Motional Feedback On A Loudspeaker Woofer

Department of Precision and Microsystems Engineering
Control of Voicecoil transducers

R. Valk

Report no : MSD 2013.030
Coach : Prof.ir. Robert-H. Munnig Schmidt
Professor : Prof.ir. Robert-H. Munnig Schmidt
Specialisation : Mechatronic System Design
Type of report : Master of Science Thesis
Date : 12 November 2013

2
R. Valk Master of Science Thesis
Design and implementation of a Motional Feedback
Loudspeaker Woofer
R. Valk
M
a
s
t
e
r
o
f
S
c
i
e
n
c
e
T
h
e
s
i
s
Design and implementation of a Motional Feedback Loudspeaker
Woofer
Master of Science Thesis
For the degree of Master of Science in Mechanical Engineering at Delft
University of Technology
R. Valk
November ,
Faculty of Mechanical, Maritime and Materials Engineering (3mE) Delft University of
Technology
The work in this thesis was supported by the department Precision Microsystems Engineering.
Their cooperation is hereby gratefully acknowledged.
Copyright c
All rights reserved.
Abstract
This master thesis describes the dynamical behaviour of a loudspeaker woofer and the mod-
elling of it. By the use of the knowledge gained from the model, a feedback controller is
implemented in order to enhance the woofers performance in terms of acoustic total har-
monic distortion (THD) and low frequency bandwidth. The rst objective of the research is
to build a woofer set-up that suppresses the THD below 1% during operation. The second
objective is to evaluate diaphragm break-up and compensate for this break-up by the use of
feedback.
The woofer is equipped with an accelerometer sensor. By the use of feedback, the motion of
the voice-coil of the loudspeaker is controlled. This enhanced motion results into an acoustic
enhancement.
Since the motion of only a single point on the woofer diaphragm is measured and used for
feedback, only the local distortion is reduced. Even when feedback is applied, the surround of
the woofer is radiating acoustic distortion. Initially a woofer is used that has some very specic
characteristics in terms of acoustic radiation. The contribution of distortion by the surround
is large with the chosen woofer. Therefore, the increase in performance measured on the
accelerometer is not identical to the increase in performance measured by the microphone.
Where the accelerometer is mounted, the motion of the diaphragm is improved, but the
acoustic distortion radiated by the surround is not reduced.
After analysing the rst woofer, the knowledge gained from the experiments is used to de-
termine a better candidate for the experiments. The experiments are repeated on a second
set-up.
Two important observations have led to the nal result. One observation is that the surround
of the woofer is a large contributor of the acoustic radiation. Choosing a woofer that is very
uniform in terms of the distortion prole across the diaphragm, is advised. In that situation,
when the motion of the location where the sensor is mounted is improved, the same holds for
the surround of the woofer. Secondly, it is observed that when using a piezoelectric sensor, the
sensor output is not only determined by the acceleration. Stress leading to deformation of the
sensor is measured too. When the contribution of measured deformation becomes dominant
over the measured acceleration, the magnitude of the signal is that of the deformation instead
Master of Science Thesis R. Valk
ii
of the acceleration. For low frequencies, this leads to a limitation in terms of potential loop
gain.
To some extend, both problems have been solved by the use of a dierent type of sensor mount
and by a careful selection of the woofer. The second woofer used, has an acoustic distortion
prole that is nearly identical across the membrane. This observation indicates that for the
operating bandwidth of the woofer, the motion of the diaphragm is a close match to that of a
rigid piston. Increasing the performance in terms of the motion of the centre, therefore leads
to an increase in performance throughout the entire diaphragm. In order to solve the problem
of the measured deformation of the sensor, a dierent type of sensor mount is designed. This
sensor mount reduces the deformation of the sensor. This leads to a steeper roll-o slope in
the sub-resonant frequency band of the woofer. This steeper slope makes it possible to design
a controller, that leads to higher distortion suppression.
The achieved reduction of harmonic distortion measured on the accelerometer sensor is up
to 22 dB, a factor 12,5. The acoustic reduction of harmonic distortion measured with a
microphone is up to 22 dB too. In practice, the frequency band of this high suppression is
narrow. The frequency band in which the suppression is over 17 dB, a factor 7, is between
40 Hz and 150 Hz
When the woofer excursion becomes large, the THD without feedback can be over 12,5%.
Suppressing the distortion by a factor 12,5 therefore does not lead to the target THD of under
1%. The diaphragm break-up is evaluated, but a controller that suppresses this motion is not
implemented. A controller would only compensate for the break-up eect on the diaphragm
location where the sensor is mounted. The actual break-up in the diaphragm would remain.
Table of Contents
Preface vii
Acknowledgements ix
List of Symbols and Abbreviations xi
1 Introduction 1
1-1 The loudspeaker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1-2 Voice-coil transducer dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1-3 Thiele/Small parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1-4 Feedback Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1-5 Diaphragm breakup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1-6 Research objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2 Loudspeaker dynamics 7
2-1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2-2 Forces that act upon the diaphragm . . . . . . . . . . . . . . . . . . . . . . . . 8
2-2-1 Lorentz actuator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2-2-2 Suspension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2-3 Mass-Spring-Damper system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2-4 Woofer selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 System model identication 13
3-1 Test set-up initial testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3-2 Laser measurement on the movement of the woofer . . . . . . . . . . . . . . . . 14
3-3 Sensor selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3-3-1 Microphone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3-3-2 Laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3-3-3 Accelerometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3-3-4 Sensor choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3-4 Data accelerometer output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
iv Table of Contents
4 Controller design 23
4-1 Linear feedback control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4-1-1 Controller topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4-1-2 Stability Consideration . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4-1-3 Practical limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4-2 Controller objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4-3 Controller design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4-4 Closing the loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5 Controller implementation 31
5-1 Digital implementation, dSpace environment . . . . . . . . . . . . . . . . . . . . 31
5-1-1 Why use a DSP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5-1-2 Drawbacks of using a digital implementation . . . . . . . . . . . . . . . . 31
5-2 Sample-rate selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5-3 Measuring the disturbance rejection . . . . . . . . . . . . . . . . . . . . . . . . 32
6 Acoustic measurements 35
6-1 Initial measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
6-2 Interpretation of initial measurements . . . . . . . . . . . . . . . . . . . . . . . 37
7 Enhancing the acoustic performance 41
7-1 Woofer selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7-2 Measurements of dierent locations of membrane . . . . . . . . . . . . . . . . . 41
8 Feedback setup of SEAS L26RO4Y 45
8-1 Changes in setup based on experience with Peerless woofer . . . . . . . . . . . . 45
8-1-1 Advanced sensor gain stage . . . . . . . . . . . . . . . . . . . . . . . . . 45
8-1-2 Low frequency magnitude error . . . . . . . . . . . . . . . . . . . . . . . 46
9 System Identication of SEAS L26RO4Y 49
10 Performance of SEAS L26RO4Y 51
10-1 Measurements on the accelerometer . . . . . . . . . . . . . . . . . . . . . . . . 53
10-1-1 20 Hz test frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
10-2 Measurements on the microphone . . . . . . . . . . . . . . . . . . . . . . . . . 57
10-2-1 Microphone data at 20 Hz . . . . . . . . . . . . . . . . . . . . . . . . . 58
10-3 Evaluating the measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Table of Contents v
11 Reection and recommendations 63
A Appendix 65
A-1 System identication signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
A-2 Reshaping the measured time domain data . . . . . . . . . . . . . . . . . . . . . 66
A-3 Testing FFT code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
A-4 Designing a model to the ETFE . . . . . . . . . . . . . . . . . . . . . . . . . . 67
A-5 Microphone performance test . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
A-5-1 20 Hz comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
A-5-2 30 Hz comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
A-5-3 40 Hz comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
A-5-4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
A-6 Power supply inuence on microphone measurement . . . . . . . . . . . . . . . . 72
A-7 3D printer lament Youngs modulus experiment . . . . . . . . . . . . . . . . . 72
A-8 THD of components other than the loudspeaker . . . . . . . . . . . . . . . . . . 73
A-9 Sensor Gain stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
A-10 Micro-Epsilon, LD1630 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
A-11 Peerless, SLS-P830946 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
A-12 SEAS, L26RO4Y . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
A-13 Kepco BOP-36-6M . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
A-14 Micro-Epsilon optoNCDT 1401 . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
A-15 Panasonic, WM-61A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
A-16 Polytec, PSV-400 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
A-17 National Instruments USB-6211 . . . . . . . . . . . . . . . . . . . . . . . . . . 83
A-18 Makerbot Replicator 2, 3D printer . . . . . . . . . . . . . . . . . . . . . . . . . 84
A-19 Measurement Specialities, ACH-01 . . . . . . . . . . . . . . . . . . . . . . . . . 85
A-20 dSpace, DS1103 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
A-21 Drawing sensor mount, Sensorcarrier1 . . . . . . . . . . . . . . . . . . . . . . . 89
A-22 Beyerdynamic MM1 reference microphone . . . . . . . . . . . . . . . . . . . . . 90
A-23 Philips NatLab improved MFB . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Bibliography 99
vi Table of Contents
Preface
This research is inspired by a product developed by Philips. In 1970, Philips developed the
MFB series of loudspeakers in which a woofer is equipped with a piezoelectric accelerometer.
This accelerometer is used as a sensor to measure the acceleration of the centre of a woofer.
That sensor signal is used for feedback control that makes the loudspeaker perform better by
means of Total Harmonics Distortion (THD) in the rst octaves and extend the low frequency
bandwidth.
After reading on that series of loudspeakers, interest grew large on how that control might
be enhanced by the use of modern digital electronics.
The Philips MFB series implemented feedback that reduced the harmonic distortion up to 10
dB, a factor 3,3. The Philips research facility [33], later on improved the motional feedback
control loop and reported signicant improvements over the control loop that was imple-
mented in the regular production units. This enhanced control loop however, was never
implemented in a commercially available product.
The enhanced control loop is dierent from the original circuitry in terms of the method
used in order to achieve a high loop-gain. It uses two signals that are added to the sensor
output. These two signals, depending on the frequency content of the input, can be dominant
or sub-dominant over the sensor output. This method thereby creates a low frequency and
high frequency masking eect. The frequency band in which the sensor output is dominant
over the two masking signals is the control bandwidth of the woofer. This dierent approach
to loop-shaping results in a loop gain, that is higher than the loop gain that was achieved
before. In appendix A-23, more information about this improved control loop is available.
With the implementation of modern Digital Signal Processors (DSP), the new method de-
veloped by Philips is not required in order to obtain a high loop gain. Due to the exibility
a DSP introduces, regular loop-shaping can be applied, as is done in the initial MFB series.
Since the limitations in terms of analogue circuit design are reduced by the DSP exibility,
the remainder of limitations are the dynamics of the loudspeaker woofer itself. Both achieving
a high loop gain and extending the control bandwidth, are the objectives of this research.
The research led to satisfactory results and was performed with great pleasure. Performing
research on a topic that is close to the heart, eased the perceived workload.
viii Preface
Acknowledgements
I would like to thank my supervisor prof.ir. R.H. Munnig Schmidt for his assistance during
the process of this research. Having a supervisor with similar interests and enthusiasm, was
a great stimulus and help in working on the project.
I would like to thank Philips for giving acces to the NatLab report on the application of MFB
loudspeakers.
I would like to thank the following companies for donating equipment in order to help me out
on this project; Analog Devices, Tymphany, SEAS and RMS Acoustics & Mechatronics.
Delft, University of Technology R. Valk
November 1, :o1
x Acknowledgements
List of Symbols and Abbreviations
Symbol Description
1
a
Acoustic power
d
Diaphragm speed
d
Surface area of diaphragm
1
a
Real part of acoustic radiation impedance
Air density
Acoustic frequency
c Speed of sound
1 Current ow through voice coil
1
L
Lorentz actuator force
/ Length of voice coil wire
1 Flux density through the voice coil
1
s
Spring force
r
d
Excursion of the diaphragm
)
s
Eigen-frequency of mass-spring-damper system
/
s
Spring constant of system
`
d
Moving mass of woofer
o
d
Acceleration of the diaphragm
c
s
Damping constant of system
Q Peaking factor of mass-spring-damper system
\
M
Matlab magnitude value
p
a
Voltage to current gain of current amplier
: Audio reference signal
j System output value
c Error between reference value and output value
O1
ye
(:) Open loop transfer function
C1
yr
(:) Closed loop transfer function
p
m
Gain margin
m
Phase margin
:
m
Modulus margin
xii List of Symbols and Abbreviations
Abbreviation Description
THD Total Harmonic Distortion
DSP Digital Signal Processor
MFB Motional Feedback
SPL Sound Pressure Level
MDF Medium Density Fiber
DAQ Data Acquisition system
ADC Analogue to digital converter
DAC Digital to analogue converter
Those who have not heard, cannot imagine the joy of really hearing
Greg Calbi
Chapter 1
Introduction
1-1 The loudspeaker
Loudspeakers are devices that use electrical power and turn that power into motion of a
diaphragm. This motion causes pressure changes in the air that are perceived as sound.
The transfer between these two quantities, electrical power and sound, is one that is very
complex. Within this statement lies the very existence of manufactures of loudspeakers.
Every individual manufacturer of loudspeakers or acoustic equipment, has a dierent method
on trying to approach a perfect transfer. One of the problems in the reproduction of the
recorded wave front, is the acoustic property of the environment. Some frequencies of the
propagated wave front remain identical in magnitude, while others are boosted or delayed by
standing waves and other acoustical inuences. Toole [30] performed research on this subject
and concluded that the listening room, loudspeaker position and recording highly inuence
the perceived sound. For example, Figure 1-1 shows the results from a series of measurements
performed on dierent locations in a listening room. The gure indicates the large deviations
from the average response, based on the location of the measurement. The location of the
loudspeakers, the location of the microphone and the shape and volume of the room, all
contribute to the measured sound. This issue, however, is not the focus of this research. This
research focusses on the mechanical- and control-aspect of the chain of the earlier mentioned
transformation from electrical power to sound.
1-2 Voice-coil transducer dynamics
A typical loudspeaker woofer consists of a diaphragm and a Lorentz actuator. The diaphragm
is suspended by a centring spider. A exible surround is used to connect the moving edges of
the diaphragm to the static frame of the woofer. This creates a barrier between the pressure
on the front side and back side of the diaphragm. The spider and surround ensure that the
motion of the diaphragm is kept within one axis. Next to that, the spider is often used as a
carrier for the wires that connect the moving voice-coil to the static frame of the woofer. As
2 Introduction
Figure 1-1: Acoustic measurements in a room with loudspeakers, on dierent locations. Devia-
tions in the order of 20 dB are measured. [30]
with all moving systems, the mass of the moving diaphragm and all other physical properties of
the moving parts create a dynamic system. Al-Ali [1] performed an analysis on the dynamics
of the actuator and the moving mass. Both in the mechanical domain and electrical domain,
the motion of the voice-coil and diaphragm are described with high-order transfer functions.
The surround, which connects the moving diaphragm to the static frame of the driver, can be
modelled as a non-linear spring and damper. The spider can be modelled as a similar com-
ponent as the surround. The Lorentz actuator has a dynamic behaviour with non-linearities
and frequency dependence too. This, in combination with the mass, shape and stiness of
the diaphragm, creates a combined behaviour of the transducer, that is non-linear.
1-3 Thiele/Small parameters
The most dominant parameters that describe the low frequency behaviour of a woofer, are
measured by the manufacturer of a loudspeaker driver. Due to the success of two pioneers
in this line of analysis, these parameters are known as the Thiele/Small parameters. The
advantage of using these parameters is that they are universally used and fairly accurately
describe the low frequency behaviour of a loudspeaker driver. These parameters however, can
only describe the linear behaviour of a woofer within a narrow frequency band.
1-4 Feedback Control
The Thiele/Small parameters describe the linear low frequency behaviour of a loudspeaker
driver, while in practice the linear range of this system model is limited. In order to reduce
1-4 Feedback Control 3
the inuence of non-linear components, feedback control or feedforward control can be applied
to the system. By suppressing the non-linear motion, the remaining motion is a closer match
to the reference audio signal.
Schurer [27] developed a method to suppress non-linear disturbances by designing a non-linear
model of the motor dynamics. By implementing the inverse of the non-linear model in a DSP,
the distortion was reduced by up to 12 dB, a factor 4. If a perfect non-linear model was
developed, this method would be interesting to use. However, due to ageing of components
or a change in temperature, the non-linear model no longer is a match to the actual motion
of the membrane.
Therefore, in this research, feedback control is used instead of feedforward control.
The mismatch between a desired response and the measured actual response can be described
as a disturbance or error. Disturbances on electrodynamic systems can be reduced by ap-
plying a well designed feedback control loop. A control loop measures the motion of the
loudspeaker and compares that to the reference. The error between the two values can be
used to compensate for the disturbance. This classic single input single output control system
is one of the most common types of control.
There are many elds of application of feedback control. The method of improving the
performance of a loudspeaker driver has been implemented by Philips in the 1970s and 1980s.
That system uses a woofer with a piezoelectric accelerometer as a sensor. The sensor data
is fed back to an analogue circuit, that compares it to the reference signal and implements
control by multiplying the error by a linear transfer function. In practice, the transfer function
is a circuit with a non-at magnitude and phase over frequency that changes the open loop
transfer function, as shown in Figure 1-2.
Figure 1-2: Open loop gure of controller dynamics, with a peak of 10 dB around 80 Hz, courtesy
of Philips
Besides the research team of Philips, other engineers have been exploring the possibilities
of using feedback in a loudspeaker design too. Al-Ali [1] performed both the analysis of the
dynamics of the loudspeaker and the application of feedback control, in order to obtain higher
performances of the system used. The research uses the Total Harmonics Distortion (THD)
as a performance measure of the woofer. Applying the control loop increased the performance
4 Introduction
in the low frequency regions by a maximum of 6 dB, a factor 2. Al-Ali advises to perform
additional research on the modelling and control of the woofer. Other engineers focused on
the identication methods. Wei [35] focused on the Volterra lter for parametric loudspeaker
system identication. Results show that this method can eectively and accurately predict
the sound pressures and THD within the bandwidth of the identication signal. Loutridis
performed a comparison between loudspeaker system identication methods. Yali [15, 20] too,
evaluated alternative identication methods and found that the empirical mode decomposition
(EMD) method performs well in the identication of loudspeaker dynamics. The concept of
performing feedback on loudspeaker woofers is applied by Breden [3], Chiu [5] and Dzisiewski-
Smith [8]. Breden used velocity feedback of a dual voice-coil subwoofer to extend the low
frequency capabilities of the woofer. He does not mention an increase in THD performance.
Chiu has successfully performed a feedback loop by the use of a microphone. He achieved a
maximal reduction in THD of 10 dB, a factor 3,3. Chiu advises to perform more research on
the control of the frequencies at which the diaphragm break-up starts, since that limited his
control performance. It is clear that this topic is explored widely.
1-5 Diaphragm breakup
Since the diaphragm of a loudspeaker driver is made out of a physical material, it has a
dynamic behaviour when excited. At certain frequencies, the diaphragm no longer moves as a
rigid piston, but shows break-up modes within the surface. These frequencies are considered
to fall outside of the practical bandwidth of a loudspeaker driver or, in some cases, are
ignored. Designers of loudspeakers try to stay away from this phenomenon in order to make
sure that this movement is not interfering with the acoustics. In typical feedback systems
of acoustic transducers, the non-rigid deformation of the diaphragm is not monitored by a
separate sensor.
Research is performed on the mapping of non-linearities of the loudspeaker. Zhang [29]
performed analysis on the diaphragm break-up and measured the acoustic inuence of the non-
linear behaviour of the transducer. The non-linearity is acknowledged, but the control design
is not described. Suykens [31] performed a linearisation of a typical non-linear transducer
model. The results show that under strict assumptions, the linearisation holds. Gelat [12]
analysed a method of describing the break-up of a woofer diaphragm. The dynamic stiness
method showed to have high resemblance to actual measurements. In that research, a feedback
control loop is not of interest. Ravaud [26] analysed the non-linearities of a loudspeaker-like
structure. The focus of Ravaud is on the time-varying inuence of the dynamics. Even though
diaphragm break-up is not the focus, it clearly shows the inuence on the dynamics of the
transducer.
As described above, many of the researches do not apply feedback control. A lot of insight
on the motion and non-linear behaviour of the woofer is available, but the application of that
knowledge in terms of designing a feedback loop is limited. Then again, the research that is
available on applied feedback, focuses on the electronics and achieves only a limited increase
in performance in terms of THD.
1-6 Research objective 5
1-6 Research objective
This research is about evaluating the limitations of applying feedback on a loudspeaker woofer.
The potential of applying feedback has been proven before. With the information available
in literature, there should be no problem in achieving a THD-reduction of 10 dB within the
control bandwidth of 40 Hz to 200 Hz.
One of the objectives is to get a reduction in THD of the woofer, so that it is below 1% during
operation. Another objective is to extend the control bandwidth beyond the frequency at
which diaphragm break-up occurs.
Mechanical and electrical limitations are evaluated and attempted to overcome. With every
limitation evaluated and possibly solved, only the fundamental mechanical limitations remain.
As an example of the problems occurring, Figure 1-3 shows a time domain measurement of a
sensor signal that is mounted on the voice-coil of a woofer. Ideally, the sensor would show a 20
Hz sinusoidal shape on the output. However, due to non-linearities, distortion is observed in
the form of harmonics of the fundamental frequency at 40 Hz and 80 Hz. Since the reference
value in this situation is merely the fundamental frequency, all deviations from that signal
are considered to be disturbances. The design and implementation of a feedback controller
has the goal of reducing these disturbances, while maintaining the fundamental frequency.
Figure 1-3: Signal output from an accelerometer mounted on the actuator of a loudspeaker
woofer. The deviations from the 20 Hz sine wave are clearly visible.
6 Introduction
Chapter 2
Loudspeaker dynamics
2-1 Overview
In Figure 2-1, a model of a loudspeaker woofer shows the components that determine the most
dominant dynamical behaviour. The moving part of the woofer that causes the dierence in
air pressure, is the diaphragm or cone, due to resemblance of the commonly used shape of
it. The diaphragm is forced to move by a Lorentz actuator, connected to the centre of the
diaphragm. In order to keep the actuator free from dust, in many woofers a dust cover is used
to seal the actuator from the environment. To keep the motion of the diaphragm in a straight
path and to seal the internal volume of the loudspeaker enclosure from the environment, a
suspension is required. The suspension consists of a surround, on the outer ring of the cone,
that seals the volume and suspends the outer part of the diaphragm. A spider is used to keep
the diaphragm aligned in the air gap of the Lorentz actuator. The frame is the static part of
the woofer that is connected to the loudspeaker enclosure.
Figure 2-1: Model of a loudspeaker, with the cone shaped diaphragm, spider, dust cap and
surround, courtesy of RMS Acoustic & Mechatronics
8 Loudspeaker dynamics
2-2 Forces that act upon the diaphragm
In order to evaluate the motion of the diaphragm, the forces that act upon the diaphragm
are evaluated.
2-2-1 Lorentz actuator
A Lorentz actuator transforms electric current into a force. This type of actuation is not
the only type available in loudspeakers, but for low frequency acoustic power, it is required
to move a large quantity of air. Equation (2-1) shows the acoustic power, 1
a
, expressed by
the speed of the diaphragm,
d
, diaphragm surface area,
d
, and real part of the acoustic
radiation impedance, 1
a
.
1
a
= (
d

d
)
2
1
a
(2-1)
Equation (2-2) shows, for low frequencies, the real part of the acoustic impedance, 1
a
, ex-
pressed in the air density, , the acoustic frequency, , and speed of sound in air, c.
1
a
=

2
2 c
(2-2)
Equation (2-1) and (2-2) show that for low frequencies, the diaphragm velocity and excursion
need to be large. Lorentz actuators are capable of large excursions. Therefore, the Lorentz
actuator is used in most woofers and subwoofers.
The actuator consists of two elements; the voice coil, a wire wound around the voice-coil
carrier that is attached to the diaphragm, and a magnet that causes a magnetic ux through
the voice coil. The current, 1, in the wires in the voice-coil results in a force, 1
L
, by equation
(2-3), where 1 is the ux density through the voice coil, /, is the length of wire inside the
magnetic eld and 1 is the current in the wire.
1
L
= 1 / 1 (2-3)
For now, B and / are considered to be constants.
2-2-2 Suspension
The spider, air volume in the enclosure and the surround, can be combined in the physical
model as a pair of non-linear springs and dampers in parallel. The stiness is low at very
small excursions, while increasing exponentially for larger excursions. This behaviour can be
described by a third order polynomial. Equation (2-4) indicates that the spring force, 1
s
, is
not linear to the excursion, r
d
, of the diaphragm. Since the excursion at low frequencies is
large, the eect of this non-linearity is large at these frequencies. For the damping, a similar
non-linear eect is noticed.
1
s
= /
1
r
d
+ /
2
r
2
d
+ /
3
r
3
d
(2-4)
2-3 Mass-Spring-Damper system 9
The combination of Equation (2-1) and Equation (2-2) show that for low frequencies, the
speed of the diaphragm,
d
, must be increased in order to maintain acoustic power, 1
a
.
Substituting the diaphragm speed by the excursion, r
d
, leads to Equation (2-5) that shows
the quadratic relationship between the frequency, , and excursion.
1
a
= (r
d

2

d
)
2

2 c
(2-5)
Equation (2-4) indicates that for higher values of excursion of the diaphragm, the contribution
of non-linear behaviour increases. Therefore, for low frequency acoustic power, where the
excursions need to be large, the non-linearities from the suspension are more dominant than
in the higher frequencies.
2-3 Mass-Spring-Damper system
The combination of forces described in Section 2-2, lead to a mass-spring-damper mechanical
behaviour, as shown in Figure 2-2. This is the most dominant mechanical behaviour of a
loudspeaker woofer in a sealed enclosure. This behaviour describes the linearised model from
a force to an acceleration of the mass. Since the force from the actuator is proportional to
the current through the voice-coil, the model holds for a current input, 1, too. Equation
(2-6) describes the eigen-frequency, )
s
, of the woofer motion, expressed in Hz, for a combined
stiness, /, and mass, `
d
.
)
s
=
1
2

/,`
d
(2-6)
Figure 2-2: Mass spring damper system actuated by a force from the Lorentz actuator.
2-4 Woofer selection
The rst woofer evaluated is the Peerless SLS-P830946, a 6.5 inch diameter woofer of which
the specications are detailed in Appendix A-11. Figure 2-3 shows many characteristics that
are typical for loudspeaker woofers. However, there is one distinct frequency band in which
the driver has very low Sound Pressure Level (SPL) output. Between 600 Hz and 800 Hz,
there is a 20 dB dip of SPL output. It is expected that within that frequency band, the rst
diaphragm break-up mode, where the outer part of the diaphragm is no longer following the
motion of the centre part, is causing this reduced output.
The objective of this research is not only to suppress non-linear inuences on the acoustic
output, but to compensate for diaphragm break-up too. Because the woofer SPL gure shows
all the signs of problematic diaphragm break-up, the woofer is chosen to be the test subject
of this research.
The parameters, provided by the manufacturer of the woofer, are noted in Table 2-1. In order
to make an accurate linear model of the woofer that is used for further research, the dynamic
linear behaviour of the woofer must be measured in an identication process.
Figure 2-3: SPL expressed in decibels with varying frequency excitation, as provided by the
manufacturer, Peerless. Additional specications are noted in Appendix A-11
2-4 Woofer selection 11
Table 2-1: Peerless SLS-P830946 parameters
Unit Symbol Value
DC resistance R 2,7 Ohm
Inductance L 0,65 mH
Moving mass M 30,5 gr.
Motor force factor B / 7,6
Compliance 1/k 467 jm/N
Mechanical Q factor Qms 2,81
Electrical Q factor Qes 0,80
Total Q factor Qts 0,62
Resonant frequency Fs 42 Hz
Ratio fs/Qts Fs/Qts 119
Energy bandwidth product (1/Qes) Fs 110
Motor Eciency factor 21,18
Sensitivity dB at 1 W/1m 82 dB
Eective piston area Sd 123,5 c:
2
Chapter 3
System model identication
In order to obtain an accurate linear model, the parameters of the manufacturer are merely
used as a guideline. There can be deviations on the parameters and these parameters only
provide accurate information up to the rst resonance frequency. A measurement is required
to match the model to the real system. This measurement can be performed by measuring
the movement of the diaphragm under various input signals. The process of obtaining a linear
model is required in order to design a controller later on in Chapter 4.
3-1 Test set-up initial testing
In order to get a good grasp of the dynamical behaviour of the loudspeaker, a test set-up is
built. The woofer is mounted inside an enclosure, made out of 12 mm Medium Density Fiber
(MDF) board. This enclosures inner dimensions are 220 by 220 by 220 mm. The inside of
the enclosure is tted with sheets of damping material on the surface area. This damping
material is used in order to dampen the acoustic enclosure dynamics, which are expected to
have a resonance in the 1.300 Hz to 1.500 Hz region. It is advised to dampen these dynamics
so that the inuence on the woofer dynamics is limited.
The woofer is driven by a current feedback amplier. Since it is known that the force of the
Lorentz actuator is linearly coupled to the electric current, as described in Equation (2-3),
driving an ideal current amplier leads to a scaled force on the centre of the woofer diaphragm.
When the input to a mass-spring-damper system, as shown in Figure 2-2, is a force, there
is a resonating peak at the eigen frequency, according to Equation (2-6). This resonating
peak can be used to have a high loop-gain around the eigen frequency. The other advantage
of using a current feedback amplier, is that when the electric resistance of the woofer or
cables change, this does not inuence the open-loop characteristics. For a voltage feedback
amplier, a change in electric resistance of the load changes the magnitude of the open-loop.
The current feedback amplier used is a Kepco BOP 36-6ML, as described in Appendix A-13.
This amplier can drive an inductive load, like that of a woofer, to 6 kHz. The bandwidth of
the usable amplication is well beyond the acoustic bandwidth of a typical woofer of 500 Hz.
14 System model identication
The exact specications of the amplier, driving inductive loads, are not described in detail.
Any dynamics introduced by the amplier are captured in the system identication process.
3-2 Laser measurement on the movement of the woofer
As described in Section 2-4, the woofer is expected to have diaphragm break-up between
600 Hz and 800 Hz. In order to evaluate the behaviour between 600 Hz and 800 Hz, a
measurement is performed. This measurement is performed with a laser tracking multiple
spots of the moving diaphragm. The laser set-up used is a Polytec PSV-400 Scanning Head
vibrometer, as described in Appendix A-16. This device measures the dierence in wavelength
between an outgoing and reecting laser beam. Tracking the dierence in wavelength can be
calculated into a speed of the measured surface.
The signal of interest is the acceleration of the diaphragm. From Section 2-2-1 and Equation
(3-1), it is known that the radiated power of a low frequency signal is proportional to the
speed of the diaphragm and the angular frequency.
1
a
= (
d

d
)
2

2
2 c
(3-1)
For a sinusoidal motion of the diaphragm, Equation (3-1) can be rewritten to Equation (3-3)
by substituting the speed of the diaphragm,
d
, for the acceleration of the diaphragm, o
d
, as
described in Equation (3-2).
o
d
=
d
(3-2)
1
air
= (o
d

d
)
2

2 c
(3-3)
Figure 3-1 shows the woofer in the enclosure, with thin pieces of reective tape attached to
the diaphragm. The laser head shown in Figure 3-2 can track the individual spots in order
to measure the local movement of the woofer. The measurement is performed by exciting
the woofer with dierent frequencies and magnitudes and tracking the resulting motion of
the reective tape on the woofer. The signal used to excite the woofer is generated by the
laser vibrometer software and send to the current amplier. The amplier outputs a current
proportional to the voltage level on the input.
The low frequency behaviour of the woofer is that of a typical second order high pass with
a peak at the resonance frequency and corresponding phase change from +180 degrees to 0
degrees. The amount of damping corresponds to the height of the resonance peak and the
slope of the transition between +180 degrees and 0 degrees phase. Equation (3-4) describes
the peaking factor, Q, that is the magnitude of the peak at the eigen frequency divided by
the magnitude beyond the eigen frequency. The value of Q is calculated from the stiness of
the system, /
s
, the moving mass of the diaphragm, `
d
, and the damping of the system, c
s
.
Q =
/
s
`
d
c
(3-4)
3-2 Laser measurement on the movement of the woofer 15
Figure 3-1: Woofer under test cir-
cumstances, with reective tape to en-
hance laser reectivity
Figure 3-2: Laser scanner used to
track the movement of the spots on
the woofer
The peak value at the resonance frequency is one of the advantages of using a current feedback
amplier. This peak is later on used to achieve high open-loop gain around the eigen frequency.
A voltage feedback amplier leads to a lower value for Q at the eigen frequency.
The measurement with the laser set-up conrms the 600 Hz to 800 Hz diaphragm break-up
dynamics that lead to the reduction of SPL in the specication sheet of the manufacturer.
Figure 3-3 shows that between 600 Hz and 800 Hz, the centre of the diaphragm and the
surround no longer move synchronised. Around 800 Hz, the motion of the surround area of
the diaphragm is 180 degrees out of phase to the motion of the centre of the diaphragm.
This implies an opposite motion of the centre of the diaphragm and the outer part of the
diaphragm. Since the acoustic radiation of the two surface areas is inverted, the sum of the
acoustic radiation at 800 Hz is very low. This observation was expected from the SPL peak
and dip shown in Figure 2-3. Additional measurements indicate that the break-up at 800 Hz
is axisymmetric.
Figure 3-4 and Figure 3-5 show a visualisation of a typical rst diaphragm break-up at which
the diaphragm no longer moves as a piston, but when there is movement within the surface
area.
When increasing the frequency, at around 3 kHz, a similar behaviour can be noticed. Moving
further up the frequency band indicates higher order diaphragm break-ups, possibly with
non-axisymmetric shapes. The frequencies of diaphragm break-up modes are beyond the
usable acoustic bandwidth of the woofer and control bandwidth of the controller. There is no
advantage in controlling the centre of the diaphragm, where the Lorentz actuator is mounted,
if the rest of the diaphragm is still deforming.
The measurement conrms that there are more dynamics involved than just the linear second
order mass-spring-damper system behaviour, that is described by the Thiele/Small parameters
noted in Table 2-1.
Figure 3-3: Laser measurement showing the magnitude and phase of the acceleration of the
centre and surround of the diaphragm
Figure 3-4: 2D visualisation of a typ-
ical rst diaphragm break-up mode
Figure 3-5: 3D visualisation of a typ-
ical rst diaphragm break-up mode
3-3 Sensor selection 17
3-3 Sensor selection
Even though the laser measurement is a useful tool to measure the dynamic behaviour of the
woofer, it is not designed to provide real-time control.
In order to implement real-time feedback control, real-time data must be available to the
controller. There are multiple types of sensors available for this purpose. Three sensor types
that can provide the resolution and speed are evaluated for the use of real-time feedback.
3-3-1 Microphone
A microphone capsule that can be used for feedback, is the Panasonic WM-61A. That micro-
phone capsule has a wide bandwidth, but the behaviour at low frequencies is not provided by
the manufacturer. According to the specication sheet in Appendix A-15, the low frequency
usable band is 20 Hz. Typically, manufacturers put bandwidth limitations at the frequency
of the -3 dB frequency. A microphone, by its nature, has the property of capturing not only
pressure waves nearby, but pressure waves from the environment too. This means that an
acoustic disturbance surrounding the microphone is inuencing the observed motion of the
loudspeaker.
Mounting a microphone inside the woofer enclosure is an option too. In that case, a high
pressure level microphone must be used.
3-3-2 Laser
A close range laser distance sensor can be used for tracking the position and, by double
derivation, the acceleration. Micro Epsilon, a manufacturer of laser displacement and posi-
tioning sensors has a sensor that is capable of capturing a measurement every 0.00001 second.
The LD1630, as described in Appendix A-10, with an analogue output, has a stated usable
bandwidth between 0 Hz and 100 kHz and a measurement range of up to 50 mm.
3-3-3 Accelerometer
An accelerometer is a sensor that measures the accelerations of the body in one or multiple
degrees of freedom. The Lorentz actuator can only exert force in one degree of freedom so
only one axis of acceleration measurement is required. A sensor of interest is the Measurement
Specialities ACH-01. This is an analogue sensor with a bandwidth from 2 Hz to 20 kHz, as
described in Appendix A-19. This sensor has been successfully implemented in MFB set-ups
by other engineers.
3-3-4 Sensor choice
Analysing the previously mentioned options, leads to the choice of using an accelerometer.
Using a microphone in front of the woofer diaphragm is expected to lead to an undesired
additional time delay and is sensitive to environmental acoustic changes. If the microphone
is used at a distance of 1 cm from the centre of the diaphragm, an acoustic signal of 1 kHz
would, by the corresponding delay, introduce an added phase of 14 degrees. Mounting a
microphone in front of the woofer therefore is not the chosen option. Mounting a high level
microphone inside of the enclosure is an option, but the low frequency roll-o is higher than
desired for control purposes.
Only a couple of laser sensors can be used in the wide frequency range and range of excur-
sion of the woofer. Therefore, the sensors that have the required specications are expensive
compared to the other types of sensors. However, the real diculty with using a position-
ing sensor is that the sensor output must be dierentiated twice before the acceleration is
obtained. This introduces additional noise to the measurement.
The accelerometer of choice is the Measurement Specialities ACH-01, shown in Figure 3-8
and described in Appendix A-19. The typical application gain stage, shown in Figure 3-6, is
implemented in order the match the performance of the sensor, stated by the manufacturer.
In order to attach the sensor to the moving part of the woofer, the dust cap is removed. An
aluminium mounting disc is made on the lathe and is glued on the voice coil carrier. The
mounting disc and sensor introduce extra weight to the diaphragm, but that changes the
dynamics within acceptable boundaries above the resonance frequency of the system. The
added mass lowers the eciency of the woofer. The sensor and aluminium disc weigh just
under 12 grams in total. With an original moving mass of 30,5 grams, the moving mass
of the woofer is increased by 40%. The eciency of the woofer is reduced by 40%. This
introduces no problems since the power amplier can deliver enough power to compensate
for the additional mass. The resonance frequency of the woofer is decreased by 18%. The
woofer that was not equipped with the additional mass has a resonance frequency of 57 Hz.
The resonance frequency with the additional weight will reduce to 47 Hz.
The mounting disc has a dynamic behaviour too, that is not desired to inuence the mea-
surement in the operating area of interest. Figure 3-7 shows that the rst inuence of the
mounting disc is at 6.200 Hz, well beyond the predicted crossover frequency of the controller.
The additional mass and the eect on the dynamic behaviour of the woofer, outweigh the
disadvantage of the two alternative types of sensors.
Figure 3-6: Accelerometer gain stage, as in the specication sheet, A-19
3-4 Data accelerometer output 19
Figure 3-7: Mounting disc with the
rst resonance at 6.200 Hz
Figure 3-8: Measurement Specialities
ACH-01 Accelerometer
3-4 Data accelerometer output
A Data Acquisition system (DAQ) is used in order to measure the accelerometer output signal
and drive the power amplier. The National Instruments USB 6211, as described in detail in
Appendix A-17, has two analogue outputs that can be programmed by the use of Mathworks
Matlab 2013. Two of the sixteen analogue inputs are used to capture both the identication
signal and the output from the accelerometer.
Figure 3-3 shows a system with higher order dynamics, well outside of the expected control
bandwidth. The goal of the system identication is to obtain an accurate linear model of the
dynamics of the movement of the centre of the woofer diaphragm.
In order to construct the linear model, a logarithmic sine sweep is used to drive the woofer.
The sweep excites the woofer from 30 Hz to 6.000 Hz with a xed peak magnitude, that
excites most dynamics of the system. The power amplier output is a current, scaled to the
input voltage. The signal generated in Matlab therefore is a scaled representation of the force
generated by the Lorentz actuator. Equation (3-5) indicates the ratio between the values used
in Matlab, \
M
, the gain of the power amplier, p
a
, and the motor force factor of the woofer,
1 /, to the resulting motor force, 1
L
. The ratio between the value in Matlab to a force in
Newton is \
M
4.56.
1
L
= \
M
p
a
1 / (3-5)
1
L
= \
M
4.56 (3-6)
The DAQ samples the analogue inputs at a rate of 100.000 samples per second. Both the
generated sine sweep signal and the sensor output are recorded in order to keep the data
synchronised.
With the captured data set from the sensor, an initial impression of the dynamics is acquired
by computing an Empirical Transfer Function Estimate (ETFE).
The Nyquist frequency, when sampling at 100 kHz, is 50.000 Hz, half of the sample frequency.
Since the highest frequency of excitation is 6.000 Hz, re-sampling by a factor ten removes the
estimated output at frequencies which are not excited and limits the ETFE to 5.000 Hz.
An ETFE of the dataset is shown in Figure 3-9. Analysing the ETFE plot shows a series of
distinct characteristics which can be related to a combination of mechanical phenomena.
1. 180 degrees phase up to 20 Hz
2. peaking high pass dynamic behaviour at 45 Hz.
3. 0 degrees phase between 70 Hz and 500 Hz.
4. relatively small magnitude change between 70 Hz and 500 Hz.
5. dynamic coupling and decoupling behaviour in the magnitude plot from 600 Hz to 800
Hz.
6. dynamic coupling and decoupling behaviour in the magnitude plot around 1.300 Hz.
7. steep drop in phase at the 3.000 Hz diaphragm resonance
Figure 3-9: ETFE of measurement data
The dynamics described above can be converted to series of linear transfer functions that
have a similar behaviour. Figure 3-10 shows the overlay of both the bode plot of the tted
model and the ETFE of the measured data.
3-4 Data accelerometer output 21
There are dierences between the acquired ETFE and the tted model. The low frequency
magnitude and phase are dierent. This is caused by the fact that the distinct peak at the
rst eigen frequency is only excited for a very small amount of time. Additional measurements
that focus on the low frequency behaviour, show a better match in both magnitude and phase
around the rst eigen frequency of 45 Hz.
There is a mismatch between the model and the data at 600 Hz and from 3.000 Hz. The
design of the model is performed based on rst principle models that not describe the motion
at frequencies over 3.000 Hz. The expected operational bandwidth of the woofer is 40 Hz to
300 Hz. For the purpose of control, an estimation of the magnitude and phase outside of the
control bandwidth is of interest, only in order to determine stability
Figure 3-10: ETFE measurement data and the tted model
Chapter 4
Controller design
4-1 Linear feedback control
In order to apply feedback, a controller is designed. The dynamics of this controller, in
combination with the dynamics of the loudspeaker system, determine the overall system
performance.
4-1-1 Controller topology
For this controller, a topology is chosen in the form of Figure 4-1. In this topology, : is the
reference. In practice, : is the audio signal to be tracked by the woofer. The C-block is
the implemented controller. This block, when implemented, consists of a discretized transfer
function. The P-block is the Plant or system, identied in the system identication process.
In practice, this block is the transfer from the signal send to the amplier to the measured
acceleration by the sensor. The output of the system, j, is compared to the reference, :. The
dierence between the reference and the measured output is the output error, c.
Figure 4-1: Feedback topology
The transfer function from the feedback error, c, to the output, j, is the Open Loop Transfer
Function. This transfer function consists of the multiplication of the transfer function of the
controller with the transfer function of the linear system model. From the open loop transfer
function, a lot of information regarding the closed loop performance can be acquired.
24 Controller design
4-1-2 Stability Consideration
When designing the controller, the stability of the system is one of the limitations of the system
that needs to be taken into account. Since the frequency response of both the controller and
the plant are known, or to be designed, the closed loop system can be evaluated for stability.
A classic way of determining stability of a system, is to evaluate the polynomial of the close
loop transfer function. When the poles of the polynomial all are in the left half of the complex
plane, the closed loop is stable.
The Nyquist plot is a graphical tool that enables to predict the degree of stability of a closed
loop system from the open loop response. Analysing stability from the open loop is useful
when instability, after closing the loop, can cause damage to the plant. In that case measuring
only the open loop frequency response is sucient to predict the closed loop behaviour.
Equation (4-1) is the notation of the open loop transfer function, O1
ye
(:). It is the multipli-
cation of the controller transfer function by the linear model of the plant.
O1
ye
(:) = 1(:) C(:) (4-1)
When the loop is closed, the transfer function from the input, :, to the output, j, is described
by equation (4-2). The transfer function of the closed loop is determined by the open loop
characteristics.
C1
yr
(:) =
1(:) C(:)
1 + 1(:) C(:)
(4-2)
Maintained oscillation is acquired when the output of the open loop, j, is inverted to the
input of the open loop, c. In that situation, the output is inverted again by subtraction and
becomes the new input, c, that is identical to the previous input. This situation therefore
occurs when the open loop transfer function, O1
ye
(:), at the given input frequency, is -1.
The characteristics of the open loop transfer function can be plotted in a Nyquist plot. Both
the magnitude and the phase of O1
ye
(:), are captured in the Nyquist plot. The critical
point of maintained oscillation is used as a guideline to check whether the system is stable
or unstable. The Nyquist stability theorem says that an open loop transfer function does
not have have poles in the right half complex plane, when the Nyquist plot does not encircle
the critical -1 point. In that situation the roots of the polynomial of the closed loop transfer
function are all in the left half complex plane and the system is stable.
The distance from the Nyquist curve to the critical -1 point is expressed by the gain margin,
phase margin and modulus margin and describe the degree of stability. Gain margin, p
m
, is
the gain that can be added before the -1 point is encircled. The phase margin,
m
, is the
phase that can be added before the curve is encircling the -1 point. The modulus margin,
:
m
, is the shortest distance between from the Nyquist curve to the critical -1 point.
4-1-3 Practical limitations
The practical implementation of the feedback loop is limited by the set-up. These limitations
need to be taken into account.
4-2 Controller objective 25
1. The loudspeaker cannot track a static acceleration.
2. The loudspeaker has a limited amount of excursion.
3. The amplier has a current-, bandwidth- and voltage-limitation.
4. Signals must be suciently large in order to obtain an acceptable bit-depth.
5. The digital sample-rate and latency of DSPs are limited.
4-2 Controller objective
The controller must be designed such, that within the acoustic bandwidth of the woofer, the
performance is enhanced. The woofer dimensions and characteristics indicate an acoustic
operating bandwidth between 40 Hz and 400 Hz. Below 40 Hz, a reasonable acoustic power
output requires large excursions from a woofer of this size, as described in Section 2-2-1. Due
to the limited excursion of 8 mm, it is decided not to track the lowest frequencies of the audio
spectrum. Beyond 400 Hz, the woofer no longer radiates acoustic power as spherical as in the
lower frequencies due to the ratio between the size of the diaphragm and the wavelength of
the pressure waves.
The objective of the nal controller is to obtain an output disturbance rejection that sup-
presses the THD below 1%. The other objective is to compensate for the diaphragm break-up.
In order to reach the objectives, the open loop must have a peak magnitude of over 20 dB, a
control bandwidth extended to the diaphragm break-up and reasonable stability margins.
Stability margins are considered to be reasonable under the following conditions.
1. The phase margin,
m
, is more than 20 degrees
2. The gain margin, p
m
, is more than 3 dB
4-3 Controller design
The controller is designed, keeping in mind the limitations of the loudspeaker, the hardware
and the frequency bandwidth in which the model of the loudspeaker is accurate. Between
30 Hz and 3.000 Hz, the model is a match to the linearization of the system within 2 dB
deviation.
Analysing the practical- and stability-limitations, leads to two required unity-gain crossover
frequencies. One crossover frequency is at a low frequency and one at a high frequency. The
frequency band in between the two crossover frequencies, except for the frequencies close to
these frequencies, is where the controller is suppressing disturbances.
The low frequency crossover frequency is close to 30 Hz. When the slope of the open loop
transfer function at the low frequency crossover frequency is close to 40 dB per decade, the loop
gain can be high at frequencies at which most disturbances are expected. The frequencies at
which the controller must have a high loop gain are the frequencies of the harmonic distortions
of 30 Hz to 60 Hz. At these low fundamental frequencies, the woofer needs a large excursion in
order to generate the required acoustic power. When the excursion of the diaphragm is large,
the non-linear behaviour of the woofer is a big contributor to the overall acoustic power. The
loop gain is designed such that it is over 10 dB between 60 Hz and 240 Hz, which captures
both the second and third harmonics distortion of all frequencies in between 30 Hz and 60
Hz.
Figure 4-2 shows the bode plot of the transfer function of the controller. The controller is
designed by multiplication of a series of transfer functions.
1. Two zeros at 44 Hz and two poles at 50 Hz. This creates an antiresonance at 44 Hz,
adding phase in the low frequency band to increase the low frequency phase margin.
2. A compensation for the diaphragm break-up from 600 Hz to 800 Hz. This is a lead-lag
compensator consisting of two poles and two zeros.
3. A notch lter for a resonance at 7.100 Hz, two poles and two zeros at 7.100 Hz
4. Two zeros at 60 Hz and two poles at 200 Hz, increasing the magnitude from 60 Hz to
200 Hz
5. Two poles at 90 Hz, leading to a descending slope of the magnitude
6. A zero at 300 Hz and a pole at 1.000 Hz in order to reduce the phase, so that the high
frequency margin is larger
Figure 4-2: Bode plot of controller block transfer function
Multiplying the control block transfer function by the linear model of the system, leads to
the open loop characteristics. This open loop is plotted in a Bode plot in Figure 4-3.
4-4 Closing the loop 27
Figure 4-3: Bode plot of open loop transfer function
Figure 4-4 shows the stability margins of the open loop. For the crossover frequency of 30
Hz, the phase gain,
lf
, is 20 degrees and the gain margin, p
lf
, is 3 dB. The gain margin is
smaller than the open loop model suggests. The low frequency magnitude slope of the model
is not a perfect match to the real system. The gain margin therefore is smaller, because the
magnitude below 30 Hz is larger than the magnitude of the linear model. For the crossover
frequency of 500 Hz, the phase gain,
hf
, is 25 degrees and the gain margin, p
hf
, is 8 dB.
The open loop transfer function magnitude is over 10 dB, between 50 Hz and 270 Hz. The
maximum loop gain is 23 dB at 95 Hz.
The objective of designing a controller that has an open loop gain that is over 10 dB is
accomplished, while the stability margins are reasonable. The objective to have loop gain at
the diaphragm break-up is not accomplished due to high frequency diaphragm break-up and
model uncertainties. In order to guarantee stability, the crossover frequency is reduced to 500
Hz. For the initial experiments, this controller is implemented.
4-4 Closing the loop
The closed loop transfer function is the transfer from the reference value, :, to the system
output, j, as shown in Figure 4-1 and Equation (4-3).
C1
yr
(:) =
1(:) C(:)
1 + 1(:) C(:)
(4-3)
Figure 4-4: Nyquist plot of open loop transfer function
Designing the characteristics of the closed loop transfer function as in Equation 4-2 is the
goal of the open loop shaping. Designing in the open loop is a tool in order to achieve the
desired closed loop characteristics.
The performance of the closed loop transfer function can be described by analysing the sen-
sitivity transfer function, o(:). This transfer function can be interpreted as the residual of a
disturbance, d, on the output, j, when the feedback loop is closed.
o(:) =
1
1 + C(:) 1(:)
(4-4)
The lower the value at a certain frequency, the higher the suppression of that disturbance.
Figure 4-6 shows that the disturbances with a frequency content between 50 Hz and 270 Hz
are suppressed by a minimum of 10 dB. Between 75 Hz and 130 Hz, the suppression is over 20
dB, a factor 10. One of the objectives is to implement a controller that reduces the THD to
below 1%. When a 40 Hz tone is played at loud volume and the second and third harmonics
are 80 Hz and 120 Hz, according to the Bode plot shown in Figure 4-6, these distortions are
suppressed by a factor ten. If the controller is not implemented and the THD is 10%, it is
calculated that with the implementation of the controller, the THD is reduced to 1%.
4-4 Closing the loop 29
Figure 4-5: Bode plot of the closed loop transfer function
Figure 4-6: Bode plot of the sensitivity transfer function
Chapter 5
Controller implementation
The sensitivity transfer function from Chapter 4 is only a simulated controller performance.
In order to test the actual performance, the controller is implemented and tested on the actual
set-up.
5-1 Digital implementation, dSpace environment
In order to implement the controller in the set-up, the Laplace notation of the controller
block, C(:), is rewritten in a discrete time state-space notation and implemented in a dSpace
DS1103 PPC controller board, as described in Appendix A-20. This controller board can
be programmed within the environment of Matlab and Simulink. The real-time module of
Matlab compiles the model of Simulink into a DSP format and can upload the compiled code
to the DSP of the dSpace DS1103.
5-1-1 Why use a DSP
Using a DSP with the exibility of the dSpace DS1103 brings advantages over building a
prototype with analogue electronics. Altering a controller implementation, adding signal
magnitude limitations or adding non-linear implementations of controllers is easier in the
digital domain than in the analogue domain.
5-1-2 Drawbacks of using a digital implementation
Using a DSP for feedback systems introduces new problems over analogue circuits.
1. An analogue-to-digital (ADC) conversion is required in order to provide the DSP with
digital information. These ADC have a limited bit-depth and a time delay due to
ltering expressed by latency.
32 Controller implementation
2. A digital-to-analogue (DAC) conversion is required in order to transform the DSP out-
come to a analogue signal. These DAC have a limited bit-depth and a time delay due
to ltering expressed by latency.
3. DSPs have a limited computational power that limits the amount of calculations a DSP
can perform per sample.
For this research, the dSpace DS1103 DSP is used since the limitations of the digital imple-
mentation do not outweigh the advantage of having the exibility of tuning the controller.
5-2 Sample-rate selection
The latency of the dSpace board is programmable and for this purpose small enough to
introduce only a small amount of added phase in the control bandwidth. The sample-rate is
set to 100 kHz. The measured latency is veried to be 1/100.000th of a second. For a 100
kHz sine wave, this period is as large as a full cycle, or a phase of 360 degrees. In order to
verify that the added phase due to sampling is suciently small, the added phase at 1 kHz is
calculated. The phase added by introducing the delay is 3,6 degrees at 1 kHz. This amount is
acceptable for this application. If a DSP is used that has a latency of 1/10.000th of a second,
the added phase at 1 kHz would be 36 degrees, an additional phase to be taken into account
when designing the open loop dynamics.
5-3 Measuring the disturbance rejection
The gures in chapter 4-4 are obtained from the linear model of the system and the designed
controller. In order to verify the actual performance of the controller, the controller is im-
plemented in the real set-up. Figure 5-1 shows the result from a measurement at which a
disturbance is introduced on the accelerometer by adding a signal on the amplier input.
When closing the loop, the disturbance is reduced or increased depending on the frequency
of the disturbance. The remaining fraction of the disturbance is a match to the calculated
sensitivity function, shown in Figure 4-6. At 40 Hz and around 450 Hz, the disturbance
introduced is not suppressed and is identical on the output. Within 40 Hz and 450 Hz, the
remainder of the disturbance is lowest at 90 Hz at just under 8%. This value is very close to
the predicted -23 dB, which in fraction results in a 7,1% remainder. The slope from 90 Hz to
the 0 dB suppression frequency at 40 Hz and 450 Hz are identical.
A second test is performed to verify the predicted suppression of disturbances from the woofer
motion. A sine wave is used as a reference signal. The magnitude of the 40 Hz sine is such
that the output at the accelerometer is 700 mV RMS at 40 Hz. A fast fourier transform is
used in order to measure the power at frequencies from 1 Hz to 5.000 Hz.
Figure 5-2 shows the frequency content measured from the accelerometer sensor. The funda-
mental frequency of 40 Hz is clearly visible and the second and third harmonics at 80 Hz and
120 Hz are -18 dB respectively -17 dB. The fourth and fth harmonics at 160 Hz and 200 Hz
are -40 dB respectively -35 dB. It is expected that when implementing the controller, these
harmonic distortions are reduced by the corresponding magnitude of the sensitivity function.
5-3 Measuring the disturbance rejection 33
Figure 5-1: Measured fraction on output when introducing a disturbance
The 80 Hz and 120 Hz content is introduced by the signal generator, amplier, woofer, ac-
celerometer and the ADC of the computer set-up. The contribution of the harmonic distortion
by the woofer is the dominant source of harmonic distortion, as described in Appendix A-8.
Figure 5-3 shows the result from the measurement, with the implementation of the feedback
controller. The second and third harmonics are no longer in de -20 dB scale but reduced to
-40 dB. This suppression of -20 dB is the exact value that was calculated for that bandwidth.
The fourth and fth harmonic distortions are suppressed by the calculated magnitude too
and are reduced to -55 dB and -50 dB, 15 dB lower than the original value.
34 Controller implementation
Figure 5-2: Fast fourier transform of 40 Hz fundamental frequency without feedback controller
Figure 5-3: Fast fourier transform of 40 Hz fundamental frequency with feedback controller
Chapter 6
Acoustic measurements
The accelerometer sensor conrms the increase in performance of the motion of the diaphragm
by the application of feedback. The dierence between Figure 5-3 and Figure 5-2, as described
in Section 5-3, indicates the enhanced performance of the motion of the accelerometer attached
to the voice-coil. In order to measure the acoustic increase of performance of the woofer,
measurements are performed with a microphone.
6-1 Initial measurements
In order to get an initial impression of the acoustic performance with and without the imple-
mentation of the controller, a measurement is performed. The actuated frequency is 40 Hz.
The microphone used is a Panasonic WM-61A, described in detail in Appendix A-15, that is
equipped with a gain stage and modication in order to be capable of higher sound pressure
levels.
The rst measurement is performed without feedback. This is the reference performance
of the woofer without any corrections performed by the feedback loop. Figure 6-1 shows
the fundamental frequency captured by the microphone at 40 Hz and the content of the
harmonic distortions. Due to the suppression of disturbances measured in Chapter 5, the
harmonic distortion should be reduced by 20 dB between 60 Hz to 200 Hz. Figure 6-2 shows
the measurement, with the feedback controller implemented. The peak of the fundamental
frequency is identical to the peak in Figure 6-1. The harmonics at 80 Hz and 120 Hz are, in
contrast to the expectations, not reduced by 20 dB, but by 7 dB and 10 dB, noted in Table
6-1.
36 Acoustic measurements
Figure 6-1: Acoustic measurement without feedback corrections
Figure 6-2: Acoustic measurement with feedback corrections
Table 6-1: Microphone measurement at 40 Hz
Frequency No feedback Feedback Dierence
40 Hz 81 dB 81 dB 0 dB
80 Hz 54 dB 47 dB -7 dB
120 Hz 52 dB 42 dB -10 dB
6-2 Interpretation of initial measurements 37
6-2 Interpretation of initial measurements
In contrast to the disturbance suppression measured on the accelerometer, the acoustic sup-
pression of distortion is of a smaller magnitude. The increased performance of the measured
sensor output is not a representation of the increase in acoustic performance. This observa-
tion leads to the theory that the motion and distortion measured by the accelerometer is not
fully representative of the acoustic distortion of the set-up.
From Section 3-2, it is known that the surround of the woofer diaphragm is of large inuence on
the acoustic radiation. In order to nd the source of the distortion, an acoustic measurement
is performed locally on several locations of the woofer diaphragm. These measurements
indicate that the surround of the woofer contributes to the distortion. Since the contribution
of the propagated acoustic power of the surround is, by its larger diameter, as large as that
of the centre of the woofer, the harmonic distortion is only reduced by a small amount. Since
the surround cannot be directly controlled by the Lorentz actuator, it will still inuence the
acoustics.
Figure 6-3 and Figure 6-4 show the reference measurement, without applied feedback. The
measured distortion at 80 Hz is 20 dB larger on the surround compared to the measurement
on the centre of the woofer. Figure 6-5 and Figure 6-6 show that with the implementation of
feedback, the surround is still radiating acoustic distortion. The combined acoustic radiation
of both the centre and the surround lead to the observation of Chapter 6-1.
The problem is not caused by the concept of the feedback loop, but by the woofer evaluated
in the research. The character that makes the woofer interesting as described in Chapter 2-4,
is the fact that the surround of the diaphragm does not move identical to the centre of the
diaphragm. It is believed that this character backres when a feedback loop is applied that
only enhances the performance on one specic point. Distortion introduced by the surround
is still present, even though the motion of the centre of the diaphragm is controlled.
The results from the acoustic measurements on the Peerless woofer set-up are not satisfac-
tory. The accelerometer sensor is indicating a suppression of harmonic distortion, but the
microphone is not indicating equal suppression. The objective to suppress the THD below
1% therefore cannot be accomplished with the current controller and set-up.
Figure 6-3: Acoustic measurement reference, microphone close to center of the woofer
Figure 6-4: Acoustic measurement reference, microphone close to rubber surround of the woofer
6-2 Interpretation of initial measurements 39
Figure 6-5: Acoustic measurement with feedback applied, microphone close to center of the
woofer
Figure 6-6: Acoustic measurement with feedback applied, microphone close to rubber surround
of the woofer
Chapter 7
Enhancing the acoustic performance
In order to test the theory from Chapter 6-2, a dierent woofer is measured in a similar
manner. The woofer of choice is a SEAS L26RO4Y, shown in Figure 7-1 and described in
detail in Appendix A-12.
7-1 Woofer selection
The SEAS L26RO4Y has a sti diaphragm due to the shape and materials used. The material
of the diaphragm is aluminium. This is a stier material, compared to the coated paper
diaphragm of the Peerless woofer. Secondly, the Peerless driver has a diaphragm shape that
is relatively at, compared to the deeper cone shape of the SEAS diaphragm.
These observations are backed up by the manufacturer, claiming that there are no signs of
diaphragm edge resonances and distortion.
It is expected that the acoustic radiation of the SEAS woofer is less inuenced by a-synchronised
local movement of the diaphragm. If that is the case, applied feedback control leads to a larger
improvement of acoustic performance, compared to the Peerless woofer.
7-2 Measurements of dierent locations of membrane
As is described in Chapter 6-2 for the Peerless woofer, the SEAS woofer is excited at 40 Hz
with an excursion of 5 mm one-way. Since the SEAS woofer has a larger diameter, the SPL
is higher compared to the measurements on the Peerless woofer.
Figure 7-2, Figure 7-3 and Figure 7-4 show that the distribution of distortion across the
woofer diaphragm is within a 12 dB dierence of the second harmonic distortion and 4 dB
of the third harmonic distortion. The largest deviation from the distortion prole of the
centre, is a 12 dB lower second harmonic distortion measured on the surround. The increase
in performance of the motion of the centre of the diaphragm is expected to lead to a closer
increase of acoustic performance.
42 Enhancing the acoustic performance
Figure 7-1: SEAS L26RO4Y
Figure 7-2: Acoustic measurement, SEAS woofer, close to the center of the woofer
7-2 Measurements of dierent locations of membrane 43
Figure 7-3: Acoustic measurement, SEAS woofer, between the centre and surround of the woofer
Figure 7-4: Acoustic measurement, SEAS woofer, close to the surround of the woofer
44 Enhancing the acoustic performance
Chapter 8
Feedback setup of SEAS L26RO4Y
As described in Section 7-2, the expected increase of performance of the SEAS woofer is
higher then the obtained increase in performance of the Peerless woofer, as is described in
Section 6-1. It is therefore decided to implement a feedback controller on the SEAS woofer.
8-1 Changes in setup based on experience with Peerless woofer
The Peerless woofer set-up introduced problems that have not been dealt with due to the
nature of a more troublesome issue of reduced increase of performance, as described in Chapter
6-2. These problems are related to the sensor data.
1. The noise level of the sensor output is high, due to the use of a wrong type of op-amp
in the gain stage, shown in Figure 3-6
2. At low frequencies, the magnitude of the signal from the accelerometer is larger than
the linear model of the system suggests.
8-1-1 Advanced sensor gain stage
Since the SEAS woofer is equipped with a new sensor and gain stage, the sensor gain stage
is redesigned such that the noise level of the gain stage is lower, compared to the gain stage
used earlier.
The gain stage consists of an instrumentation amplier with a pair of output op-amps so that
the output signal is a low noise balanced signal, as described in Appendix A-9
The new gain stage has a noise oor of 0.5mV RMS. The gain stage used before has a noise
oor of 2mV RMS, with identical gain settings.
46 Feedback setup of SEAS L26RO4Y
8-1-2 Low frequency magnitude error
As observed in Chapter 3, in the 1 Hz to 30 Hz frequency band, there is a mismatch between
the measured sensor output magnitude and the modelled output value. This error results in a
limited amount of open loop gain due to stability limitations at the low frequency unity gain
crossover frequency. It is believed that this error is caused by stress leading to deformation
of the sensor. The sensor is not only measuring the acceleration, but deformation by forces
on the sensor too.
In order to test this theory, a similar mount to that shown in Figure 3-7 is used as a mount for
the sensor. This reference measurement indicates a similar low frequency behaviour. Figure 8-
1 shows the steep slope of the linear model of the system and the blue lines are the ETFEs
of several measurements. At 10 Hz, there is an error in magnitude of 16 dB.
Figure 8-1: Low frequency sensor output estimated transfer function compared to model
In order to test that the linear model is accurate, the low frequency movement is measured
by a triangular laser position sensor by Micro Epsilon. This sensor, the optoNCDT1401, as
described in Appendix A-14, can track the movement of the diaphragm at low frequencies.
The ETFE indicates a nearly at response, shown in Figure 8-2. When translating the
position measurement of the laser to the acceleration, that leads to the modelled low frequency
magnitude.
It can be concluded that the sensor output from the accelerometer is not only of the accelera-
tion, but from deformation of the sensor too. From the data sheet of the sensor in Appendix
A-19, this observation is conrmed. The sensitivity due to deformation of the sensor is 0.3 g
per c. This deformation sensor output does not t the model and leads to diculties when
designing a high gain, high damping feedback controller.
An alternative sensor mount is designed and implemented. This mount, as shown Figure 8-3,
reduces the deformation of the accelerometer sensor. The table-like structure, on which the
sensor is mounted, is connected to the voice coil carrier by a structure that is less sti in the
8-1 Changes in setup based on experience with Peerless woofer 47
Figure 8-2: Low frequency laser measurement of position
lateral plane while keeping stiness in the axis of the measured acceleration. This makes sure
that the strain from the ring glued in the voice coil carrier, is transferred to the sensor table
by a smaller amount. The holes in the structure ensure that the air ow is not restricted as
well.
The physical implementation of the redesigned sensor mount, as shown in Figure 8-4, is
constructed by a rapid prototyping 3D deposition printer, as described in Appendix A-18.
Figure 8-3: Model of new sensor mount
48 Feedback setup of SEAS L26RO4Y
Figure 8-4: Physical implementation of new sensor mount
Chapter 9
System Identication of SEAS
L26RO4Y
As described in Section 3-4, the dynamics of the open loop system are required in order to
design a feedback controller. The steps performed in order to identify the dynamic behaviour
of the Peerless woofer are repeated for the SEAS woofer.
Three measurements are performed that, when combined, lead to the modelled system fre-
quency response shown in Figure 9-1.
Figure 9-1: System model of the dynamics of the SEAS woofer
50 System Identication of SEAS L26RO4Y
The ETFE from the identication dataset shows a series of characteristics that are related to
a combination of mechanical phenomena and are implemented in the linear system model.
1. A high pass lter at 1 Hz, caused by the sensor
2. A high pass peaking lter at 45 Hz
3. A small magnitude change between 70 Hz and 500 Hz.
4. A coupling and decoupling behaviour in the magnitude plot from 600 Hz to 800 Hz.
5. A resonance in the magnitude plot around 7.000 Hz.
When identifying the model of the Peerless woofer in Chapter 3-4, these characteristics were
observed too. The low frequency dynamics are a close match to the second order high pass
lter observed before in Section 3-4. For the 500 Hz to 1.000 Hz regions, a coupling and
decoupling behaviour is observed. This phenomenon is observed at a dierent frequency, but
this coupling and decoupling seems to be present in many woofers in the 500 Hz to 1.000
Hz region. The impact of this diaphragm break-up on the measured acoustic radiation is
dierent for every woofer.
In contrast to the observations in Section 8-1-2, the low frequency magnitude and phase are
a closer match to the characteristics of a second order high-pass lter. The magnitude of the
ETFE at 10 Hz is over 15 dB lower with the redesigned sensor mount. The implementation of
the new sensor mount, as shown in Figure 8-4, solved the low frequency problem by reducing
the deformation of the accelerometer.
Chapter 10
Performance of SEAS L26RO4Y
In order to test the theory that a dierent woofer is more suitable for the use of feedback,
a controller is designed and implemented for the new woofer. The process of developing the
controller is identical to the process used in Chapter 4 in terms of practical- and stability-
considerations.
The controller is developed by combining a series of transfer functions.
1. First order low frequency high pass lter in order to reduce open loop DC gain.
2. A compensation for the diaphragm break up at 600 Hz to 800 Hz. This is a lead-lag
compensator consisting of two poles and two zeros.
3. A notch lter for the resonance at 2.300 Hz, two poles and two zeros at 2.300 Hz
4. A pole at 2 Hz, in order to get additional phase
5. Two zeros at 45 Hz, to increase the loop gain between 45 and 200 Hz
6. Two poles at 200 Hz, in order to have a descending slope of magnitude
The Bode plot of the controller block in Figure 10-1 shows the transfer function of the con-
troller. Figure 10-2 shows the open loop and its potential performance, of the SEAS set-up.
Between 25 Hz and 250 Hz, the open loop gain is a minimum of 10 dB. The frequency band
between 40 Hz and 150 Hz, has a minimum gain of 20 dB.
In contrast to the Peerless woofer test frequency of 40 Hz used in Chapter 6, the SEAS woofer
driver is capable of driving lower frequencies. This is possible due to the following dierences
to the set-up used initially.
1. The woofer diaphragm surface is larger by a factor 3.
2. The maximum excursion is larger by a factor 3,5.
52 Performance of SEAS L26RO4Y
Figure 10-1: Controller transfer function for SEAS L26RO4Y
Figure 10-2: Open loop transfer function of new controller in combination with SEAS L26RO4Y
Due to the dierences mentioned above, the test frequencies used are 20 Hz, 25 Hz and 30 Hz.
At these frequencies, the measured harmonic distortion is high due to the large excursion and
corresponding non-linear behaviour. It can therefore be demonstrated that in that bandwidth,
the highest amount of performance gain can be obtained.
10-1 Measurements on the accelerometer 53
10-1 Measurements on the accelerometer
The accelerometer output is the signal used for feedback and should therefore be a close match
to the predicted performance.
The FFT is a representation of the magnitude of a specic frequency in a measured signal.
This FFT is used in order to evaluate the magnitude of distortions in the signal measured.
The measurement time is 20 seconds for all measurements.
Measured magnitudes of -70 dB and under are less reliable since these values are very close
to the noise level of both the DAQ and dSpace interface. Therefore, values close to the noise
oor of either the DAQ or dSpace interface are not included in the table.
The measurements are used to verify the predicted increase in performance on the accelerom-
eter. By comparing a reference measurement to a measurement with the implementation of
the controller, the dierence between the two measurements can be evaluated.
The magnitude of the excursion is chosen such, that both the dSpace voltage output range
and additional electronics are just below clipping voltage. In practice, this means the signal
send to the current amplier is just under 15 V peak-to-peak. With the hardware used,
this is the highest amount of excursion possible at the specic test frequency. Throughout
the measurements of the accelerometer, the input voltage range of the DAQ, described in
Appendix A-17, is equally set. Since the maximum magnitude of the specic frequency is
dierent for every measurement, the magnitude of the fundamental frequency is dierent for
dierent frequencies.
Under test circumstances, when comparing the performance with feedback to the performance
without feedback, the conditions are identical in terms of maximum excursion of the woofer
diaphragm. The accelerometer output value in both conditions is matched in peak-value.
10-1-1 20 Hz test frequency
The rst test frequency is 20 Hz. This frequency is considered to be the lowest in the audible
frequency range.
Accelerometer data at 20 Hz
The diaphragm travel is 5 mm both ways, 10 mm in total.
Table 10-1: Accelerometer measurement at 20 Hz
20 Hz 0 dB 0 dB 0 dB
40 Hz -18 dB -40 dB -22 dB
60 Hz -23 dB -43 dB -20 dB
80 Hz -48 dB - -
100 Hz -50 dB - -
Figure 10-3: FFT of 20 Hz fundamental frequency without feedback, accelerometer data
Figure 10-4: FFT of 20 Hz fundamental frequency with feedback, accelerometer data
This frequency is produced by the lowest pedal notes of a pipe organ.
The diaphragm travel is 4.5 mm both ways, so 9 mm in total.
10-1 Measurements on the accelerometer 55
50 Hz -22 dB -44 dB -22 dB
75 Hz -28 dB -50 dB -22 dB
100 Hz - - -
125 Hz -50 dB - -
30 Hz is the lowest frequency a ve string double bass can produce with the B-string.
The diaphragm travel is 6 mm both ways, so 12 mm in total.
30 Hz 10 dB 10 dB 0 dB
60 Hz -30 dB -50 dB -20 dB
90 Hz -24 dB -48 dB -24 dB
120 Hz -54 dB - -
150 Hz -48 dB -66 dB -18 dB
10-2 Measurements on the microphone 57
10-2 Measurements on the microphone
The microphone output is the signal that is used in order to evaluate the acoustic dierences
in performance. The microphone is positioned close to the diaphragm in order to ensure that
the radiated pressure waves of the loudspeaker cone are dominant over reections of the room.
The distance of the microphone to the diaphragm during all tests is 10 cm.
The FFT is a representation of the magnitude of a specic frequency in a measured signal.
This FFT is used in order to evaluate the magnitude of distortions in the signal measured.
The measurement time is 20 seconds for all measurements.
Measured magnitudes of -60 dB and under are less reliable since these values are very close
to the noise level of the microphone and DAQ. Therefore, values close to the noise oor of
either the DAQ or microphone are not included in the tables.
Depending on the magnitude of the signals measured, the input voltage-range of the DAQ is
changed. This change in input voltage range, in combination with the change in maximum
magnitude, as described in Section 10-1, leads to a variation in measured magnitude peaks.
Under test circumstances, when comparing the performance with feedback to the performance
without feedback, the conditions are identical in terms of excursion of the woofer diaphragm.
The accelerometer output value in both conditions is matched in peak-to-peak value.
The THD of the system is commonly referred to as the sum of the magnitudes of the harmonic
distortions divided by the magnitude of the fundamental frequency, as described in 10-1.
TH1 =
\
2
+ \
3
+ \
4
...\
n
\
1
100 (10-1)
10-2-1 Microphone data at 20 Hz
The measurements have a fundamental frequency of 20 Hz. Without feedback, the rst
harmonic distortion is at -12 dB under the fundamental frequency. The third harmonic
distortion is -18 dB and the fourth is -40 dB under the fundamental frequency. The frequency
content at 100 Hz is neglected due to the fact that at this frequency, the contribution of
the power supply of the microphone is more dominant than the actual measured acoustic
radiation, as described in Appendix A-6.
The THD of the reference measurement without feedback is 38.7%. With feedback, the THD
is reduced to 3.4%.
20 Hz 10 dB 10 dB 0 dB
40 Hz -2 dB -24 dB -22 dB
60 Hz -8 dB -28 dB -20 dB
80 Hz -30 dB - -
100 Hz -28 dB -28 dB 0 dB
Figure 10-9: FFT of 20 Hz fundamental frequency without feedback, microphone data
Figure 10-10: FFT of 20 Hz fundamental frequency with feedback, microphone data
distortion is -28 dB and the fth is -52 dB under the fundamental frequency. The frequency
content at 100 Hz is neglected due to the fact that at this frequency, the contribution of
the power supply of the microphone is more dominant than the actual measured acoustic
radiation, as described in Appendix A-6.
is reduced to 1.7%.
25 Hz 18 dB 18 dB 0 dB
50 Hz -4 dB -20 dB -16 dB
75 Hz -10 dB 30 dB -20 dB
100 Hz -32 dB -28 dB +4 dB
125 Hz -34 dB -44 dB -10 dB
distortion is -30 dB, the fourth is -56 dB and the fth is -54 dB under the fundamental
frequency.
is reduced to 2.1%.
60 Hz -28 dB -32 dB -4 dB
90 Hz -26 dB 46 dB -20 dB
120 Hz -52 -56 dB -4 dB
150 Hz -50 -54 dB -4 dB
10-3 Evaluating the measurements
As described in Chapter 5-3, the reduction of harmonic distortion on the accelerometer sen-
sor is a match to the predicted reduction in harmonics. Depending on the magnitude of the
distortion and the frequency, the most dominant second- and third-order harmonic distortion
magnitudes are reduced by approximately 20 dB, a factor ten. In practice, when the mea-
surements of the sensor output are compared, the reduction of harmonic distortion is indeed
20 dB, a factor ten.
In contract to Chapter 6, the performance measured by the microphone with the new set-up
is improved signicantly. The reduction of harmonic distortion measured by the microphone
is reduced up to 22 dB, but in other measurements reduced less. When analysing all mea-
surements performed by the microphone, the average gain in performance is a factor seven.
One of the objectives of this research is to suppress the THD to a level below 1%. When
evaluating the increase in performance measured by the microphone, this objective is not
accomplished. The reduction in THD, up to a factor 11, is just not enough to suppress the
THD below 1%, but still signicantly better than normal subwoofer systems.
Chapter 11
Reection and recommendations
Looking back on the the research, I am pleased with the nal results and with the knowledge
gained on this topic. The nal measurements, as described in Chapter 10, show that the
concept of MFB is one with potential when analysed properly.
The objectives set are not reached. Depending on the excursion of the woofer, the THD is not
reduced below 1% under testconditions. The improvement in terms of THD are up to a factor
11, but even with that suppression, the THD can reach 4% at a 20 Hz high level reference
signal. Tuning the controller can make the controller suppress the harmonic distortion by a
higher magnitude, but at the cost of low frequency stability. The second objective was to
control the diaphragm break-up. The controller however, was limited in control bandwidth
due to stability issues. The analysis of the diaphragm break-up remains a interesting subject,
but with the accelerometer sensor used, it was not possible to measure and control both the
centre and the surround of the diaphragm.
The process of building a test set-up from scratch is a real challenge. One can in advance
only imagine the obstacles that need to be overcome in order to get to a satisfactory result.
Noisy sensors, faulty readings or unforeseen mechanical issues all contributed to evaluation
of each and every aspect of the concept. Encountering all these diculties and xing them,
has increased the know-how on the topic of MFB.
For further research, it is advised to further evaluate the eect of strain of a piezo based
sensor on the sensor output. This eect can be explained, but in practice needs to be dealt
with in order to design high performance (sub)woofer set-ups.
For further research, it is advised to evaluate the possibility of measuring the movement of
several locations on the diaphragm. Possibly, with more advanced controllers, the diaphragm
break-up can be reduced in acoustic inuence.
For further research it is advised to re-evaluate the concept of a woofer for feedback. Com-
mercially available woofers are designed to be be used in conventional loudspeakers, not for
feedback purposes. For instance, if the Lorentz actuator is attached to the diaphragm on a
location where the rst diaphragm break-up is not monitored, the break-up does no longer
inuence the measured system dynamics that limit potential loop gain.
64 Reection and recommendations
Appendix A
Appendix
A-1 System identication signal
1 %Create signal parameters
2 Fs = 100000;
3 t_end = 20;
4 t= 0: 1/ Fs : t_end ;
5 fo=30; f1=6000;
6 u_scale = 0 . 1 5 ;
7
8 %Create Sweep
9 u=chirp ( t , fo , t_end 0. 1 , f1 , logarithmic) ;
10 u = uu_scale ;
11
12 %Remember, the last value output will remain output, even after stop(AO)
13 %command. Therefore , force the last couple values of the test signal to
0.
14 u( length ( u) ) = 0;
15
16 %Puts info from DAQ in separate variable
17 daq=daqhwinfo ;
18
19 %Define Adapter ID as AO, Analog Out and same for AI
20 AO=analogoutput ( daq . InstalledAdaptors {1} , Dev1) ; %nidaq
21 AI=analoginput ( daq . InstalledAdaptors {1} , Dev1) ; %nidaq
22
23 %Set the sampletime matching the in and outpu of the signal
24 set ( AO , SampleRate , Fs) ;
25 set ( AI , SampleRate , Fs) ;
26 AI . inputtype=SingleEnded ;
27
28 %Determine which channels are selectted
29 AOcurrent = addchannel ( AO , 1 , Current) ;
66 Appendix
30 AIcurrent = addchannel ( AI , [ 1 2] , Current) ;
31
32 set ( AI , SamplesPerTrigger , Fst_end ) ;
33
34 %Define what signal is used for output
35 putdata ( AO , u ) ;
36
37 %Start the output signal
38 display ( Starting measurement)
39 display ( Starting output)
40 start ( AI) ;
41 start ( AO) ;
42
43 %Let it run for t_end seconds
44 wait ( AO , Fst_end+1)
45
46 %Force stop the measurement
47 stop ( AI) ;
48 stop ( AO) ;
49
50 display ( Stopped signal output and measurement)
51
52 %Get the data from the measurement and put it in a data-vector
53 y = getdata ( AI) ;
A-2 Reshaping the measured time domain data
1 %Remove LF drift
2 ydetrend = detrend ( y) ;
3
4 %Resample signal so only actuated dynamics are shown
5 r_scale = 10;
6 yresample = resample ( ydetrend , 1 , r_scale ) ;
7
8 %Put data in ID Data object
9 iddat = iddata ( yresample ( : , 1 ) , yresample ( : , 2 ) , r_scale/Fs)
10
11 %Define ETFE and plot VS model
12 s1 = etfe ( iddat , 1000 , 1000)
13 h1 = bodeplot ( s1 , PLANT5 , {252 pi , 50002 pi}) ;
14 setoptions ( h1 , FreqUnits , Hz) ;
A-3 Testing FFT code
1 clear all
2 clc
3
4 %Testing FFT functionality
5
6 Fs=200;
7 t_end = 20;
A-4 Designing a model to the ETFE 67
8 t = 1/Fs : 1/ Fs : t_end ;
9
10 signal1 = 10sin ( t2pi50) ;
11 signal2 = 10cos ( t2pi10) ;
12
13 y = ( signal1+signal2 ) ;
14
15 %Plot FFT
16 N = length ( y) ;
17 xdft = fft ( y) ;
18 xdft = xdft ( 1 : N/2+1) ;
19 psdx = ( 1/( FsN) ) . abs ( xdft ) . ^2 ;
20 psdx ( 2 : end1) = 2psdx ( 2 : end1) ;
21 freq = 0: Fs/length ( y) : Fs /2;
22 plot ( freq , 10 log10 ( psdx ) ) ; grid on ;
23 title ( Periodogram Using FFT) ;
24 xlabel ( Frequency (Hz)) ; ylabel ( Power/Frequency (dB/Hz)) ;
A-4 Designing a model to the ETFE
1 %Define plant
2 s = tf ( [ 1 0 ] , [ 1 ] ) ;
3
4 %Introduce fase by sampling at 100kHz
5 DELAY = ss( exp(1e5s) ) 10^( 8/20) ;
6
7 %Resonance peak subwoofer at 44Hz
8 PLANT1 = DELAY s^2/( s^2+(442pi) /5s + (442 pi) ^2) ;
9
10 % Damping on system, showing presencence from 1 000 Hz
11 pole = 6000/( s+6000)
12
13 %600 to 800 Hz coupling and decoupling of suspension
14 PLANT2 = PLANT1 ( s^2+4450/5s+4450^2) /( s^2+4900/6s + 4900^2) pole
15
16 %HF resonance
17 PLANT3 = PLANT2 ( s^2+30000/7s+30000^2) /( s^2+23000/5s + 23000^2)
18
19 PLANT4 = PLANT3 36000^2/( s^2+36000s/14+36000^2) . . .
20 ( s^2+13152pis/50+(13152pi) ^2) /( s^2+13102pis/50+(13102pi)
^2)
21
22 %Details
23 PLANT5 = PLANT4 ( s^2+28000s/1+28000^2) /( s^2+28000s/2+28000^2)
24
25 %Plant defined
A-5 Microphone performance test
In order to verify that the measurements from the Panasonic microphone capsule used, A-15,
are correct, a reference measurement is performed with a Beyerdynamic MM1 microphone,
68 Appendix
A-22. The Beyerdynamic MM1 is designed for the purpose of measuring the performance of
loudspeaker systems. This microphone too, has an unknown distortion prole, but due to the
reputation of the microphone, it is used as a reference.
The comparison consists of measuring a acoustic signal with both microphones close together.
The microphone output is passed through the dedicated pre-ampliers and the analogue
output is measured with the DAQ, A-17. The test frequencies are 20 Hz, 30 Hz and 40
Hz. Since the microphone is used for measuring harmonic distortion, that performance is
evaluated.
The sensitivity and gain stages of the microphones are not identical during the tests. There-
fore, a change in measured magnitude is expected. The deviation from that oset is the
measurement error.
A-5-1 20 Hz comparison
The measured magnitude of the signal is within +/- 2 dB of a 4 dB gain dierence.
Figure A-1: FFT of 20 Hz fundamental frequency, Beyerdynamic
Frequency Beyerdynamic MM1 Panasonic WM-61A Dierence
20 Hz 10 dB 14 dB 4 dB
40 Hz -4 dB 1 dB 5 dB
60 Hz -10 dB -4 dB 6 dB
80 Hz -32 -27 dB 5 dB
100 Hz -31 -28 dB 3 dB
A-5 Microphone performance test 69
Figure A-2: FFT of 30 Hz fundamental frequency without feedback, microphone data
The measured magnitude of the signal is within +/- 2 dB of a 9 dB gain dierence.
30 Hz 11 dB 20 dB 9 dB
60 Hz -25 dB -14 dB 11 dB
90 Hz -25 dB -16 dB 9 dB
120 Hz -59 -52 dB 7 dB
150 Hz -45 -35 dB 10 dB
70 Appendix
Figure A-3: FFT of 30 Hz fundamental frequency, Beyerdynamic MM-1
Figure A-4: FFT of 30 Hz fundamental frequency without feedback, Panasonic WM-61A
The measured magnitude of the signal is within +/- 3 dB of a 9 dB gain dierence. Only
one measurement is outside this range. Since all other measurements are within a +/-3 dB
dierence, the 80 Hz measurement is ignored.
A-5 Microphone performance test 71
40 Hz 11 dB 20 dB 9 dB
80 Hz -42 dB -22 dB 20 dB
120 Hz -36 dB -27 dB 9 dB
160 Hz -64 -51 dB 8 dB
200 Hz -49 43 dB 6 dB
Figure A-5: FFT of 40 Hz fundamental frequency, Beyerdynamic MM1
Figure A-6: FFT of 40 Hz fundamental frequency without feedback, Panasonic WM-61A
72 Appendix
A-5-4 Conclusion
The Panasonic WM-61A, used in the report, deviates in terms of sensitivity from the Beyer-
dynamic MM1, but is within acceptable magnitudes of error.
A-6 Power supply inuence on microphone measurement
Evaluating the power spectral density gures of the microphone data used throughout the
report, lead to the conclusion that for 100 Hz, the improvement obtained by the feedback
loop is lower than expected. It is believed that due to the simplistic microphone power supply
and gain stage, the mains frequency is present in the analysed time domain data captured.
The test is performed with the exact same Matlab code as the other microphone measure-
ments. A relatively quiet room is used. Depending on the amount of gain used in the gain
stage, the 100 Hz content is around -35 dB compared to the reference. When the 20 dB
gain stage is not used, which during some measurements was required, the 100 Hz content is
lowered by the predicted amount of -20 dB to -55 dB.
This indicates that for measuring harmonic distortions of 100 Hz, the setup is not reliable
since the inuence from the power supply is contributing to the actual harmonic acoustic
distortion.
Figure A-7: FFT of microphone data in quiet room, Panasonic WM-61A
A-7 3D printer lament Youngs modulus experiment
The 3D printer used, A-18, uses a PLA lament to build the model. This material has some
unknown material properties. Due to the use of this material in the sensor mount, shown in
A-8 THD of components other than the loudspeaker 73
Figure 8-4, it is desired to know the Youngs modulus in order to evaluate the stiness of the
structure.
A 100% lled beam 3D print is used for the experiment. The beam measures 150 mm long and
has two sides of 10mm. The beam is clamped in a vise and a load is applied. By measuring
the deection of the beam, the Youngs modulus is calculated.
This experiment indicates a Youngs modulus of around 1,5 GPa. When the sensor mount is
designed, this value is used in order to model the dynamic behaviour of the mount.
A-8 THD of components other than the loudspeaker
When evaluating the performance of the woofer, it is required to know whether the distortion
measured is introduced by the ampler, signal generator or woofer. In order to evaluate the
performance in THD of the amplier and signal generator, two tests have been performed.
The signal generator is tested by making a 40 Hz sine wave and directly recording it with the
National Instruments DAQ, A-17. Figure A-8 shows that the second harmonic distortion is -
80 dB under the fundamental frequency. This is equivalent to 0.01% THD. This is a factor 100
under the target total THD with the implementation of the woofer. The THD contribution of
the signal generator to the THD in test conditions with the woofer, can therefore be neglected.
Figure A-8: FFT of signal generator with a fundamental frequency of 40 Hz, second harmonic
distortion 80 dB under fundamental frequency
The amplier THD is tested by using it to drive current through a power resistor. The
input signal is the signal generation from which the performance in terms of THD is known.
The second harmonic distortion is -76 dB under the fundamental frequency of 40 Hz. This
leads to a THD of 0.016%. Again, this is a factor 63 lower than the target THD with the
implementation of the woofer. The THD contribution in terms of THD introduced by the
amplier and signal generator can be neglected.
74 Appendix
Figure A-9: FFT of signal generator and amplier, with a fundamental frequency of 40 Hz,
second harmonic distortion 76 dB under fundamental frequency
A-9 Sensor Gain stage 75
A-9 Sensor Gain stage
20-9-2013 7:39:30 D:\DropBox\Afstuderen MFB\Eagle\Gain Stage\gainstage_smd.sch (Sheet: 1/1)
OPA1642
OPA1642
OPA2131P
OPA2131P
1
8
0
k
-24V +24V
+15V
-15V
GND
GND
+24V
-24V
G
N
D
GND
1
0
-15V
2
M
GND
2k7
4
n
7
1
0
+
1
5
V
1
k
0
r
GND
2k7
4n7
2k7
2k7
2k7
2k7
1
0
-
1
5
V
10
+15V
4n7
2
k
7
GND
2k7
4
n
7
G
N
D
2k7
4
n
7
2k7
4
n
7
GND
MOUNT-HOLE3.2 MOUNT-HOLE3.2
MOUNT-HOLE3.2 MOUNT-HOLE3.2
G
N
D
G
N
D
GND
GND
MB1S
OUT-1
OUT-2
OUT-3
IN-1
IN-2
IN-3
2
3
1
8
4
IC1B
6
5
7 2
3
1
8
4
IC2A
IC2B
6
5
7
C
1
C
2
C3
C4
C5
C6
C
7
C8
C9
R
1
IC3
GND
IN OUT
IC4
GND
OUT IN
L
1
2
N
1
2
GND
1
2
C10
TR1
1
5 6
7
9
10
C11
R
2
R
3
R4
C
1
2
R
5
R
6
R
7
R8
C13
R9
R10
R11
R12
R
1
3
R14
C14
R
1
5
R16
C
1
5
R17
C
1
6
R18
C
1
7
H1 H2
H3 H4
B2
SMDFUSE
A
B
C
D
E
F
1 2 3 4
A
B
C
D
E
F
1 2 3 4
+
P
R
I
S
1
S
2
+
ACH-01 Gain Stage
Valk
RMS Acoustics and Mechatronics
2013
76 Appendix
A-10 Micro-Epsilon, LD1630
31
Model
LD
1610-0.5
LD
1610-2
LD
1610-4
LD
1610-10
LD
1610-20
LD
1610-50
LD
1610-100
LD
1610-200
Measuring range 0.5mm 2mm 4mm 10mm 20mm 50mm 100mm 200mm
Start of measuring range 23.75mm 23mm 22mm 40mm 55mm 115mm 170mm 240mm
Linearity
1m 4m 8m 20m 40m 100m 200m 400m
0.2 % FSO
Resolution (dynamic*) 0.3m 1.3m 2.6m 6.5m 13.0m 32.5m 65m 200m
Resolution (static**) 0.02m 0.1m 0.2m 0.5m 1m 2.5m 6m 20m
Spot diameter 0.1mm 0.2mm 0.3mm 0.6mm 0.9mm 1.5mm 1.5mm 2mm
Frequency response 10kHz (-3dB)
Light source laser, wavelength 670 nm, red
Laser safety class class 2
Vibration 10 g ... 1 kHz (sensor, 20 g option)
Operation temperature 0 ... +50C
Storage temperature -20 ... +70C
* Measurement on white target with 10kHz ** Measurement on white target with 20Hz
Model
LD
1630-0.5
LD
1630-2
LD
1630-4
LD
1630-10
LD
1630-20
LD
1630-50
Measuring range 0.5mm 2mm 4mm 10mm 20mm 50mm
Start of measuring range 23.75mm 23mm 22mm 40mm 55mm 115mm
Linearity
1.5m 6m 12m 30m 60m 150m
0.3 % FSO
Resolution (dynamic*) 0.8m 3.5m 7m 17.5m 35m 50m
Resolution (static**) 0.05m 0.2m 0.4m 1m 2m 7.5m
Spot diameter 0.1mm 0.2mm 0.3mm 0.6mm 0.9mm 1.5mm
Frequency response 100kHz (-3dB)
Light source laser, wavelength 670nm, red
Laser safety class class 2
Vibration 5g ... 1kHz (sensor, 20g option)
Operation temperature 0 ... +40C
Storage temperature -30 ... +75C
* Measurement on white target with 100kHz ** Measurement on white target with 230Hz
Controller
Digital ouput Ethernet TCP /IP factory default IP 192.168.122.245 (frequency response 1 - 30kHz)
Analogue output
displacement 10V (option 0 ... 10V / 0 ... 5V); 4 ... 20mA
impedance appr. 0 Ohm (10mA max.)
tilt with 30 object inclination (axis A): appr. 0.5% (white target)
cut off frequency DC ... 10kHz / 100kHz
temperature drift 0.02 % C FSO
intensitiy 0V ... 10V
Switching outputs
MIN +24V when distance < MIN, LED yellow
OK +24V when distance > MIN and < MAX, LED green
MAX +24V when distance > MAX, LED orange
Error +24 V, LED red
Switching hysteresis appr. 0.5 % FSO
Ambient light 20,000 LUX
Life time 50,000h laser diode
Isolation voltage 200V DC, 0V
Humidity up to 90% RH
Protection class sensor: IP 64, controller: IP 40
Power supply +24V DC / 200mA (10 ... 30V)
Connector 25 pin Sub D male connector
Cable length (standard) 2m
A-11 Peerless, SLS-P830946 77
A-11 Peerless, SLS-P830946
Model Number:
Product Line: 31-Aug-09 Date:
SLS-P830946
Peerless Silver
Transducer Specification Sheet
Rev 1_1 Revision:
Product Description:
The SLS line combines high quality performance with an affordable design. This 6.5
inch 4 ohm member of the SLS family features a rigid steel basket, a paper cone, and
a ferrite magnet motor with aluminium shorting ring for improved distortion
performance.
Mechanical 2D Drawing:
Specifications:
DC Resistance Revc 2.7 5.0% Energy Bandwidth Product EBP (1/Qes)fs 110
Minimum Impedance Zmin 3.6 7.5% Moving Mass Mms g 30.5
Voice Coil Inductance Le mH 0.65 Suspension Compliance Cms um/N 467
Resonant Frequency fs Hz 42 15.0% Effective Cone Diameter D cm 12.54
Mechanical Q Factor Qms - 4.92 Effective Piston Area SD cm
2
123.5
Electrical Q Factor Qes - 0.38 Equivalent Volume Vas L 11.196
Total Q Factor Qts - 0.35 Motor Force Factor BL Tm 7.60
Ratio fs / Qts F fs / Qts 119 Motor Efficiency Factor (Tm
2
)/ 21.18
Half Space Sensitivity @ 2.83V dB@2.83V/1m dB 86.7 +/-1.0
1
Voice Coil Former Material VCfm - GSV
Sensitivity @ 1W/1m 1W/1m dB 82.0 +/-1.0
1
Voice Coil Inner Diameter VCd mm 38.44
Gap Height Gh mm 8.00
Rated Noise Power (IEC 2685 18.1) P W 75 Maximum Linear Excursion Xmax mm 8.20
Test Spectrum Bandwidth Ferrofluid Type FF N/A
Transducer Size - inch 6.5
1 - Piston Band Sensitivity Tolerance Transducer Mass - kg 2.4
Frequency and Impedance Response:
12 dB/Oct 20Hz - 1kHz
50
60
70
80
90
100
80
90
100
m
s
]
@
1
.4
1
5
V
r
m
s
2
.8
3
V
r
m
s
/1
m
On Axis 30 Deg 60 Deg Imp
0
10
20
30
40
50
60
70
80
90
100
50
60
70
80
90
100
10 100 1000 10000 100000
I
m
p
e
d
a
n
c
e
[
O
h
m
s
]
@
1
.4
1
5
V
r
m
s
S
P
L
[
d
B
]
@
2
.8
3
V
r
m
s
/1
m
Frequency [Hz]
F088-0713A
0
10
20
30
40
50
60
70
80
90
100
50
60
70
80
90
100
10 100 1000 10000 100000
I
m
p
e
d
a
n
c
e
[
O
h
m
s
]
@
1
.4
1
5
V
r
m
s
S
P
L
[
d
B
]
@
2
.8
3
V
r
m
s
/1
m
Frequency [Hz]
Tymphany HK Ltd
Address : Room 1307-8 Domnion Centre, 43-59 Queen's Rd East, Wanchai, Hong Kong
E-mail: sales@tymphany.com
78 Appendix
A-12 SEAS, L26RO4Y
10 100 1 000 10 000
100
95
90
85
80
75
70
65
60
55
50
50
40
30
20
10
0
Frequency [Hz]
Extremely stiff and rigid aluminium cone gives tremendous bass precision. The cone
and the long throw low loss rubber surround show no sign of the familiar cone edge
resonance and distortion associated with soft cones.
Lead-out wires symetrically stitched to the spider to avoid resonances. Total suspen-
sion designed to assure stability for extreme excursions.
4-layer, extremely long, high temperature voice coil wound on an glassber voice
coil former gives a high power handling capacity.
Cu-cap around the pole pieces reduce non linear and modulation distortion and
increase overload margin.

Extra large magnet system provides high efciency and low Q.
Extremely stiff and stable injection moulded metal basket keeps the critical compo-
nents in perfect alignment. Large windows in the basket both above and below the
spider reduce sound reection, air ow noise and cavity resonance to a minimum.
L26RO4Y
D1004
The frequency responses above show measured free eld sound pressure in 0, 30, and 60 degrees angle using a 28L closed box. Input
2.83 VRMS, microphone distance 0.5m, normalized to SPL 1m.The dotted line is a calculated response in innite bafe
based on the parameters given for this specic driver. The impedance is measured in free air without bafe using a 2V
sine signal.
S
P
L

[
d
B
]
I
m
p
e
d
a
n
c
e

[
o
h
m
]
Jun 2011-1 *IEC 268-5
SEAS reserves the right to change technical data
Nominal Impedance 4 Ohms Voice Coil Resistance 3.3 Ohms
Recommended Frequency Range 20 - 1000 Hz Voice Coil Inductance 3.85 mH
Short Term Power Handling * 500 W Force Factor 18 N/A
Long Term Power Handling * 250 W Free Air Resonance 24 Hz
Characteristic Sensitivity (2,83V, 1m) 85.5 dB Moving Mass 173 g
Voice Coil Diameter 56 mm Air Load Mass In IEC Bafe 4.0 g
Voice Coil Height 38 mm Suspension Compliance 0.26 mm/N
Air Gap Height 10 mm Suspension Mechanical Resistance 5.4 Ns/m
Linear Coil Travel (p-p) 28 mm Effective Piston Area 363 cm
2
Maximum Coil Travel (p-p) 56 mm VAS 46 Litres
Magnetic Gap Flux Density 1.1 T QMS 4.90
Magnet Weight 2.34 kg QES 0.28
Total Weight 10 kg QTS 0.27
RoHS compliant product www.seas.no
A-13 Kepco BOP-36-6M 79
A-13 Kepco BOP-36-6M
80 Appendix
A-14 Micro-Epsilon optoNCDT 1401
Technical data
optoNCDT 1401 uses a semiconductor laser with a
wavelength of 670 nm (visible/red). The maximum
optical output power is 1 mW. The sensor is
classified as laser class II. A warning sign is
attached to the sensor housing.
IEC - Standard
LASER RADIATION
Do not stare into the beam
CLASS 2 LASER PRODUCT
IEC 60825-1: 2001-11
P<1mW; =670nm
ILD 1401-5 ILD 1401-10 ILD 1401-20 ILD 1401-50 ILD 1401-100 ILD 1401-200
5 mm (.20 ") 10 mm (.39 ") 20 mm (.79 ") 50 mm (1.97 ") 100 mm (3.94 ") 200 mm (7.87 ")
Start of measuring range 20 mm (.79 ") 20 mm (.79 ") 30 mm (1.18 ") 45 mm (1.77 ") 50 mm (1.97 ") 60 mm(2.36 ")
22.5 mm (.89 ") 25 mm (.98 ") 40 mm (1.57 ") 70 mm (2.76 ") 100 mm (3.94 ") 160 mm (6.30 ")
End of measuring range 25 mm (.98 ") 30 mm (1.18 ") 50 mm (1.97 ") 95 mm (3.74 ") 150 mm (5.91 ") 260 mm (10.24 ")
10 m 20 m 40 m 100 m 200 m 400 m
0,6 m 1 m 2 m 5 m 20 m 40 m
3 m 5 m 10 m 25 m 100 m 200 m
SMR 110 m 110 m 210 m 800 m 1000 m 2100 m
Midrange 450 m 830 m 335 m 110 m 130 m 2100 m
EMR 830 m 1600 m 830 m 730 m 760 m 2100 m
analog
digital
FSO = Full scale output
*) Note: with averaging factor 64
All specifications apply for a diffusely reflecting matt white ceramic target
Measuring rate
Linearity
Model
Measuring range
Mid of measuring range
Resolution
EN 50081-1
EN 50082-2
11 ... 30 VDC, typ. 24 VDC / 50 mA
integral signal processor
0.2 %
static 0.01 % FSO *
class 2 IEC 60825-1 : 2001-11
dyn. at 1 kHz 0.05 % FSO
IP 67
1 kHz
semiconductor laser 1 mW, 670 nm (red) Light source
Vibration
Protection class
Spot diameter
Laser safety class
Operating temperature
15 g / 10 Hz ... 1 kHz
appr. 100 g (without cable)
0 ... 55 C (32 ... 130 F)
Weight
Storage temperature
SMR = Start of Measuring Range
Output
RS232
-20 ... 70 C (-4 ... 158 F)
4 ... 20 mA (1 ... 5 V with cable PC 1401-3/U)
Electronics
Supply voltage
Electromagnetic
compatibility (EMC)
EMR = End of Measuring Range
A-15 Panasonic, WM-61A 81
A-15 Panasonic, WM-61A
Design and specications are subject to change without notice. Ask factory for technical specications before purchase and/or use.
Whenever a doubt about safety arises from this product, please contact us immediately for technical consultation.
Microphone Cartridges
Omnidirectional Back Electret
Condenser Microphone Cartridge
Series: WM-61A
WM-61B (pin type)
I Sensitivity
I Specifications
Sensitivity 354dB (0db = 1V/pa, 1kHz)
Impedance Less than 2.2 k
Directivity Omnidirectional
Frequency 2020,000 Hz
Max. operation voltage 10V
Standard operation voltage 2V
Current consumption Max. 0.5 mA
Sensitivity reduction Within 3 dB at 1.5V
S/N ratio More than 62 dB
I Typical Frequency Response Curve
I Dimensions in mm (not to scale)
WM-61A WM-61B
3.4 1.5
6
.
0
1
.
9
Term 1
output
Term 2
ground
3.4
6
.
0
Term 1
output
Term 2
ground
+20
+10
0
10
20
30
20 50 100 200 500 1000 2000 5000 10000 20000
Frequency (Hz)
R
e
l
a
t
i
v
e

R
e
s
p
o
n
s
e

(
d
B
)
Vs = 2.0V
RL = 2.2k
35 4dB
I Features
G Small microphones for gener al use
G Back electret type designed for high resistance to
vibrations, high signal-to-noise ratio
G High sensitivity type
G Microphone with pins for flexible PCB (WM-61B type)

82 Appendix
A-16 Polytec, PSV-400
6
Vibrometer system
Computer
OFV-5000 Vibrometer Controller, equipped with analog (PSV-400-B) and additional digital velocity decoder
(M2-20: with analog displacement decoder)
PSV-I-400 Sensor Head, includes OFV-505 Vibrometer Sensor, precision scanner and color video camera
with autofocus and 72X zoom, with transportation case
PSV-E-401 Junction Box with PSV-CL-10 main cable, 10 m
Standard System Components
Accessories
Technical Data
Additional components depend upon PSV model and configuration
Model
PSV-400-B
Decoder
VD-04
PSV-400-H4 VD-08
# of ranges

3
8
Ranges mm s
-1
/V

10 ... 1000
0.2 ... 50
Full scale (p)
m/s
0.1 ... 10
0.002 ... 0.5
Decoder frequency
range
0.5 Hz ... 250 kHz
DC ... 25 kHz
Resolution
1)
m s
-1
/Hz
0.1 ... 5
< 0.01 ... 0.2
VD-09 8 5 ... 1000 0.05 ... 10 DC ... 2.5 MHz 0.01 ... 4
PSV-400-H4-S VD-03-S 3 20 2000 0.2 20 0.5 Hz ... 1.5 MHz 0.1 ... 5
VD-07-S 6 2 100 0.02 1 DC ... 350 kHz <0.05 ... 0.2
PSV-400-M2
PSV-400-M4
PSV-400-M4-S
VD-07 6 1 ... 50 0.01 ... 0.5 DC ... 350 kHz <0.02 ... 0.2
VD-09
VD-09-S
VD-07-S
8
14
5 ... 1000
10 ... 2000
0.05 ... 10
0.1 ... 20
DC ... 2.5 MHz
DC ... 2.5 MHz
0.01 ... 4
0.04 ... 8
PSV-400-M2-20
additionally
VD-05
2)
2 100/500 nom. 0.5/2.5 0.5 Hz ... 10 MHz <3
Optional
5)
DD-900 16 0.05 ... 5000 m/V 1 m ... 100 mm
6)
DC ... 2.5 MHz <0.015 ... 1500 nm
7)
DD-300
2)
1 50 nm/V 75 nm
3)
30 kHz ... 24 MHz <0.02 pm /Hz
4)
Component PSV-I-400 Sensor Head
Optics
Dimensions
[W x L x H]
190 mm x 376 mm x 163 mm
(7.5 in x 14.8 in x 6.4 in)
PSV-A-420 Geometry Scan Unit
(optional) with PSV-I-400
238 mm x 376 mm x 163 mm
(9.4 in x 14.8 in x 6.4 in)
PSV-A-410 Close-up Unit
(optional)
124 mm x 90 mm x 75 mm
(4.9 in x 3.5 in x 3.0 in)
Weight 7 kg (15.4 lbs)
Laser type HeNe laser (633 nm)
7.4 kg (16.3 lbs)
Laser diode (620 690 nm)
0.35 kg (0.77 lbs)
Laser safety class Class 2 (<1 mW visible output)
Working distance With MR lens: 0.04 m~100 m; with LR lens: 0.35 m~100 m
Sample size From few mm up to several m
Camera Color video camera, CCD 1/4, 752x582 pixels, with Auto Focus and 72X Zoom (4X digital, 18X optical)
Scan speed Up to 30 points/s (typical)
Component OFV-5000
Vibrometer Controller
PSV-E-401
Junction Box
PSV-E-408 Junction Box
(optional for H-system)
PSV-W-401 Data
Management System
Dimensions mm
[W x L x H] (in)
450 x 360 x 150
(17.7 x 14.1 x 5.9)
450 x 360 x 135
(17.7 x 14.1 x 5.3)
482 x 303 x 23
(19.0 x 11.9 x 0.9)
450 x 550 x 190
(17.7 x 21.7 x 7.5)
Weight 10 kg (22.0 lbs)
General Specifications
Power 100 VAC...240 VAC 10 %, 50/60 Hz; overall max. 800 W
Calibration Every 24 months (shorter re-calibration intervals available upon request)
9 kg (19.8 lbs) 1.5 kg (3.3 lbs) 18 kg (39.7 lbs)
Scanner High precision scan unit (scanning range 20 about X,Y);
angular resolution <0.002, angular stability <0.01/hr
Data Acquisition/Data Processing

> 152 mm
Class 2 (<1 mW output)
Environmental conditions Operating temperature: +5 C +40 C (41 F 104 F); storage temperature 10 C +65 C
(14 F 149 F); relative humidity: max. 80 %, non-condensing
PSV-400 Decoder/Performance Specifications
PSV-W-401 Data Management System
Industrial PC with Windows
XP or Visita 64, Gigabit Ethernet, data aquisition hardware

24 Wide screen monitor, DVD writer, optical mouse and keyboard
VIB-A-T02 Tripod with tip-tilt adapter
PSV-Z-051 Handset (optional for PSV-400-B)
A-17 National Instruments USB-6211 83
A-17 National Instruments USB-6211
NI USB-621x Specifications
Specifications listed below are typical at 25 C unless otherwise noted.
Analog Input
Number of channels
USB-6210/6211/6215................. 8 differential or
16 single ended
USB-6218................................... 16 differential or
32 single ended
ADC resolution............................... 16 bits
DNL ................................................ No missing codes
guaranteed
INL.................................................. Refer to the AI Absolute
Accuracy Table
Sampling rate
Maximum................................... 250 KS/s (aggregate)
Minimum.................................... 0 S/s
Timing accuracy......................... 50 ppm of sample rate
Timing resolution ....................... 50 ns
Input coupling................................. DC
Input range ...................................... 10 V, 5 V,
1 V, 0.2 V
Maximum working voltage for analog inputs
(signal + common mode) ................ 10.4 V of AI GND
CMRR (DC to 60 Hz) ..................... 100 dB
Input impedance
Device on
AI+ to AI GND...................... >10 Gin parallel
with 100 pF
AI to AI GND...................... >10 Gin parallel
with 100 pF
Device off
AI+ to AI GND...................... 1200
AI to AI GND...................... 1200
Input bias current.............................100 pA
Crosstalk (at 100 kHz)
Adjacent channels .......................75 dB
Non-adjacent channels................90 dB
Small signal bandwidth (3 dB)......450 kHz
Input FIFO size................................4,095 samples
Scan list memory.............................4,095 entries
Data transfers...................................USB Signal Stream,
programmed I/O
Overvoltage protection (AI <0..31>, AI SENSE)
Device on ....................................30 V for up to
two AI pins
Device off ...................................20 V for up to
two AI pins
Input current during
overvoltage condition......................20 mA max/AI pin
Settling Time for Multichannel Measurements
Accuracy, full scale step, all ranges
90 ppm of step (6 LSB) ..........4 s convert interval
84 Appendix
A-18 Makerbot Replicator 2, 3D printer
7
PRINTING
Print Technology: FusedFilament Fabrication
BuildVolume: 11.2Lx 6.0Wx 6.1 Hin
[28.5x 15.3x 15.5cm]
Layer Resolution High 100microns[0.0039in]
Settings: Medium 270microns[0.0106in]
Low 340microns[0.0133in]
PositioningPrecision: XY: 11 microns [0.0004in];
Z: 2.5microns [0.0001 in]
Filament Diameter: 1.75mm [0.069in]
NozzleDiameter: 0.4mm [0.015in]
SOFTWARE
SofwareBundle: MakerBot MakerWare
Bundle1.0
FileTypes: .stl, .obj, .thing
Supports: Windows (XP/7),
Linux (Ubuntu10.04+),
MacOSX(10.7/10.8)
PHYSICALDIMENSIONS
Without Spool: 19.1 x12.8x14.7in [49x32x38cm]
WithSpool: 19.1 x16.5x14.7in [49x42x38cm]
ShippingBox: 23x 21.5x 17 in[59x 55x 43cm]
Weight: 25.4lbs [11.5kg]
ShippingWeight: 32.0lbs [14.5kg]
TEMPERATURE
Ambient Operating 1532C [6090F]
Temperature:
StorageTemperature: 032C [3290F]
ELECTRICAL
ACInput: 100240V,
~
2amps, 5060Hz
Power Requirements: 24VDC@6.25amps
Connectivity: USB, SD card [included]
MECHANICAL
Chassis: Powder-coated steel
Body: PVC Panels
Build Platform: Acrylic
XYZ Bearings: Wear-resistant,
oil-infused bronze
Stepper Motors: 1.8 step angle with
1/16 micro-stepping
SPECIFICATIONS
3D Imaging
Activision
Alaska Manufacturing Extension Partnership
Amherst County Public Schools
Bainbridge Island School District
Bartlett School of Architecture
Biola University
Bloomington Public Library
Boston College
Boston University
Bowling Green State University
Brooklyn College
Brown University
BUR Bikes
CBS Network West Coast
Chang Bioscience, Inc
City University of Hong Kong
Coinstar
Columbus School for Girls
CorcoranGalleryof Art/Collegeof Art &Design
Cornell University
Deloitte Innovation
Duke University
Edelman
Electronic Arts
Finnish Institute of Occupational Health
GE Power Conversion
Georgia Institute of Technology
Gonzaga University
Google
Gulfstream Aerospace Corporation
Hardin Marine
Harkins Custom Knives
Intel Corporation
JELD-WEN
Kennedy-Matsumoto Design
Kent State University
Kitchen Concepts LLC
Libero Jewelers
Mars Space Flight Facility
MIT
NASA
NASA Glenn Research Center
NASA Marshall Space Flight Center
NASA Goddard Space Flight Center
National Federation of the Blind
Neurosciences Research Foundation
New York Hall of Science
Northrop Grumman
The New York Times Company
Proctor & Gamble
Pixil Inc.
PPG Industries
Purdue University
Qualcomm Inc.
Rochester Institute of Technology
San Francisco Art Institute
Sandia National Labs
SeattleAcademyof ArtsandScience
SIU - School of Architecture
Sony Electronics
Southwest Energy
Stanford University
Texas A&M University
The Eli Whitney Museum
The New York Public Library
The University of Chicago
UC Berkeley
US Army
US Cutter
UnitedNationsInternational School
Valley Fine Foods
Woodbury University School of Architecture
Yale University
MAKERBOT
REPLICATOR
CUSTOMER LIST
To purchase, visit MakerBot.com/Replicator2 or call +1.347.334.6800
A-19 Measurement Specialities, ACH-01 85
A-19 Measurement Specialities, ACH-01
Application Specification
Accelerometer ACH-01 114-1089
30 DEC 00 Rev F
8Copyright 1998 by MSI. All International Rights Reserved. 1 of 4
INFORMATIONAL For drawings, technical data or samples, call 610-650-1500
NOTES AND Specifications subject to change. Consult MSI for latest specifications.
DISCLAIMER
PN: 1004568-8 LOC ER
1.0 INTRODUCTION
This specification covers the application requirements of Measurement Specialties Accelerometer ACH-01. The
ACH-01 is an inexpensive, general purpose accelerometer with outstanding performance characteristics. The use of
piezoelectric polymer film in the ACH-01 provides many cost/performance advantages that allow it to be used in a
wide range of applications where the use of traditional accelerometer technology is impractical. It is specifically
designed for high volume applications which require the permanent installation of an accelerometer, such as
machine health monitoring, modal analysis, automotive sensors, appliances, and feedback control systems.
Features of the accelerometer include:
Wide Frequency Response Low Transverse Sensitivity
Excellent Phase Response Wide Temperature Range
Wide Dynamic Range Small Temperature Dependence
Low Cost 3V to 40V Supply
Ultra-Low Power Impedance Buffered Output
Very High Resonant Frequency Excellent Linearity
ACH-01
SERIAL #
ACH-01-02 with Pins ACH-01-04 with Wires
(GND) BLACK
(OUT) WHITE
(V+) RED
(OUT)
(GND)
(V+)
.025 SQUARE PINS
.10 TYP.
CENTER LINE
.050
#26 AWG PVC INS.
LEADS (3) 12.0" LONG
SERIAL #
ACH-01-03 with Shielded Cable
SERIAL #
(V+) 3
(GND) 1
(OUT) 2 WHITE (SOURCE)
RED (DRAIN)
SHIELD
.740
CONNECTOR:
AMP P/N 104257-2
Connector Detail
.080 TYP
ACH-01
.244
.515
.700
1.000
.515
.244
40.0
Note: All dimensions in inches and are reference.
86 Appendix
Accelerometer ACH-01 114-1089
Rev F Measurement Specialties Incorporated, Norristown, PA 19403 610-650-1500 3 of 4
Printed in U.S.A. For drawings, technical data or samples, call 610-650-1500.
Specifications subject to change. Consult MSI for latest specifications.
2.6 Specifications
Performance (T=25bC) Symbol Min Typ Max Units
Sensitivity Mo 7 9 11 mV/g
Lower Frequency Limit (1) fl -- 2 5 Hz
Upper Frequency Limit (1) fu 10 20 -- kHz
Equivalent Noise Floor
10Hz
100Hz
1kHz
--
--
--
130
20
6
--
--
--
Hz g/
Dynamic Range -- 150 -- -- g
Linearity -- -- 0.1 1.0 %
Transverse Sensitivity Mt -- 2.0 5 %
Resonant Frequency fo -- 35 -- kHz
Phase Deviation (5 Limit) (6) 10 -- 10,000 Hz
Supply Voltage V+ 3 -- 40 Volts
Quiescent Supply Current lQ 10 A
JFET Pinch Off Voltage Vgss -0.6 -0.8 -1.8 V
JFET Saturation Current Idss 30 90 A
Output Impedance (6) -- -- 20 -- k
Environmental Characteristics
Operating Temperature (2) To -40 -- 85 C
Storage Temperature Ts -40 -- 85 C
Maximum Shock Level Am 1000 -- -- g
Base Strain Sensitivity (3) -- -- 0.3 -- g/
Transient Temp Sensitivity (4) -- -- 0.35 -- g/C
Physical Characteristics
Weight (5) Cable W -- 8 -- grams
(1) 3 dB limit (3) @ 250 in base plane (5) Includes 40" cable and connector
(2) 2 dB from nominal Mo at 1kHz (4) @ 3Hz LLF (6) Typical Value
A-20 dSpace, DS1103 87
A-20 dSpace, DS1103
295
DS1103 PPC Controller Board Single-Board Hardware /
2013
I
n
t
r
o
d
u
c
t
i
o
n
A
p
p
l
i
c
a
t
i
o
n

F
i
e
l
d
s
S
o
f
t
w
a
r
e
H
a
r
d
w
a
r
e
E
n
g
i
n
e
e
r
i
n
g
S
u
p
p
o
r
t

a
n
d

M
a
i
n
t
e
n
a
n
c
e
Technical Details

1)
Speed and timing specications describe the capabilities of the hardware components and circuits of our products.
Depending on the software complexity, the attainable overall performance gures can deviate signicantly from the hardware
specications.

Parameter Specication
Processor PowerPC Type n PPC 750GX
CPU clock n 1 GHz
Cache n 32 KB level 1 (L1) instruction cache
n 32 KB level 1 (L1) data cache
n 1 MB level 2 (L2)
Bus frequency n 133 MHz
Temperature sensor n Reads actual temperature at the PPC
Memory Local memory n 32 MB application SDRAM as program memory, cached
Global memory n 96 MB communication SDRAM for data storage and data exchange with host
Timer 2 general-purpose timers

n One 32-bit down counter
n Reload by software
n 15-ns resolution
n One 32-bit up counter with compare register
n 30-ns resolution
1 sampling rate timer (decrementer)

n 32-bit down counter
n 30-ns resolution
1 time base counter

n 64-bit up counter
n 30-ns resolution
Interrupt controller n 3 timer interrupts
n 7 incremental encoder index line interrupts
n 1 UART (universal asynchronous receiver and transmitter) interrupt
n 1 CAN interrupt
n 1 slave DSP interrupt
n 2 slave DSP PWM interrupts
n 1 host interrupt
n 4 external interrupts (user interrupts)
A/D converter Channels

n 16 multiplexed channels equipped with 4 sample & hold A/D converters
(4 channels belong to one A/D converter. 4 consecutive samplings are necessary to sample
all channels belonging to one A/D converter.)
n 4 parallel channels each equipped with one sample & hold A/D converter
n Note: 8 A/D converter channels (4 multiplexed and 4 parallel) can be sampled simultaneously.
Resolution n 16-bit
Input voltage range n 10 V
Overvoltage protection n 15 V
Conversion time n Multiplexed channels: 1 s
1)
n Parallel channels: 800 ns
1)
Offset error n 5 mV
Gain error n 0.25%
Offset drift n 40 V/K
Gain drift n 50 ppm/K
Signal-to-noise ratio n >83 dB
D/A converter Channels n 8 channels
Resolution n 16-bit
Output range n 10 V
Settling time n 5 s (14-bit)
Offset error n 1 mV
Gain error n 0.5%
Offset drift n 30 V/K
Gain drift n 25 ppm/K
dSPACE GmbH Rathenaustrae 26 33102 Paderborn Germany info@dspace.de www.dspace.com
88 Appendix
296
DS1103 PPC Controller Board Single-Board Hardware /
2013
Parameter Specication
D/A converter Signal-to-noise ratio n >83 dB
Imax n 5 mA
CImax
n 10 nF
Digital I/O Channels

n 32-bit parallel I/O
n Organized in four 8-bit groups
n Each 8-bit group can be set to input or output (programmable by software)
Voltage range n TTL input/output levels
Iout , max n 10 mA
Digital incremental
encoder interface
Channels

n 6 independent channels
n Single-ended (TTL) or differential (RS422) input (software programmable for each channel)
Position counters

n 24-bit resolution
n Max. 1.65 MHz input frequency, i.e.,fourfold pulse count up to 6.6 MHz
n Counter reset or reload via software
Encoder supply voltage

n 5 V/1.5 A
n Shared with analog incremental encoder interface
Analog incremental
encoder interface
Channels

n 1 channel
n Sinusoidal signals: 1 Vpp differential or 11 App differential (software programmable)
Position counters

n < 5 resolution
n 32-bit loadable position counter
n Max. 0.6 MHz input frequency, i.e., fourfold pulse count up to 2.4 MHz
A/D converter performance

n 6-bit resolution
n 10 MSPS
Encoder supply voltage

n 5 V/1.5 A
n Shared with digital incremental encoder interface
CAN interface Conguration

n 1 channel based on SAB 80C164 microcontroller
n ISO DIS 11898-2 CAN high-speed standard
Baud rate n Max. 1 Mbit/s
Serial interface Conguration

n TL6C550C single UART with FIFO
n PLL-driven UART for accurate baud rate selection
n RS232/RS422 compatibility
Baud rate

n Up to 115.2 kBd (RS232)
n Up to 1 MBd (RS422)
Slave DSP Type n Texas Instruments TMS320F240 DSP
Clock rate n 20 MHz
Memory

n 64 Kx16 external code memory
n 28 Kx16 external data memory
n 4 Kx16 dual-port memory for communication
n 32 KB ash memory
I/O channels
1)

n 16 A/D converter inputs
n 10 PWM outputs
n 4 capture inputs
n 2 serial ports
Input voltage range

n TTL input/output level
n A/D converter inputs: 0 5 V
Output current n Max. 13 mA
Host interface n Plug & Play support
n Requires a full-size 16-bit ISA slot
Physical
characteristics
Physical size n 340 x 125 x 45 mm (13.4 x 4.9 x 1.77 in)
Ambient temperature n 0 50 C (32 122 F)
Cooling n Passive cooling
Power supply

n +5 V 5%, 4 A
n +12 V 5%, 0.75A
n -12 V 5%, 0.25A
1)
The exact number of I/O channels depends on your conguration and is described in the user documentation.
dSPACE GmbH Rathenaustrae 26 33102 Paderborn Germany info@dspace.de www.dspace.com
A-21 Drawing sensor mount, Sensorcarrier1 89
A-21 Drawing sensor mount, Sensorcarrier1

R
5

R
5

B B

2
6

2
2

A
A
SECTION F-F
SECTION B-B
22

1
.
5
0

2

1
4

2 2

3

26

1
2

SECTION A-A
49
51
F F
CHK'D
sensorcarrier1
WEIGHT:
A4
SHEET1OF 1 SCALE:1:1
DWG NO.
TITLE:
REVISION DO NOTSCALEDRAWING
MATERIAL:
DATE SIGNATURE NAME
APPV'D
MFG
EDGES
FINISH:
BREAK SHARP
DEBURAND
Q.A
ANGULAR:
UNLESSOTHERWISESPECIFIED:
DIMENSIONSAREIN MILLIMETERS
SURFACEFINISH:
TOLERANCES:
LINEAR:
DRAWN

1

1
0

1

30

1
4

6

90 Appendix
A-22 Beyerdynamic MM1 reference microphone
APPLICATIONS
The MM 1 is a measurement microphone which has been
designed specifically for measuring sound reinforcement
and PA-systems. It is designed to work with spectrum
analysers for measuring frequency response and sound
pressure levels of loud speaker systems. The MM 1 is the
ideal microphone for the measurement of audio signals in
the research, development, for reverberation testings and
other applications.
The narrow tubular construction ensures that the micro-
phone has negligable influence on the sound field so that
an increase in sound pressure is avoided with high
frequencies. A natural reproduction is achieved due to
the linear frequency response.
FEATURES
Linear frequency response
Omnidirectional polar pattern
Calibrated open circuit voltage
Narrow tubular construction
Germany
Theresienstr. 8
D-74072 Heilbronn
Tel. +49 (0)7131 / 617-0
Fax +49 (0)7131 / 617-224
E-mail: info@beyerdynamic.de
Internet: www.beyerdynamic.de
United States
56 Central Ave.
Farmingdale, NY 11735
Tel. +1 (631) 293-3200
Fax +1 (631) 293-3288
E-mail: salesUSA@beyerdynamic.com
Internet: www.beyerdynamic.com
Great Britain
17 Albert Drive
Burgess Hill RH15 9TN
Tel. +44 (0)1444 / 258258
Fax +44 (0)1444 / 258444
E-mail: sales@beyerdynamic.co.uk
Internet: www.beyerdynamic.co.uk
MM 1
Measurement Microphone
Order # 449.350
OPTIONAL ACCESSORIES
GST 400 Microphone stand, 3/8", height 0.90 - 1.65 m, with G 400 boom . . . . . . . . . . . . . . . . . Order # 421.294
GST 500 Microphone stand, 3/8", height 0.85 - 1.60 m, with telescopic G 500 boom. . . . . . . . . Order # 406.252
ST 400 Microphone stand, 3/8", height 0.90 - 1.65 mm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Order # 421.286
ST 500 Microphone stand, 3/8", height 0.85 - 1.60 mm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Order # 406.643
WS 10 Windscreen, charcoal grey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Order # 403.008
A-22 Beyerdynamic MM1 reference microphone 91
Photos are non-contractual. Contents subject to change without notice. Printed in Germany.
b
e
y
e
r
d
y
n
a
m
i
c
)
)
)
)

M
M

1
FREQUENCY RESPONSE & POLAR PATTERN
This polar pattern and frequency response curve (measuring tolerance 1.5 dB) correspond to a typical
production sample for this microphone.
WIRING DIAGRAM
Frequency response 1.5 dB 0 dB 15 mV/Pa MM 1 =
^
TECHNICAL SPECIFICATIONS
Transducer type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Condenser (back electret)
Operating principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pressure
Frequency response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 - 20,000 Hz
(50 - 16,000 Hz 1.5 dB)
Polar pattern. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Omnidirectional, diffuse field calibrated
Open circuit voltage at 1 kHz . . . . . . . . . . . . . . . . . . . . . . . 15 mV/Pa ( -36.5 dBV) 1 dB
Nominal impedance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330 !
Nominal load impedance. . . . . . . . . . . . . . . . . . . . . . . . . . . " 2.2 k!
Max. SPL at f = 1 kHz, k = 1%, RL = 2.2 k! . . . . . . . . . . . . . 128 dBSPL
S/N ratio rel. to 1 Pa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . > 57 dB
A-weighted equivalent SPL . . . . . . . . . . . . . . . . . . . . . . . . . approx. 28 dB(A)
Power supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 - 48 V phantom supply
Current consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . approx. 3.4 mA
Output. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . transformer balanced
Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-pin XLR male
Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Length: 133 mm
Shaft diameter: 19/9 mm
Head diameter: 9 mm
Weight (w/out cable) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 g
=
^
92 Appendix
A-23 Philips NatLab improved MFB
A-23 Philips NatLab improved MFB 93
94 Appendix
96 Appendix
98 Appendix
Bibliography
[1] K. M. Al-Ali. Loudspeakers: Modeling and Control. PhD thesis, University of California
at Berkeley, 1999.
[2] E. De Boer. Theory on motional feedback. IRE Transactions on audio, January-
February:1551, 1961.
[3] R. Breden. Roaring subwoofer. Electronics world, February:104109, 1997.
[4] N. Quaegebeur A. Chaigne. Nonlinear vibrations of loudspeaker-like structures. Journal
of Sound and Vibration, 309:178196, 2008.
[5] S. Chiu. A direct pwm loudspeaker feedback system. Masters thesis, Massachusetts
Institute of Technology, 1996.
[6] Jr D. Davis E. Patronis. Sound System Engineering. Focal Press, 2006.
[7] D. de Greef J. Vandewege. Acceleration feedback loudspeaker. Wireless world,
September:3236, 1981.
[8] S. D. Dzisiewski-Smith. Positional loudspeaker feedback and control using dsp within a
class d audio power amplier system. Masters thesis, Imperial college London, 2007.
[9] P.A. Nelson S.J. Elliott. Active Control of Sound. Academic Press Inc., 1992.
[10] G. F. Franklin J. D. Powell A. Emami-Naeini. Feedback control of dynamics systems.
Pearson Prentice Hall, 2006.
[11] B. Friedland. Control System Design. Dover Publications, 2005.
[12] M. Petyt P. N. Glat. Vibration of loudspeaker cones using the dynamic stiness method.
Applied Acoustics, 53:313332, 1998.
[13] S.M. Kuo R. K. Yenduri A. Gupta. Frequency domain delayless active sound quality
control algorithm. Journal of Sound and Vibration, 318:714724, 2008.
100 Bibliography
[14] P. G. L. Mills M. O. J. Hawksford. Transconductance power amplier systems for current
driver loudspeakers. Audio Engineering Society, 37, 10:809822, 1989.
[15] L. Yali Y. Hongwu Z. Hui. Speaker identication based in emd. Technical report, College
of Physics and Electronic Engineering, Northwest Normal University, China, 2009.
[16] Y. Jing. Robust vibration control based on identied models. Journal of Sound and
Vibration, 269:317, 2004.
[17] M. J. Brennan S. M. Kim. Feedforward and feedback control of sound and vibration, a
weiner lter approach. Journal of Sound and Vibration, 246:281296, 2001.
[18] L. Kreitmeier. Fem simulation of loudspeakers and loudspeaker components. Technical
report, Harman/becker automotive systems, Germany, 2002.
[19] A. G. A. Mathalif R. S. Langley. Active control of high-frequency vibration: Optimisation
using the hybrid modelling method. Journal of Sound and Vibration, 331:29692983,
2012.
[20] S. J. Loutridis. Resonance identication on loudspeaker driver units: A comparison of
techniques. Applied Acoustics, 66:13991426, 2005.
[21] R. Miller. Modal analysis of loudspeaker diaphragms. Masters thesis, Faculty of Rens-
selaer Polytechnic Institute, Connecticut, 2010.
[22] K. J. Astrm R. M. Murray. Feedback systems. Princeton University Press, 2008.
[23] S. Cecchi E. Moretti F. Piazza. A new approach to bass enhancement based on pronys
method. Technical report, Universita Politecnica della Marche, 2007.
[24] T. S. Hsu K. A. Poornima. Loudspeaker failure modes and error correction techniques.
Applied Acoustics, 62:717734, 2001.
[25] A. Draeger S. Engell H. Ranke. Model predictive control using neural networks. IEEE
Control, October:6166, 1995.
[26] R. Ravaud G. Lemarquand T. Roussel. Time-varying non linear modeling of electrody-
namic loudspeakers. Applied Acoustics, 70:450458, 2009.
[27] H. Schurer. Second order volterra inverses for compensation of loudspeaker nonlinearity.
Applications of Signal Processing to Audio and Acoustics, 1:165168, 1995.
[28] J. Shaw. Adaptive control for sound and vibration attenuation: a comparative study.
Journal of Sound and Vibration, 235:671684, 2000.
[29] Z. L. Zhang Q. T. Tao. Experimental study of non-linear vibrations in a loudspeaker
cone. Journal of Sound and Vibration, 248:18, 2001.
[30] F. E. Toole. Sound reproduction. Elsevier, 2008.
[31] J. Suykens J. Vandewalle J. van Ginderdeuren. Feedback linearization of nonlinear
distortion in electrodynamic loudspeakers. Journal of the Audio Engineering Soviety,
43:690694, 1995.
101
[32] C. R. Fuller A. H. von Flotow. Active control of sound and vibration. IEEE Control
Systems, December:919, 1995.
[33] A.J.M. Kaizer C.A.M. Wesche. An improved motional feedback loudspeaker system.
Technical report, Philips NAT.LAB., 1978.
[34] J. Peifang L. Wei W. Schuaibing Y. Jun G. Woon-Seng. An alternative method to
measure the on-axis dierence-frequency sound in a parametric loudspeaker without
using an acoustic lter. Applied Acoustics, 73:12441250, 2012.
[35] J. Wei G. Woon-Seng. Identication of a parametric loudspeaker system using an adap-
tive volterra lter. Applied Acoustics, 73:12511262, 2012.
[36] Z. Limin W. Lifu Q. Xiaojun. An intuitive approach for feedback active noise controller
design. Applied Acoustics, 74:160168, 2013.
[37] U Zlzer. Digital Audio Signal Processing. Wiley, 2008.
102 Bibliography

Motional Feedback On A Loudspeaker Woofer

Uploaded by

Copyright:

Available Formats

Motional Feedback On A Loudspeaker Woofer

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Motional Feedback On A Loudspeaker Woofer

Uploaded by

Copyright:

Available Formats

Department of Precision and Microsystems Engineering

Control of Voicecoil transducers

Master of Science Thesis R. Valk

XP or Visita 64, Gigabit Ethernet, data aquisition hardware

You might also like