companies would like to use an avoiding double hits; in some cases introduced
automated hammer. However as shown in problem responses in the hammer itself, and
Table 6.2, the automated hammer tested were much slower than manual excitation. In
produced less consistent natural frequency the cases studied, the automated hammers took

Table 6.2
sae J2598 gage r&r results with automated hammer.
Transducer Measurement gage r&r, % gage r&r Value

Accelerometer Natural Frequency 17.4 24.3 Hz.

Damping 46.7 0.04%

119
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

as much time to perform one impact as all five measurement repeatability. The standard SAE
impacts could be done manually. J2598 provides an excellent recommended
practice to achieve accurate and repeatable
In the final analysis, when using SAE J2598 results.
with an accelerometer to measure response,
one should anticipate natural frequency The next step is to determine if the data
repeatability of ±5.1 Hz or 0.178% of the obtained from using this standard leads to
nominal value. Damping will be repeatable a better understanding of the importance of
to ±0.03% of critical damping or 4.695% of natural frequency and damping in the squeal
the nominal value. Using a microphone for noise characteristics of braking systems. Clearly
response measurements the natural frequency the brake pads are only one component of the
will be repeatable to ±5.6 Hz or 0.195% of the system, but if a relationship can be established
nominal value. In this case the damping will to these parameters it will be a major step
be repeatable to ±0.04% of critical damping or in understanding squeal characteristics and
6.26% of the nominal value. avoiding brake noise problems. Only with years
of experience will it be possible to see if there is
a definite relationship.
6.3 Summary
The development of a standard for pad The technology and capabilities of brake noise
damping and natural frequency measurement and vibration measurement are changing
proved to be much more difficult than anyone rapidly. Only a few years ago there was no
had imagined. In the process, a lot of lessons agreement on the simplest aspects of how to
were learned. Although many people were recreate or measure brake noise. While there
convinced that this was an “easy’ measurement, has been significant progress, there is much
it proved to be much more complex than first more to come. This will be the focus of the next
imagined and very difficult to achieve good and last chapter of this text.

120
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Chapter 7
Future Trends

121
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

7.1 Introduction It is anticipated that SAE J2521 will come up

The measurement of brake NVH has changed for review and be updated in the 2011 to 2012
rapidly in the past decade with the introduction time frame. This topic of specifying tighter
of new measurement standards that have been environmental controls will be a primary
adopted around the world. These standards consideration. It is highly likely that tighter
have led to significant improvements in environmental specifications will be issued
brake squeal. In addition, they have served with the revised J2521 in late 2010 or early 2011.
as a platform for further improvements in A humidity specification for the optional cold
procedures and the exploration of the extremes section is very likely.
of operating conditions leading to brake noise.
One automaker already has a requirement for
It is anticipated that much of the coming the ambient air in the shaft dynamometer test
change in brake squeal measurements will chamber to be 5oC (41oF) and over 80% relative
be evolutionary based on refinements to humidity. This condition is difficult to maintain
SAE J2521. On the other hand, revolutionary and generally requires a cooling system with
changes are anticipated in the measurement of larger capacity and redundant cooling cools.
other brake noises and in the incorporation of The problem is that at such conditions the
vehicle characteristics. cooling cools are running near or below
freezing temperatures. There is a rapid buildup
of ice on the cools and they become ineffective.
7.2 Environmental Conditions One approach is to use redundant cools so that
There will most likely be two changes in SAE the air flow can be switched from one coil to
J2521 that will probably be applied to other the other to allow the ice to melt on one while
brake noise measurement procedures as they the other is being used. Another approach is
are developed. Currently SAE J2521 has an to dry the air before the coil and add humidity
optional cold section where the temperature after the cooling cool. This avoids the need for
of cooling air -10 oC (14 oC) and the brake redundant coils, but is very expensive in terms
temperature of 0 oC (32oF) are specified. of equipment cost and energy usage. Whichever
The next step will be to specify humidity approach is used, it results in an expensive and
of the cooling air and perhaps the ambient complex environmental control system. Such
humidity at the brake. Common noise systems can account for nearly 50% of the cost
problems like “morning sickness” are felt to of a brake dynamometer.
be due to the combination of low temperature
and high humidity. Experience with SAE There is clearly a tradeoff between the value
J2521 has shown that many of the cooling of better humidity control and the cost and
air systems in use reach the -10 oC (14 oC) complexity of dynamometers. Specifying
requirement, but achieve this requirement humidity could require the installation
with low humidity. While a relative humidity of new environmental control systems or
of more than 80% would be desired, the significant upgrades to current systems. With
actual humidity is often less than 50%. There new environmental control systems for shaft
is a concern that such a low humidity may dynamometers costs as much as $250,000, this
not provide the necessary conditions for the is a major issue for brake NVH test laboratories
noise to be generated. around the world.

122
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Future Trends

It is also likely that a tighter specification will and humidity. It is feared that this variation
be placed on all test sections in terms of the may lead to differing results from season to
ambient air temperature and humidity. In season or between laboratories located in
the current version of J2521, these parameters different parts of the world. It seems very likely
are very loosely defined. Because of concerns that specifications will be implemented that
about moisture absorption by friction material require conditioned air within some specified
and data that suggests there is a seasonal cycle limits of temperature and humidity to be used
of brake noise, there are several investigators for brake cooling throughout J2521 to avoid
who believe it is necessary to specify the this problem. Again, this will be an added
temperature and humidity at which parts are operational cost and perhaps require new
stored before a test and the ambient conditions equipment to be added to some brake NVH
during a test. Where this debate will lead in dynamometers. Therefore, this is not a minor
terms of actual requirements is difficult to consideration.
determine right now.

There are two concerns in this debate. First is 7.3 Low-Frequency Brake Noise Tests
the storage of brake parts. It is not uncommon One of the current topics of intense interest in
for parts to be stored in unconditioned spaces brake noise testing is the development of a low-
until a short time before the start of testing. frequency test procedure. As note previously,
Depending on the location of the laboratory, the dominance of squeal noise has declined.
this can mean that the parts may have This trend has resulted in more concern
experienced a wide range of temperatures and about low-frequency noise issues and there is
humidity before the start of testing. As a tester no internationally recognized standard test
in Brazil once noted, they had tested several procedure. The SAE Brake NVH Standards
samples for a customer in Mexico and found no Committee is working on this issue and there
noise issues. The parts went to the customer in could be a standard issued in late 2010 or early
Mexico and immediately there were issues with 2011. This procedure would be a major addition
excessive noise. After some investigation, it to brake NVH test suites for brake system
was clear one of the major issues was the parts suppliers around the world.
were taken from an unconditioned space in the
laboratory in Brazil. Since it was summer, the The nature of the procedure will focus on
parts had seen temperatures in excess of 30oC low speeds in the range of 0-10 km/h (0-6.2
(86oF) for many days and relative humidity mph). Control of temperature and humidity
consistently above 80%. The temperatures in will be mandatory in this test procedure.
Mexico were not much different; however, the Early experiments suggest that water spray on
humidity was more like 40% or one-half the the brake parts may be a required part of the
value in Brazil. Testing with parts conditioned procedure. Since the focus will be on moan and
to low humidity quickly revealed some of the groan issues, completely different measurement
same noise problems seen in Mexico. practices, analyzer setting, and data collection
processes will most likely be specified. High-
The same concern is raised during the regular frequency sampling will not be required.
test process when many companies use outside However, the important noise components may
air for brake cooling. This unconditioned air not be such pure peaks that stand out from the
may again be at a wide range of temperatures background as with squeals. More advanced

123
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

signal processing and more elaborate data that one to three per year will be added over the
extraction may be required. It is too early to next five years globally.
describe the details of this test procedure, but
it clearly will be much different from J2521. It The alternative to using a chassis dynamometer
will require new practices and measurements. is to be able to simulate what the vehicle
Whether it will necessitate significant changes interior noise environment will be. Ultimately,
to existing test apparatus is not clear. this is the more efficient approach since it
relies on existing shaft dynamometers and
computer models. The limitation in such an
7.4 Representing Vehicle Effects approach is the computer model of the vehicle.
One of the topics receiving increased attention At the higher squeal frequencies—above
is the effect of the vehicle on the occupants’ 10,000 Hz—current models are generally
perception of brake noise. In many instances, only good approximations because there are
the same or very similar brake hardware is only a few noise sources of concern at such
applied to different vehicles or vehicles with frequencies. At the lower frequencies, sound
different levels of sound control trim. The is transmitted structurally and such models
result may be that the same brake is noisy are very complex. In addition, such structural
on one vehicle and not on the other. This models are not highly accurate without
concern points to the need to perform tests experimental verification. So, a vehicle will
on actual vehicles on chassis dynamometers have to be available to build an accurate model.
or mathematically apply filtering to noise Fortunately, these are the same sort of models
measured at the brake to represent what the built for understanding the impact of tire-
vehicle interior sound will be like. pavement and powertrain noise. There are good
tools available and a lot of work has already
Some investigators already use vehicles on been done in terms of refining such models. For
chassis dynamometers to evaluate brake noise. this approach to become a useful alternative,
There are currently dynamometers equipped much work will have to be done to understand
to do this. However, this is only occasionally critical brake noise transmission paths and
done now. It is likely that the frequency of this their effects on interior noise.
practice will increase and there appears to be
a trend to have more chassis dynamometers to
be able to do this sort of testing. In the course 7.5 Judder Measurement
of the next five years it is anticipated that there Judder has become a significant warranty and
will be increased installations and dedications NVH concern for most vehicle manufacturers
of chassis dynamometers for brake noise in the U.S. and Japan. While some progress
testing. has been made on defining acceptable levels of
DTV and some proprietary procedures have
Building chassis dynamometers for brake noise been developed for measuring off-brake DTV,
testing is a major investment. Considering that there is still much to be done to quantify the
current trends suggest that these dynamometer effects of DTV and customer reaction to the
chambers will have to be environmentally resulting judder.
conditioned, the prices could exceed $3,000,000
each. Therefore, such dynamometers will not Instrumentation is in place to measure DTV
become commonplace, but it is highly likely and BTV. In both cases, there is a need to use

124
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Future Trends

order analysis to understand the sources of a vehicle judder sensitivity measurement. At

the rotor distortion or torque variation. Many present, no comprehensive approach has been
experimenters have been reluctant to define a standardized. This is a difficult measurement
standard measurement procedure since there to standardize since one must measure vehicle
is little correlation between the levels of DTV transmission paths. This is a sophisticated
and the amount of judder from one vehicle to measurement requiring considerable
another. Vehicle sensitivity is a critical factor engineering knowledge. In addition, the
in understanding judder. Two vehicles with the paths involved suspension, steering, and
same amount of DTV or BTV may experience structural systems of the vehicle, which means
entirely different amounts of judder based on each vehicle measurement is unique. The
steering wheel oscillation, pedal pulsation, or development and publication of a standard
seat-track vibration measurements. In addition, vehicle sensitivity measurement will take some
the subjective reaction to the judder may differ time to develop and some more time to get an
greatly between vehicles based on the different industry consensus behind the methods used.
levels of vibration of different driver interfaces. However, it is not unreasonable to believe that
While steering wheel oscillation is generally by the 2013 or 2014 time period mentioned
considered the most important, different above for a complete DTV measurement
combinations of vibrations from all the standard, there could be a vehicle sensitivity
components can change the driver’s perception measurement standard.
of the magnitude of the judder.
The remaining challenge will be to correlate
Therefore, defining a measurement standard these objective measurements with subjective
is a difficult task. It is anticipated that the first response. This is a topic of growing importance
attempts at judder measurement standards will in the brake NVH field. When all important
begin in the next one to two years. The first brake noises were squeals and all squeals were
draft international standards should appear by pure tone and highly annoying, it was simple
2012. The first measurement procedures will to correlate the objective measurements to
focus on standardizing how DTV and BTV are subjective response. However, as other types
measured. These first standards will avoid the of brake noise become important and lower
whole issue of defining a test matrix that leads level sounds that do not stand out from the
to DTV. However, once a DTV measurement background are studied objective to subjective
has been standardized, sets of DTV generation correlation becomes much more difficult.
matrices will come shortly. An international
standard containing both an off-brake cold
DTV generation matrix and a hot DTV 7.6 Subjective Correlation
generation matrix can be expected by 2013 or As noted above, this is one of the next great
2014. challenges in brake NVH. How does one
understand how the vehicle occupant will react
The other aspect of judder that must be to judder on vehicle A versus B? An important
addressed is vehicle sensitivity. For any judder- part of this assessment is being able to obtain
related measurement to be successful there objective measures of driver interface vibration
must be a way to relate the measured DTV or such as the steering wheel oscillation, pedal
BTV to the response on a particular vehicle. pulsation, and seat-track vibration. However, a
There have been some attempts at defining further step is required. Sound quality metrics

125
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Brake NVH: Testing and Measurements

must be developed to isolate the objectionable will be to hold constant high levels of humidity
components that generate unfavorable at temperatures just above 0oC (32oF). This
customer response. is a difficult requirement that necessitates
special designs to avoid ice buildup from
Sound quality metrics have been used for a causing deteriorating performance. The
long period of time to understand how people required design modifications generally lead
will react to sounds. There have been many to greater costs and possibly greater energy
applications in architectural acoustics and usage when in operation. Cost increases for
consumer product work. As noted above, sound these adaptations are between 50 and 100% or
quality is easy to apply with brake squeals. If current environmental system costs. This is
the squeal is above the background level it will definitely something that should be planned
be objectionable. However, with judder, groan, for in considering buying new brake noise
and even lower level squeals, the application of dynamometers.
sound quality becomes much more difficult.
The other major equipment trend is the
A major new trend in brake NVH will be the increased usage of chassis dynamometers.
growth of sound quality analysis to assess While there is no mass replacement of shaft
the subjective response to brake noise. This dynamometers with chassis dynamometers
is a direct result of the near elimination of foreseen, there will be the need to use chassis
the major squeal noises that have plagued the dynamometers in some instances to recreate
industry for so many years. The new, more some of the low-frequency noise issues such
subtle noise problems will require sound as groan and judder. In addition, some
quality type analysis to assess their importance. automakers are using chassis dynamometers
As one brake engineer recently noted, “every to allow them to better determine whether a
brake noise is now important—as the squeal brake noise will present a significant problem
problems have disappeared every noise issue on a particular vehicle. Because there are
is getting the same attention. We need to significant differences between the sound
find some way to make a judgment as to how packages among different vehicles using
serious a noise may be.” The challenge will the same brake, the only way to check the
have to be met with sound quality metrics like differences between vehicles is on the road
loudness, sharpness, and measures of time- or on the chassis dynamometer. Several
based parameters. It is likely that unique sound companies have made the decision that
quality metrics will be developed for particular the controlled conditions or year-round
kinds of brake noise. There is a need for a lot testing capabilities presented by the chassis
of new work on identifying and understanding dynamometer are of value, and they have built
the annoyance of brakes noises other than or are building such facilities for brake testing.
squeal.
There is considerable expense in building a
four-wheel chassis dynamometer. This cost
7.7 New Equipment has led many to use a two-roll dynamometer
As noted above, it is highly likely that the and do both fixture and single-axle testing.
anticipated new test standards will require This reduces the costs significantly, but does
updated or new environmental systems. The not provide as good a simulation as a four-roll
critical requirements for these new systems dynamometer.

126
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Future Trends

Finally, a major consideration in building a especially at low-temperature extremes.

chassis dynamometer is the environmental These new standards could be in place as
control system. An environmental control early as 2010.
system that conditions the entire hemianechoic
space can cost a million dollars or more for There also seems to be a movement to greater
a four-roll chassis dynamometer. Some have use of chassis dynamometers for brake noise
utilized spot coolers blowing cool on each evaluations. This trend is driven by both
brake corner as an alternative. However, the the changing nature of noise problems and
option of building a chassis dynamometer for the need to make finer distinctions between
brake noise testing without environmental vehicles.
controls that can take the temperature to or
below freezing is not viable with current test The most important thing to remember is
procedures and trends. Thus building a four- that brake noise measurement is not yet a
roll chassis dynamometer for brake noise mature field. There is still much development
testing is an expensive undertaking that cannot and refinement to be done. Much more
be taken lightly. sophisticated tests, measurements, and analyses
will be introduced in the next few years. It
must be remembered that the first brake noise
7.8 Summary measurement standard was only introduced in
It is clear that there has been a reduction in 2003. There is much more work to be done.
brake squeal noise issues. This change will
lead to new measurement standards focused Finally, the reduction in dominance of brake
on lower frequency noises such as groan squeal issues will lead to greater usage of sound
and judder. The measurement and analysis quality and similar tools to assess brake noise.
requirements for dealing with these sorts of Unlike the past, one will not be able to say any
issues will be much different. brake noise is a squeal and all squeals are bad.
There will be different degrees of brake noise,
At the same time, new standards are going to and sometimes it will not be obvious whether
require better environmental control systems a measured noise or vibration will be annoying
to maintain both temperature and humidity to vehicle occupants.

127
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

References
1. Thompson, William T. 1981. Theory of Vibration with Applications.
Englewoods Cliff, New Jersey: Prentice-Hall, Inc.
2. Nashif, Ahid D., D. I. G. Jones, and J. P. Hedderson. 1985. Vibration
Damping, New York: John Wiley & Sons.
3. Oppenheim, Alan V. and R. W. Schafer. 1975. Digital Signal Processing,
Englewood Cliffs, New Jersey: Prentice-Hall, Inc.
4. SAE J2786 Automotive Brake Noise and Vibration Standard Nomenclature.
October 2009. Warrendale, Pennsylvania: Society of Automotive Engineers.
5. IEC 61094-1, Measurement microphones—Part 1: Specifications for
laboratory standard microphones. 2000. Geneva, Switzerland: International
Electrotechnical Commission 3.
6. Bendat, Julius D. and Allan G. Piersol, 1971. Random Data: Analysis and
Measurement Procedures, New York: John Wiley & Sons.
7. Thompson, J. K., A. Marks, and D. Rhode. 2002. Inertia Simulation in Brake
Dynamometer Testing. Paper No. SAE 2002-01-2601, pp 151-154, Proceedings
of 20th Annual Brake Colloquium, October 2002, P-379. Warrendale,
Pennsylvania: Society of Automotive Engineers.
8. SAE J2521 Disc and Drum Brake Dynamometer Squeal Noise Matrix, June
2005. Warrendale, Pennsylvania: Society of Automotive Engineers.
9. SAE J2625 Automotive Vehicle Brake Squeal Test Recommended Practice,
November 2007. Warrendale, Pennsylvania: Society of Automotive
Engineers.
10. SAE J2598 Automotive Disc Brake Pad Natural Frequency and Damping Test.
June 2005. Warrendale, Pennsylvania: Society of Automotive Engineers.

129
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Index
Acceleration, 3, 5 Displacement, 4, 5
and velocity and displacement, 5f Dynamometers, 57–60, 81f
Accelerometers, 40, 41–42 chassis, 59f, 60f, 83f, 84f, 124, 126–
charge type, 42–43 127
integrated electronics piezoelectric examples of, 78–85
(IEPE), 43 and inertia simulation, 63f
for road tests, 89 shaft versus chassis, 84–85
Acoustics, 13–19, 56 special considerations for, 66–67
environments, 19–21 typical enclosure for, 64f
Amplitude, and frequency, 5
Apply squeal, 32
Averaging, 44, 45f Eigenfrequency, 6
Electric motor, 58, 60–61
Encoder module, 105
Brake noises, frequency range, of 29f
Brake NVH, 28–30
Brake torque variation (BTV), 35, 125–126 Fade and recovery, 71
Fast Fourier transform (FFT), 22, 24
Free-field environment, 19, 20
Charge-type accelerometer, 42–43 Frequency
Chirp, 32 and amplitude, 5
City traffic tests, 47f, 95–101 natural, 6
Coherence, 49–50, 51f Frequency analysis, 21–25
Coherent output power (COP), 47–53 Frequency correction, 18f
Cold backward/forward, 70 Frequency data, typical, 48f
Cold judder, 35–36 Frequency domain, moving from time
Cold matrix B, 70 domain to, 21f
Constant percentage bandwidth (CPB), 24 Frequency range, of common brake noises,
Continuous non-stationary signals, 26 29f
Creep-groan tests, 66–67 Frequency response function (FRF), 50–
Cross spectrum, 48–49 51, 113–114
Frequency spectrum, 8, 22f, 23f
Frequency weighting, 18
Damping, 3, 8 Friction characterization test, 69, 70
one-degree-of-freedom system with, Future trends, 122–127
8f
pad, 110–120
Data acquisition and analysis, 40f, 43–47, Green brake noise, 68
90f, 103–108 Groan, 33–34
Deceleration module, 70
Decibel scale, 17, 19
Deterministic signals, 25 Hanning window function, 43
Diffuse-field environment, 20 Harshness, definition of, 2
Disc thickness variation (DTV), 35–36, Hemianechoic environment, 21, 56
125–126 Hot judder, 35

131
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Index

Hum, 34 Non-stationary signals, 25–26

Human perception of sound, 19t NVH, basic terminology of, 2–3
Humidity, 56–57, 122–123

Order tracking, 24–25

Inertia, simulation of, 61–63 Oscillation, 3
Inertia discs, 58 Overlap sampling, 43
Initial brake temperature (IBT) matrix,
92, 93–94
Initial conditioning, 67–68 Pads, natural frequency and damping
Instrumentation testing of, 110–120
future, 126–127 Parking lot maneuver, 92–93, 94
for road tests, 88–91 Peak hold average, 45
Integrated electronics piezoelectric (IEPE) Pinch-out squeal, 32
accelerometer, 43 Pure tone, 30
Interior noise, 30, 51
Intermediate conditioning and warm-up,
70 Random signals, 26
Road testing
establishing criteria in, 107–108
Judder, 35–37, 124–125 goals for, 88
instrumentation for, 88–91, 102f
measurement systems for, 103–108
Los Angeles City Traffic (LACT) test, procedures for, 91–103
95–101 Rotor bending mode shape, 32f
daily summary report, 98f, 99f, 100f Rotor resonant frequencies, 31f
instrumentation for, 97f
traffic route, 96f
Low-frequency tests, 123–124 SAE Brake NVH Standards Committee,
Low-speed drag, 66 110
SAE J2521, Disc and Drum Brake
Dynamometer Squeal Noise Matrix,
Mass, changes in, effect of, 7f 67–78, 79f, 80f, 81f
Mass and spring system, 4f cumulative noise occurrence plot,
Mechanical inertia, 61, 62 77f
Microphone, 40–41 microphone placement
directivity pattern of, 42f specification, 71f
Mils, 6 pressure apply ramp specification,
Moan, 33–34 69f
frequency spectrum of, 35f standard report, 74f, 76f
time waveform of, 34f test matrix summary, 68t
Models, future, 124 test results, 78f, 79f, 80f, 81f
SAE J2598, Automotive Disc Brake Pad
Natural Frequency and Damping Test,
Noise 110, 112–120
definition of, 2 analyzer display, 117f
interior, 30, 51 excitation and response
vehicle, sources of, 3f measurement locations, 115f

132
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

Index

improvement in damping Test fixtures, 73t

measurement, 119f Tests
improvement in natural frequency environmental conditions of,
measurement, 118f 122–123
results with automated hammer, city traffic, 95–101
119t creep-groan, 66–67
results with manual hammer, 118t friction characterization, 69, 70
SAE J2625, Automotive Vehicle Brake laboratory requirements for, 56–57
Squeal Test Recommended Practice, Los Angeles City Traffic (LACT),
92–95 95–101
example results report, 94f low-frequency, 123–124
test matrix summary, 93t noise measurement standards,
Signal averaging, 44 67–78
Signals SAE J2521, 67–78, 79f, 80f, 81f
non-stationary, 25–26 SAE J2598, 110, 112–120
stationary, 25–26 SAE J2625, 92–95
Sound Thermocouple module, 105
binaural measurements, 89, 90f Time domain, moving to frequency
definition of, 2 domain, 21f
human perception of, 19t Time signal, real-world example, 23f
ideal sound fields, 20f Torque, order plot for, 49f
measurement parameters of, 14f Transducers, 40–43
quality metrics for, 125–126 Transient signals, 26
terminology of, 14f
types of signals, 25–26
as vibration in air, 13f Vehicle bus interface module, 105–106
Sound power, 13, 15 Velocity, 3–4, 5
versus pressure, 16f and acceleration and displacement,
Sound pressure, 13, 15 5f
computing levels of, 17f Vibration
range of amplitudes of, 17f amplitude of, 5, 6t
of spectrum, 52f definition of, 2, 3
versus power, 16f different sources of, 12f
Source, path, and receiver model, 10, 26, forced, 11–12, 26
30 frequency of, 5
Spectral analysis, 9–12 multi-degree-of-freedom system, 9f
Squeal, 29, 30–33 potential sources of, 11f
example of, 33f, 46f real-world system, 10f
problem resolution, 101–103 resonant, 11, 26
Stationary signals, 25–26 single-degree-of-freedom system,
Stiffness, 4 6–9
Stop length, 106–107
Strain gage module, 105
Winter parking lot maneuver, 92
Wire brush noise, 32
Temperature, 56–57
Test enclosures, 63–66, 78–80
variations in, 82f

133
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

About the Author

Dr. James Thompson has more than 35
years of experience in noise, vibration, and
harshness (NVH), finite-element analysis,
experimental mechanics, and testing. He was
the executive director of sales, marketing,
and NVH at Link Engineering Co. for nearly
10 years. In this capacity, he played a major
role in the development of brake NVH
dynamometers and vehicle test systems. He
also has worked as a consultant on several
brake NVH issues and has lead testing
programs successful in resolving such issues
for major OEMs, Tier 1, and Tier 2 suppliers.

Currently, Dr. Thompson chairs the SAE Brake NVH Standards Committee. In
addition, he is the current chairman of the SAE Noise and Vibration Conference.
This committee was a pioneer in developing the first brake NVH test standard
and has an outstanding record of developing innovative test and technology
standards for the brake industry.

Dr. Thompson also developed the SAE Brake NVH course, which he has taught
for the last five years in cooperation with SAE and Link Engineering. Over this
time period, more than 200 students have attended the course. In addition, he
has organized the NVH sessions at the SAE Brake Colloquium for the last 12
years.

Outside of his brake-specific activity, Dr. Thompson is also the president of the
Institute of Noise Control Engineering in the USA. Dr. Thompson was elected
to SAE Fellow Grade in 2006. In 2009, he received the SAE Lloyd L. Withrow
Distinguished Speaker Award. He has more than 30 publications and regularly
organizes and chairs NVH-related conference sessions.

Dr. Thompson received his PhD in mechanical engineering from Purdue

University in 1979 and his BS and MS in mechanical engineering both from
Virginia Tech in 1974. Both his master’s thesis and PhD dissertations were on the
topics of acoustics and noise control.

135
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019
Downloaded from SAE International by Univ of Nottingham - Kings Meadow Campus, Monday, August 26, 2019

BrakeTesting and
NVH Measurements
James K. Thompson

As other vehicle systems have become more About the Author

refined, more attention must be placed on brake Dr. James K. Thompson
NVH issues because they can cause a negative has more than 35 years of
customer experience. From the laboratory to the experience in the fields of
road, the use of technology as well as further study NVH, finite element analysis,
by engineers is helping to lessen noise, judder, and experimental mechanics and
vibration in cars. testing. He was the executive
director of sales, marketing, and NVH at
This book provides readers with a fundamental Link Engineering Company and served as a
understanding of current practices for measuring consultant on brake NVH issues to OEMs
and testing brake NVH. From coverage of basic and Tier 1 and Tier 2 suppliers. Dr. Thompson
definitions and concepts to in-depth analysis chairs the SAE International Brake NVH
of on-road testing procedures, it will serve as a Standards Committee and the SAE Noise
comprehensive reference guide for brake test and Vibration Conference. He developed
technicians, test engineers, lab managers, and SAE's brake NVH course, which he has
others who work on making brakes quieter, taught for the past five years.
smoother, more refined, and more reliable.
Readers will learn how to test for brake noise,
what tools to use, and which recent standards and
practices have led to the successful measurement
of brake noise and vibration.

R-399

ISBN 978-0-7680-3480-6

9780768 034806