U.S. patent application number 12/006108 was filed with the patent office on 2008-07-24 for power amplification for parametric loudspeakers.
Invention is credited to Jeevan G. Bank, James J. Croft.
Application Number | 20080175404 12/006108 |
Document ID | / |
Family ID | 31993723 |
Filed Date | 2008-07-24 |
United States Patent
Application |
20080175404 |
Kind Code |
A1 |
Bank; Jeevan G. ; et
al. |
July 24, 2008 |
Power amplification for parametric loudspeakers
Abstract
A method for minimizing the reactance of a reactive load
transducer at a carrier frequency in an amplifier with a power
output stage for amplifying a signal, comprises: supplying a signal
to the power output stage of the amplifier which includes at least
one reactive component coupled to the amplifier; and counteracting
transducer reactance at a carrier frequency by interacting the
signal with the at least one reactive component.
Inventors: |
Bank; Jeevan G.; (San Diego,
CA) ; Croft; James J.; (Poway, CA) |
Correspondence
Address: |
THORPE NORTH & WESTERN, LLP.
P.O. Box 1219
SANDY
UT
84091-1219
US
|
Family ID: |
31993723 |
Appl. No.: |
12/006108 |
Filed: |
December 28, 2007 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10332659 |
Oct 3, 2003 |
7319763 |
|
|
PCT/US01/21749 |
Jul 11, 2001 |
|
|
|
12006108 |
|
|
|
|
60217373 |
Jul 11, 2000 |
|
|
|
Current U.S.
Class: |
381/77 |
Current CPC
Class: |
H03F 2200/541 20130101;
H03F 3/2171 20130101; H03F 3/181 20130101; H03F 3/2173 20130101;
H03F 3/2178 20130101 |
Class at
Publication: |
381/77 |
International
Class: |
H03F 3/181 20060101
H03F003/181; H03F 3/20 20060101 H03F003/20; H03F 3/217 20060101
H03F003/217 |
Claims
1. A method for minimizing the reactance of a reactive load
transducer at a carrier frequency in an amplifier with a power
output stage for amplifying a signal, comprising: supplying a
signal to the power output stage of the amplifier which includes at
least one reactive component coupled to the amplifier; and
counteracting transducer reactance at a carrier frequency by
interacting the signal with the at least one reactive
component.
2. A method as in claim 1, wherein counteracting transducer
reactance further comprises counteracting the transducer reactance
of a capacitive transducer load using at least one inductively
reactive component.
3. A method as in claim 1, wherein counteracting transducer
reactance further comprises counteracting the transducer reactance
of an inductive transducer load using at least one capacitively
reactive component.
4. A method for minimizing the reactance of a reactive load
transducer at a transducer's off-resonant frequency in an amplifier
with a power output stage for amplifying a signal, comprising the
steps of: supplying a signal to the power output stage of the
amplifier which includes at least one reactive component coupled to
the amplifier; and counteracting transducer reactance at the
transducer's off-resonant frequency by interacting the signal with
the at least one reactive component.
5. A method as in claim 4, wherein counteracting transducer
reactance further comprises counteracting the transducer reactance
of a capacitive transducer load using at least one inductively
reactive component.
6. A method as in claim 4, wherein counteracting transducer
reactance further comprises counteracting the transducer reactance
of an inductive transducer load using at least one capacitively
reactive component.
7. A power amplification system configured for improved efficiency
in a parametric loudspeaker system reproducing an amplitude
modulated carrier signal that can be used to produce a parametric
acoustic output, including a switching power stage that amplifies
the modulated carrier signal over one of a carrier frequency and a
sideband frequency at a switching frequency that is essentially an
integer multiple of the carrier frequency within a frequency
tolerance limit that is no greater than the lowest audible
frequency of the parametric loudspeaker system.
Description
PRIORITY CLAIM
[0001] This is a continuation of copending U.S. patent application
Ser. No. 10/332,659, filed Jan. 10, 2003, which is a 371 of
PCT/US01/21749, filed Jul. 11, 2001, which claims benefit of U.S.
Provisional Patent Application Ser. No. 60/217,373, filed Jul. 11,
2000, each of which is hereby incorporated herein by reference in
their entireties.
TECHNICAL FIELD
[0002] The present invention relates generally to parametric
loudspeaker systems. More specifically, it relates to power
delivery systems for parametric loudspeakers.
BACKGROUND ART
[0003] The parametric loudspeaker is an electroacoustic system that
operates by producing an ultrasonic carrier frequency, for example
40 kHz, that is then modulated by an audio input signal. The
modulation shifts the audio frequency up to the frequency of the
carrier plus the audio frequency. This upshifted frequency (f1)
interacts with the carrier frequency (f2) thus generating an
audible reproduction of the audio input signal by driving the air
to non-linearity which produces the audible signal of interest
(f1-f2) plus other components (such as f1+f2). The ultrasonic upper
frequency requirement of a parametric system is typically at least
60 kHz because this allows 20 kHz of the audio signal to be
modulated on top of the 40 kHz carrier signal.
[0004] Historically, the use of parametric loudspeakers has been
limited. This is partially due to their general inefficiency
because the sound output is based on a second order effect of the
demodulation of ultrasonic sound waves in the air into audible
sound. This second order effect needs a greater amount of power to
drive the system and deliver the audio output.
[0005] Parametric power delivery systems also have further reduced
efficiency because the parametric system requires a continuous
carrier frequency output. At full audio output, the carrier
frequency is operated at a constant 1/4 power output level, which
causes high power dissipation in the amplifier. Even at lower audio
levels or during a break in the music the carrier signal must be
driven at high constant power levels.
Further, most parametric loudspeaker transducers exhibit highly
reactive loads. In the prior art, parametric transducers are driven
using a conventional linear power amplifier to directly drive the
transducer, and they require very large power amplifiers that
dissipate significant power and heat in the output stage.
[0006] Due to the high continuous power levels that can be
required, the transducers that work best for parametric or
ultrasonic loudspeakers tend to have dominantly reactive
(capacitive and/or inductive) characteristics. This is in contrast
to conventional electromagnetic speakers which tend to have a
dominant resistive characteristic. One of the reasons for using
reactive speakers in a parametric system is that the high average
level of the carrier frequency can cause high thermal dissipation
in the resistive element of any transducer. A purely reactive
transducer dissipates very little heat in the device itself because
of the reactive load it provides to the amplifiers.
Correspondingly, the output stage of the power amplifier
(particularly a linear amplifier) coupled to a reactive transducer
or speaker has significant thermal losses. These losses are caused
because the power amplifier must amplify highly reactive charging
currents when driving the reactive load directly. The problem is
particularly detrimental at the frequencies of greatest output,
such as the carrier frequency and frequencies associated with
lowest audio frequencies to be reproduced.
[0007] A related major issue with prior art parametric loudspeakers
is that the reactive load transducers require significant reactive
charging power. In turn, that power requirement has forced the use
of much higher output power amplifiers to supply this wasteful
power.
[0008] Prior art parametric loudspeakers have used what is commonly
known in the art as a linear or Class B amplifier topology which
reaches maximum efficiency at full power and is at its most
thermally inefficient mode at the 1/4 power level or the equivalent
half voltage level. As an example, a 100 watt Class B amplifier
when operating at 1/4 power may dissipate 50 watts into wasted heat
while outputting only 25 "useful" watts to the load. This is both
an inefficient waste of power and a costly system to build because
it can require extensive cooling systems.
[0009] A serious contributor to the inefficiency of a linear power
amplifier in a parametric system is the fact that the common
transducer type used in parametric loudspeakers has a reactive
impedance that must be driven by the power amplifier. It is well
known that linear amplifiers have a significant reduction in
efficiency and increase in heat when driving a reactive load.
Accordingly, it is desirable to provide a system which would allow
a more efficient use of amplifier power in a parametric speaker
system.
SUMMARY OF THE INVENTION
[0010] In accordance with one embodiment, the invention provides a
method for minimizing the reactance of a reactive load transducer
at a carrier frequency in an amplifier with a power output stage
for amplifying a signal, including: supplying a signal to the power
output stage of the amplifier which includes at least one reactive
component coupled to the amplifier; and counteracting transducer
reactance at a carrier frequency by interacting the signal with the
at least one reactive component.
[0011] Additional features and advantages of the invention will be
set forth in the detailed description which follows, taken in
conjunction with the accompanying drawings, which together
illustrate by way of example, the features of the invention.
BRIEF DESCRIPTION OF THE DRAWINGS
[0012] FIG. 1 is a block diagram of a Class D pulse width
modulation amplifier;
[0013] FIG. 2 is a schematic of a parametric loudspeaker system
with a H-Bridge amplifier, output filtering, and transformer load
coupling;
[0014] FIG. 3 is a schematic of a half bridge amplifier with output
filtering;
[0015] FIG. 4 is a graph of the pulse width modulation (PWM)
frequency at approximately 23 times the carrier frequency;
[0016] FIG. 5 is a graph of distortion frequencies created by the
PWM frequency in FIG. 4;
[0017] FIG. 6 illustrates a frequency spectrum graph with a
non-synchronous pulse width modulation frequency that creates
distortion in the upper side band of the audio;
[0018] FIG. 7 illustrates a frequency spectrum graph with a
synchronous PWM frequency which avoids distortion in the upper
audio side band;
[0019] FIG. 8 is a block diagram of a parametric system with a
Class D amplifier which avoids distortion spurs using a frequency
multiplier to create a PWM frequency that is a selected multiple of
the carrier;
[0020] FIG. 9 illustrates the generation of an ideal signal sine
wave;
[0021] FIG. 10 illustrates the generation of a signal wave using
bipolar pulse width modulation at 3 times the carrier
frequency;
[0022] FIG. 11 illustrates the generation of a signal wave using
unipolar pulse width modulation at 2 times the carrier
frequency;
[0023] FIG. 12 illustrates the generation of a signal wave using
bipolar pulse width modulation at 5 times the carrier
frequency;
[0024] FIG. 13 illustrates the unipolar pulse width modulation at 4
times the carrier frequency;
[0025] FIG. 14 is a schematic of a parametric loudspeaker system
with a linear power amplifier and a power transformer;
[0026] FIG. 15 is a graph of the frequency response curve for the
parametric loudspeaker system in FIG. 14;.
[0027] FIG. 16 is a graph of the frequency response curve for the
system in FIG. 14, including the acoustic output curve;
[0028] FIG. 17 is a parametric loudspeaker system with an inductor
tuned to an off-resonant transducer frequency;
[0029] FIG. 18 is a graph of the frequency response of the system
of FIG. 17 with an inductor tuned to an off-resonant transducer
frequency, and a curve showing the response without the
inductor;
[0030] FIG. 19 is a parametric loudspeaker system with an inductor
in parallel, tuned to an off-resonance transducer frequency;
[0031] FIG. 20 is a parametric loudspeaker system with an inductor
and variable capacitor to tune the frequency to an optimal
frequency;
[0032] FIG. 21 is a graph of the frequency response curve for FIG.
20;
[0033] FIG. 22 is a schematic of a parametric loudspeaker with an
inductive speaker and a parallel resonating capacitance;
[0034] FIG. 23 is a schematic of a parametric loudspeaker system
with an inductive speaker and a resonating capacitance in
series;
[0035] FIG. 24 is a graph of the frequency response curves of FIGS.
22 and 23 when reactive components are tuned to a frequency
off-resonance of the transducer;
[0036] FIG. 25 is a graph of the frequency response curves of FIGS.
22 and 23 when the reactive components are tuned at the resonance
of the transducer;
[0037] FIG. 26 is a schematic of a parametric loudspeaker system
with reactive circuitry including a group of inductors and a
capacitor;
[0038] FIG. 27 is a schematic of a parametric loudspeaker system
with reactive circuitry including a group of inductors and a group
of capacitors;
[0039] FIG. 28 is a variable saturable inductor for use in FIGS.
17, 19, 20, and 26-32;
[0040] FIG. 29 is a schematic of a parametric loudspeaker system
with reactive circuitry including a group of inductors in series
and a capacitor in parallel;
[0041] FIG. 30 depicts a parametric loudspeaker system with
reactive circuitry including an inductor before a step-up
transformer, a parallel capacitor, and an inductance contained in
the transformer;
[0042] FIG. 31 is a schematic of a parametric loudspeaker system
with reactive circuitry including an inductor before a step-up
transformer, a parallel capacitor, and an inductor after the
transformer;
[0043] FIG. 32 is a schematic of a parametric loudspeaker system
with reactive circuitry including an inductor and a parallel
capacitor before the transformer and an inductor group and
capacitor after the transformer;
[0044] FIG. 33 is a schematic of a power transformer isolation
configuration for a linear or switch-mode power amplifier;
[0045] FIG. 34 illustrates a multi-level switching amplifier for
parametric loud-speakers;
[0046] FIG. 35 illustrates a multi-level switching amplifier for
parametric loud-speakers using an output coupling transformer.
DETAILED DESCRIPTION
[0047] For the purposes of promoting an understanding of the
principles of the invention, reference will now be made to the
exemplary embodiments illustrated in the drawings, and specific
language will be used to describe the same. It will nevertheless be
understood that no limitation of the scope of the invention is
thereby intended. Any alterations and further modifications of the
inventive features illustrated herein, and any additional
applications of the principles of the invention as illustrated
herein, which would occur to one skilled in the relevant art and
having possession of this disclosure, are to be considered within
the scope of the invention.
[0048] Since the normal operating mode of a parametric loudspeaker
can be 1/4 output power even at idle, and the transducers used in
parametric loudspeakers usually have a significant reactive
component in their load impedance, conventional linear amplifiers
tend to operate inefficiently in parametric loudspeaker systems. As
a result of the problems described before, it can be seen that a
major contributor to the inefficiency of prior art parametric
systems has been the linear Class B amplification systems used.
[0049] Another type of amplifier is a switch-mode or Class D power
amplifier. This amplifier uses a switching output stage whose
output power signal is in either an "on" or "off" condition as
opposed to the linear signal mode of a Class B design. With a
switch-mode power amp, the efficiency is very high and the
efficiency can be maintained even with a reactive load. This
technology has been known to improve the efficiency of low
frequency systems significantly but has only recently been capable
of operating effectively with audio devices up to 20 kHz (i.e.,
maximum audible frequencies). Even then, it has been difficult to
operate without greater distortion or high frequency response
errors, even at frequencies below 20 kHz when used with capacitive
transducers.
[0050] FIG. 1 illustrates one embodiment of a Class D power
amplifier or a pulse width modulation (PWM) amplifier in a
parametric system with output filtering before the signal is sent
to the load or transducer. An input signal source 100 is provided
to an error amplifier 102 which receives an inverting input from a
feedback amplifier 103 to control signal gain. The signal is then
provided to a loop filter 104 which can be a lag-lead filter to
stabilize the feedback loop. A pulse width modulator 106 is
included to generate, control, and vary the pulse width or length
of the pulses generated. Bridge drivers 108 control the switching
of the output power switches 110. The bridge drivers and pulse
width modulator can be combined into one integrated chip or reside
on separate chips as depicted. The output power switches receive
power from the power supply 112 and deliver a switched
amplification signal using either a half or full bridge switching
configuration. Filter components 114 are used to convert the pulse
width modulation to an analog signal and to reduce the switching
noise. One set of filters is needed if a half bridge switching
configuration is used and a second set of symmetrical filters 116
is used if a full bridge switching configuration is used. A
transducer load 118 is shown connected to the filter
components.
[0051] One issue with prior art Class D power amplifiers is for the
high frequency performance of the amplifier to equal or even
approach that of a Class B design, the switching frequency is
required to be at least 10 to 25 times the highest frequency to be
reproduced and preferably even greater. This is a difficult design
criteria to meet even with an audio signal bandwidth of only 0-20
kHz, because it requires a switching frequency of 200 kHz to 500
kHzClass D amplifiers with this frequency range have been used but
not with extensive success.
[0052] Using even higher switching frequencies to achieve greater
performance is desirable from the standpoint of linearity, output
impedance, and lower cost filter design, but higher switching
frequencies are prohibitive from the standpoint of switching losses
and thermal dissipation. In fact, some Class D amplifiers where
higher switching frequencies (above 400 kHz) have been attempted
have resulted in amplifiers that are as inefficient as a standard
Class B amplifier.
[0053] Another issue with prior art Class D amplifiers is that they
generally use one or more LC filters on the output to produce a
linear signal. These LC filters are designed to have a large enough
value to minimize the switching noise of the Class D amplifier
while at the same time having the conflicting requirement to be
small enough to not interact with the loudspeaker load at high
audio frequencies.
[0054] Yet another problem with prior art Class D switching
amplifiers is that they can recycle energy from the power supply to
the load and then back into the power supply again. Because of this
issue most Class D amplifiers must use a bridged output stage that
has twice as many power transistors, drivers, rectifiers and output
filters. If this configuration is not used, then it is possible to
have energy flow from a positive power supply terminal to the load
and recycle back into the negative power supply. This recycling can
produce an overvoltage and cause a catastrophic overvoltage
condition. Either the more costly bridged output topology or else a
complex power supply balancing system may be required.
[0055] Despite the drawbacks of a Class D amplifier, it is
desirable to use in a parametric loudspeaker system because of the
amplifier efficiency increase in general, and in particular with
reactive loads because of the efficiency increase at 1/4 power. The
minimum 10 to 25 times requirement of prior art Class D amplifiers
when applied to a parametric loudspeaker system (whose highest
frequency to be reproduced is usually at least 3 times higher than
the audio band which is already difficult for a Class D amplifier)
at 60 kHz requires the switching frequency of the Class D power
amplifier to be 600 to 1500 kHz. These high switching frequencies
can be difficult and expensive to realize when using Class D
amplifiers. An additional drawback with a more expensive and
complex Class D amplifier is when it is used to produce 600-1500
kHz frequencies, the efficiency of the Class D amplifier degrades
and a significant power loss may be produced that approaches a
linear amplifier. Using a Class D amplifier at ultrasonic
frequencies can increase the cost and complexity of an amplifier
system, without retaining the efficiency of Class D amplifiers
utilizing a lower switching frequency. As can be seen, it would be
valuable to utilize the efficiency of a Class D amplifier in a
parametric loudspeaker system while still maintaining the
efficiency available at its lower switching frequencies.
[0056] As illustrated in FIG. 2, a parametric loudspeaker system
includes a Class D amplifier with a H-bridge amplifier, output
filtering, and a transformer load coupling. Four switching power
devices are shown as 120a-120d which switch the power received from
the power source line 122. A number of pulse driver lines 130a-130d
are connected to the bridge drivers to drive the switching devices.
The switch devices can be MOSFETs, high speed bipolar transistors
or another fast power switching device known in the art. A typical
voltage component rating for use in this system is a 60 V MOSFET.
Higher voltage MOSFETs can be used, such as 200 V, but these are
more expensive.
[0057] Two switch filters 124a and 124b are also used in the full
bridge configuration to filter out the high frequency switching
noise. A capacitor 126 is included to filter out any DC signal. To
match the voltage to the transducer 132, a transformer 128 is used
to step-up or step-down the voltage provided by the switching
devices 130a-d. A transformer is especially helpful where an
ultrasonic signal will be produced because the ultrasonic
transducer can require higher voltages or higher currents than the
direct output for which the amplifier is optimized. Of course, it
is possible to omit the transformer and match high voltage power
devices (e.g., MOSFETs, etc.) to the high voltage transducer but
this depends on which method is desired. Two pairs of two pole
filters 131a and 131b are included to provide further high
frequency filtering and/or they can be tuned for reactive load
matching.
[0058] FIG. 3 is a schematic of a half bridge amplifier with output
filtering. Two switching power devices are shown as 140a and 140b
which switch the power received from the power source line 142.
Pulse driver lines 144a and 144b are each connected to the bridge
drivers to drive the switching devices. The switch devices can be
MOSFETs, high speed bipolar transistors or similar fast power
switching devices known in the art. In a full bridge configuration,
a four pole filter 145 is included to filter out the high frequency
switching noise. Additionally, a transformer coupling can be
located between the Class D amplifier and the transducer 148 to
match the output of the amplifier to the impedance of the
transducer.
[0059] FIG. 4 is a chart illustrating a pulse width modulation
frequency 150 that is approximately 20-25 times greater than the
carrier frequency signal 152 or a frequency of 800 kHz to 1 MHz. An
ultrasonic signal uses a carrier signal above 20 kHz. The preferred
carrier frequency is 30 to 60 kHz and that is modulated with audio
signal sidebands which can add another 15 or 20 kHz to the
frequency. A Class D power amplifier uses a switching frequency
approximately 20 times or more than the highest reproduced
frequency, and this requires switching frequencies of 1.2 MHZ or
more which are not practical with present switch-mode or Class D
technology. If a lower switching frequency is used, distortion is
created within the audio related sideband which produces distorted
audible sounds and/or a greater potential for more filter
interaction because larger filters must be used to filter these
lower distortion frequencies. Moreover, frequency response
anomalies can be created when a lower switching frequency is used
due to the interaction between filters and the loudspeaker
load.
[0060] Referring again to FIG. 4, the harmonics 153 which are an
integer multiple of the carrier wave are also shown. FIG. 5
illustrates a close-up view of the carrier frequency in FIG. 4 with
the harmonics 153 shown as dotted lines. When the pulse width
modulation (PWM) frequency is at a level that is less than 25 times
greater than the carrier, in-band distortion spurs 154 are created.
These in-band distortion spurs can be heard in the audio reproduced
by the parametric transducer. FIG. 6 illustrates a pulse width
modulation frequency 156 at approximately 170 kHz. Even when the
pulse width modulation is closer to the carrier 158, a distortion
spur 160 is still created in the upper side band audio frequency
162. It should be realized that the figures illustrate single
sideband (SSB) (i.e. upper sideband (USB) audio signals modulated
onto the carrier but a lower side band (LSB) or double sideband
(DSB) can also be used. These illustrations represent the single
sideband (or upper sideband) case but the distortion spurs would be
mirrored when lower or double sideband modulation is used.
[0061] Instead of picking an arbitrary frequency value for the
pulse width modulation, it has been discovered that it is
surprisingly advantageous to use a pulse width modulation frequency
that is a multiple of the carrier frequency. The valuable result is
that the secondary distortion harmonics from the PWM frequency fall
directly on the carrier and any other smaller harmonics coincide
with higher harmonics of the carrier. FIG. 7 illustrates the
situation where a PWM frequency 164 is set at a frequency of four
(4) times the carrier frequency 158. Since the distortion harmonics
of the PWM frequency fall outside the upper side band audio 162 (or
lower side bands when provided), then no audible distortion is
produced. FIG. 7 also shows other integer harmonics 166 which could
be used for the PWM frequency.
[0062] Synchronization of the PWM frequency to the appropriate
multiple of the carrier frequency also allows for smaller filter
components since the second harmonic is canceled out without
filtering. In other words, the synchronization reduces the required
cut off rate and/or increases the frequency where the cut off
filter must be set. A major advantage of using a PWM frequency at a
harmonic of the carrier frequency is that the same Class D
amplifier which is used with audio frequencies, for example 400
kHz, can also be used at 400 kHz with ultrasonic frequencies
instead of the expected 800 kHz-1.5 MHz. Using synchronization also
allows for tuned bandpass filtering where the frequencies below 20
kHz can be filtered out. This allows for a smaller transformer
because the lower frequencies are not needed.
[0063] Although using an exact integer multiple of the carrier
frequency is the most advantageous configuration, the PWM frequency
can be slightly displaced from the integer multiple of the carrier
without audible distortion. Even if the PWM frequency is only
substantially a carrier multiple, advantageous effects are still
produced. In this situation, the switching frequency of the power
stage of the switch-mode power amplifier corresponds substantially
to a multiple of a carrier frequency of the parametric loudspeaker
system within a "frequency tolerance limit." The frequency
tolerance limit is defined by a correspondence with the lowest
audible frequency of operation for the parametric loudspeaker. In
other words, a parametric transducer has a threshold frequency
below which it cannot effectively reproduce audible sound.
Distortion products produced by the PWM below that threshold
frequency are not audible. This lowest audible frequency is often
between 200 Hz to 400 Hz or less and the distortion components
associated with the frequency tolerance limit are less than the
lowest audible frequency of operation. The frequency tolerance
limit T.sub.L is less than or equal to the multiple of the carrier
(xC) times the lowest frequency limit (LFL). This can also be
written as:
T.sub.L<=(xC)*LFL
[0064] So, if the multiple of the carrier is 3 and the lowest
audible frequency is 300 Hz, then the frequency tolerance limit is
900 Hz. This is the maximum amount which the PWM may be shifted off
the harmonic without audible distortion.
[0065] An example of this is using a 100 kHz carrier frequency and
a switching PWM frequency at 400 kHz to amplify the carrier (i.e.,
4 times the carrier). If the PWM frequency is displaced slightly
off the 4.sup.th integer multiple of the carrier (400 kHz) by 200
Hz, this creates a distortion product at 50 Hz. To determine the
distortion frequency, the displacement (200 Hz) is divided by the
integer multiple 4 (the PWM multiplier or (xC)). This resultant
distortion at 50 Hz is not audible in a parametric speaker system
that only has low frequency capability down to 200 Hz. If the
displacement were 2000 Hz, then the distortion harmonic would be
500 Hz and just above the reproducible threshold which is
undesirable.
[0066] FIG. 8 is a block diagram of a Class D amplifier with the
required circuitry for generating a pulse width modulated (PWM)
power signal at selected frequencies. First, a carrier wave
frequency is generated 200. The carrier wave frequency is delivered
to the carrier frequency reference input 202 and the generalized
amplitude modulation (including the carrier) 204. The generalized
amplitude modulation is modulated with the audio input 206 to
create a carrier modulated audio signal. The carrier modulated
signal passes through an error amplifier 208 before it is passed
onto the comparator 210.
[0067] Concurrently, the carrier reference input is sent to the
phase locked loop frequency multiplier 212 and then onto the
triangle wave generator 214. The phase locked loop reference
frequency is the same frequency as the carrier that is used in the
modulator. Hence, the triangle wave is an integer multiple of the
carrier frequency. This provides a clocking signal for the switch
drivers. The signal from the triangle wave generator creates a
pulsed signal (at the selected integer multiple of the carrier
frequency) as it is passed to the comparator 210 along with the
signal from the error amplifier 208. This signal is then passed to
the power amplifier switch drivers 224 which in turn control the
power amplifier switching devices 216. The switching devices can be
MOSFET switches in a half or full bridge configuration. A feedback
loop 222 is also provided to control the signal gain. After the PWM
power signal has been generated, an output filter 218 is used to
remove the high frequency switching noise. Matching reactive
components 218 are also included which will be discussed later. The
final ultrasonic signal is delivered to the transducer 220 which
emits the composite acoustic waves.
[0068] A pulse width modulation (PWM) amplifier must reproduce an
analog wave containing audio information as accurately as possible.
A perfect sine wave is illustrated by FIG. 9. There are two common
types of PWM architecture. One type of PWM is called bipolar
modulation because the pulses for each part of the sine wave are
created with only two levels. In other words, there is a high level
and a low level but no mid-level. The other type of PWM is unipolar
modulation where only positive pulses referenced to zero are used
for the positive part of the sine wave and only negative pulses
referenced to zero are used for the negative part of the wave being
reproduced. In other words, three states are available.
[0069] FIG. 10 depicts a graph of a sine wave being generated with
bipolar PWM where the pulse count is an integer multiple of three
(3) times the carrier wave. It is important to note that the pulses
are generated symmetrically about the 180-degree point with the
pulses for the negative part of the signal mirroring the positive
part of the signal. In contrast, the unipolar or 3-state pulse
modulation (positive, negative and zero state) only generates
positive pulses for the positive part of the wave and negative
pulses for the negative part of the wave. FIG. 11 illustrates pulse
generation at two (2) times the carrier wave frequency. It should
be noted that although a smooth wave form is shown, when the
carrier is modulated with the audio signal the wave will be
irregularly shaped and the appropriate pulses are delivered to
reproduce a given wave form.
[0070] FIG. 12 illustrates bipolar pulse width modulation (PWM)
where the pulse frequency is 5 times the carrier frequency. FIG. 13
depicts a unipolar pulse width modulation where the PWM frequency
is four (4) times the carrier frequency. The pulse symmetry shown
between the positive and negative parts of the wave helps to cancel
out the even harmonics of the carrier frequency. As described
earlier, using a PWM frequency at an integer multiple of the
carrier frequency solves the problem of unwanted distortion
products within the audible range.
[0071] Another significant problem with parametric audio systems is
that a capacitive transducer is frequently used for output. A
purely capacitive load presents a difficult load to an amplifier. A
large amount of energy is stored in the capacitance. This energy
must be provided by the amplifier. The capacitive impedance varies
widely and cycles unwanted voltage and current throughout the
circuit.
[0072] As illustrated in FIG. 14, a power amplifier parametric
speaker system is coupled to an A/C power source 230 through a
step-up or step-down power transformer 232. Power is supplied to a
bridge rectifier 234 which in turn supplies power to a linear power
amplifier 238 through two storage capacitors for DC energy storage
236a and 236b. Modulation electronics 240 are connected to the
linear power amplifier to supply an ultrasonic signal modulated
with audio related sideband signals. The amplified modulated signal
is supplied to a capacitive transducer 242, which in this case is a
piezoelectric transducer. When the carrier frequency 244 is aligned
with the transducer resonant frequency 246, as depicted in FIG. 15,
an increased level of acoustic output or efficiency gain is
achieved. FIG. 16 illustrates a graph of a capacitive transducer
frequency response (i.e., output in decibels as compared to the
audio frequency being reproduced). The first curve 248 represents
the output curve of the transducer when the carrier is not placed
at the resonant frequency of the transducer. The second curve 250
represents the increased acoustic output of a transducer when the
carrier wave 244 is placed at the resonant frequency of the
transducer. Despite this increased output, the capacitive nature of
the transducer is not changed and the amplifier must dissipate a
significant amount of reactive energy into heat.
[0073] An additional inductance between the amplifier and the
transducer improves this by providing a positive reactance which
counteracts or counterbalances the negative reactance of the
transducer capacitance. FIG. 17 illustrates a parametric
loudspeaker power amplifier to provide improved power efficiency.
The parametric loudspeaker system includes a switching power stage
in the power amplifier 260, and modulation electronics 263 to
supply a signal. Reactive circuitry elements are coupled between
the switching power stage and at least one transducer 264.
Specifically, a series inductance 262 is coupled to the switching
power stage. The switching power stage is preferably a Class D
amplifier. Of course, a linear power amplifier could be used with
the inductor but the Class D amplifier is significantly more
efficient.
[0074] The reactive circuit elements counteract or neutralize the
effects of the reactive part of the transducer load impedance and
increases energy efficiency. The reactive energy storage components
provide reactive power that is stored in the reactive part of the
load impedance. This means the reactive matching network constantly
exchanges reactive energy with the transducer which relieves the
amplifier from having to provide that energy. So, the amplifier
only provides the energy to drive the transducer. In addition, the
reactive elements alternately exchange the reactive energy with the
reactive part of the load impedance. The reactive energy is stored
alternately in the reactive part of the load and then in the
reactive element(s) provided.
[0075] As depicted in FIG. 18, the inductance of FIG. 17 is tuned
to produce a reactance counteracting frequency which is at an
off-resonant frequency 266. This shifts the acoustic output curve
272 from its original output to the acoustic output curve 270
created by tuning the inductor. This is valuable because it creates
greater output at high frequencies which tend to attenuate. The
inductance can also be tuned to move the reactance counteracting
frequency to the resonant 268 frequency of the transducer. An audio
frequency range which carries the audio signal is modulated onto
the carrier frequency and passed to the transducer for
reproduction. FIG. 19 illustrates that an inductance 262 can also
be connected in parallel with the circuit and then tuned to the
carrier frequency, a transducer off-resonant frequency, or the most
preferred frequency of operation.
[0076] Over a narrow band, the transducer capacitance appears to
disappear as a result of the added reactive components. However,
the reactance values of the capacitance and inductance vary
oppositely with frequency, so the reactance damping or cancellation
occurs only over a narrow band. The magnitude of the reactance
actually varies more abruptly than before. In fact, it
theoretically can present a short circuit at the resonant
frequency.
[0077] A better approach is to employ a reactive matching network
comprised of a multiplicity of inductors and capacitors. Such a
network includes a minimum of two inductors and one capacitor. This
multiple element configuration for reactive circuitry elements is
illustrated in FIG. 20. The modulation electronics 280 are coupled
to the switching power amplifier 282 and reactive elements are
coupled between the switching power amplifier and the transducer
290. The reactive elements in this configuration are an inductor in
series 284 and a variable capacitor 286 in parallel. The preferred
tuning for the inductor in this configuration (before the capacitor
is taken into consideration) is to tune the inductor higher or
lower than the resonant frequency of the transducer. Then the
capacitor can be tuned to bring the frequency substantially close
to or on the resonant frequency of the transducer. This provides an
adaptable tuning that can be dynamically adjusted for parameter
changes in the transducer, such as thermal dissipation,
temperature, humidity, or barometric changes. The dynamic tuning
also allows other frequencies to be maximized that are off the
transducer resonant frequencies and for changes to the carrier
frequency, when needed to track substantial changes in the
transducers acoustic resonance. FIG. 21 illustrates the increased
audio output when the frequency is tuned to be near the carrier or
resonant frequency. Using a reactive matching network with a
multiplicity of inductors and capacitors minimizes the power
requirements for the amplifier.
[0078] FIGS. 22 and 23 represent a parametric loudspeaker system
with a predominantly inductive transducer. The modulation
electronics 300 provide the ultrasonic carrier signal modulated
with the audio sideband signal which is delivered to the linear or
Class D power amplifier 302. The signal is delivered to the
inductive transducer 304 through a reactive matching network. This
transducer represents transducers such as electromagnetic voice
coil transducers, magnetostrictive transducers, and similar
transducers which exhibit an inductive load. Resonating
capacitances 306 and 308 are used to tune each circuit to
counteract the reactance in the inductive load at frequencies of
interest. FIG. 22 shows a capacitance 306 in parallel and FIG. 23
shows a capacitance 308 in series. Both capacitances resonate with
the inductive transducer and use the transducer inductance as part
of the reactive matching network. FIG. 24 illustrates the use of
the reactive network to tune the carrier frequency 307 off the
resonant transducer frequency. FIG. 25 illustrates using the
reactive matching network of FIGS. 22 and 23 to tune the signal
frequency 309 to the carrier or resonant transducer frequency.
[0079] In the case of a Class D switching amplifier, the reactive
matching network can be combined with the low pass filter normally
employed to minimize electromagnetic interference (EMI), provide
the proper filtering to minimize switching noise, and allow proper
operation of the PWM function. Due to potential interaction, these
elements would need to work together, even if designed separately.
The reactive matching network can be designed to minimize the
product of maximum voltage and maximum current required at the
output of the amplifier. When the low pass filter is combined with
the reactive matching network, the resulting network filter is
designed to meet the needs of both the PWM network and
counteracting the reactive load.
[0080] As represented by the schematics of FIGS. 26 and 27, the
modulation electronics 310 provide the ultrasonic carrier signal
modulated with the audio related sideband signals which are then
delivered to the power amplifier 312. The signal is delivered to
the transducer 318 through a reactive matching network.
Specifically, both reactive matching networks include at least one
inductor in series 314 and one capacitor 316 in parallel. FIG. 26
includes an additional inductor 320 in parallel. Using groups of
inductors and capacitors to counteract or counterbalance the
reactive load is valuable in optimizing the power requirements of
the amplifier. Multiple inductors and capacitors are preferably
used because they can be tuned to counteract not only reactance for
a narrow frequency range but also reactance over a broader range of
frequencies. Similar to a multi-pole bandpass filter, it is also
possible to tune multiple inductors and capacitors to counteract
reactance even more dramatically over a very narrow range than is
possible with a single inductor. Using a group of capacitors and
inductors to counteract reactance over a broad or very narrow
frequency range optimizes the amplifier's ability to drive the
transducer. In contrast, a single inductor only has limited tuning
or frequency control.
[0081] FIG. 27 includes an additional inductor in series 322 and a
variable capacitor 324 in parallel as part of the reactive network.
Using a variable capacitor allows the frequency tuning to be
adjusted to certain frequencies desired to be reinforced. A
reactive network with multiple elements is especially effective for
a Class D amplifier but it can also be used with other types of
amplifiers where the reactance of the load needs to be
counteracted.
[0082] FIG. 28 depicts a variable saturable inductor which can be
substituted for or added to any of the inductors shown in the
preceding figures. The inductor 330 included in the reactance
circuit has a variable inductance as the power supplied 332 is
changed. As the magnetic field of the permeable core is changed,
the inductance in the reactive circuit fluctuates. A variable
inductor can be controlled electronically and can be connected to a
feedback loop in the circuit to determine what inductance values
will optimize the reactance of the circuit at any given time.
[0083] FIGS. 29-32 illustrate several configurations for reactive
matching networks. Each of FIGS. 29-32 include modulation
electronics 310 to provide the ultrasonic carrier signal modulated
with the audio sideband signals (single sideband or double
sideband), which are then delivered to the switching power
amplifier 312. The signal is delivered to the transducer 318
through a reactive matching network. Each of these circuits
includes at least one series inductor 314 and one parallel
capacitor 316. FIG. 29 shows the use of an additional inductor 340
in series with the other elements of the reactive matching network.
In addition, the carrier frequency can be varied to accommodate
changes the resonant frequency of the transducer that are related
to changes in barometric pressure and altitude. The carrier
frequency can be varied in order to maximize efficiency and output
levels for parametric loudspeaker output.
[0084] Some or all of the required inductance can also be included
in the matching transformer if desired. FIG. 30 shows a step-up or
step-down transformer 342 which may include a portion of the
desired inductance within the transformer. The inductance of the
matching transformer is tuned to work with the other elements in
the reactive matching network. FIG. 31 is a variation of FIG. 30
where the inductance 346 has been moved out of a transformer 344 to
the transducer side of the transformer. FIG. 32 shows an inductor
314 and capacitor 316 used for reactive matching and a transformer
356 to scale the voltage. Additional elements of the reactive
network between the transformer and the transducer are two
inductors 350 and 352 and a resonant capacitor 354.
[0085] Another issue with parametric transducers is the isolation
of the AC power line from the load circuit. This large amount of
power can be dangerous to users if the high voltage leads from the
amplifier are exposed because a consumer might be seriously
shocked. Conventional amplifier systems isolate the power amplifier
from the AC power lines using an expensive power isolation
transformer before the bridge rectifier. FIG. 33 illustrates a
power amplifier configuration for isolating the power amplifier
without using an expensive power isolation transformer. A
parametric signal from the modulation electronics 360 is passed
through a first isolation transformer 362 to the power amplifier
364. The power to the power amplifier is supplied by input power
circuitry which is a bridge rectifier 370 connected to the AC power
line 372. Then the amplified signal is passed from the power
amplifier through a second isolation transformer 366 to the output
transducer. This way the power amplifier and the two isolation
transformers can be sealed in the same container, which is
inaccessible to the user. The second transformer 366 can be a
step-up, step-down or 1:1 transformer depending on the matching
needs of the transducer such as electrostatic, piezoelectric or
other transducer types. Isolating the power amplifier using two
smaller transformers minimizes the cost and size of a parametric
power amplifier by eliminating the typically larger and more costly
power transformer or power isolation system. Another embodiment of
this isolation configuration is to use an opto-isolator in the
place of the input transformer.
[0086] One switching power amplifier embodiment uses a switched
amplitude or multi-voltage level power amplifier. An ultrasonic
carrier frequency is modulated with audio to generate sideband
signals, which along with the carrier are amplified by the
switching power stage. An ultrasonic transducer is connected to the
output of the power amplifier. A feature of the multi-level
switching power stage is that it has multiple power supplies or
transformer taps (see FIG. 35) and uses amplitude power switching
to provide controlled power to, and minimize dissipation in, the
output stage of an amplifier. A multi-level switching amplifier
preferably includes at least two switchable power delivery levels
per polarity.
[0087] One embodiment of the multi-level switching amplifier for
parametric loudspeakers is shown in FIG. 34. A power stage is
illustrated that includes multiple level voltage supplies 380, 382,
for one polarity and 384, 386 for the opposite polarity, in
combination with the energy storage supply capacitors 390, 392,
394, 396.
At low program signal levels, transistors 402 and 404 operate the
"inner" supplies 382 and 384 drawing current from supply capacitors
392 and 394. At higher levels, transistors 400 and 406 are switched
on and the diodes 408 and 409 are reverse biased which terminates
current flow through the circuit path that includes the diodes.
This multi-level approach provides greater efficiency and reduced
dissipation in the power stages. The outputs 410 from the switching
power stage 490 can be coupled to a transducer. It should also be
mentioned that the two power supplies for each polarity in FIG. 34
can also be three or more power supplies or two or more taps on an
output coupling transformer for each polarity.
[0088] Another embodiment of a multi-level switching amplifier for
parametric loudspeakers is shown in FIG. 35. A transformer 450 is
coupled to an amplifier with multiple level taps 452, 454, for one
polarity and 458, 460, for the opposite polarity, and 456 as a
center tap. This configuration provides output and efficiency
advantages comparable to the system of FIG. 34, but only requires
the use of a single energy storage supply capacitor 462.
[0089] At lower program signal levels, transistors 470 and 472
(shown as MOSFETS) operate through lower level transformer taps 452
and 460 drawing current from supply capacitor 462. At higher
levels, transistors 474 and 476 are switched on and transistors 470
and 472 are switched off and diodes 478 and 480 are reverse-biased
which terminates current flow through the circuit path that
includes the diodes. This draws current from the two higher level
voltage taps 454, 458. The outputs 482a and 482b from the
transformer can be coupled to a transducer. This multi-level
approach provides greater efficiency and reduced dissipation in the
power stage 491.
In the multi-level power supply embodiments of FIG. 34, the 4 sets
of power supplies can be replaced with 4 sets of taps on a
transformer output. This replaces the power supplies in the same
manner as the four supplies of FIG. 34 are replaced with the four
taps of FIG. 35.
[0090] The systems discussed relative to FIG. 34 may have the low
levels set by the power supply voltages of 382 and 384 and FIG. 35
may have the low level set by the ratio of turns of the lower taps
454 and 458 compared to the upper taps 452 and 460. The selection
of power levels may be chosen based on a number of parameters such
as the predicted average levels of operation, the peak to average
ratio of the program material, etc. A preferred embodiment is to
set the high to low voltage ratios to be two to one. In addition,
the multiple power level configurations can be combined with the
power amplifier configuration for isolating the power amplifier
without using a power isolation transformer. The combination of
both of these configurations provides multiple power levels that
are isolated from user access, reduced size, reduced weight and
reduced cost.
[0091] With the systems shown in FIGS. 34 and 35, significant
improvements in power amplifier efficiency can be achieved over
prior art parametric loudspeaker power amplifiers.
[0092] Even more complex multiple power level, switched amplitude
embodiments can be designed, such as the implementation of a
multi-bit power amplifier. For example, there may be 4 separate
power supplies for each polarity that produce 2 times 4 power
supply levels, but when used in a multi-bit approach they provide
two times 16 power supply levels. These power supplies can be used
in various combinations to provide different power levels. In a
multi-bit power supply, there can be four power supplies which each
have a power switch (e.g., a MOSFET). Any number of these switches
can be on or off in various combinations so that there are 2.sup.4
or 16 power supply levels. This is advantageous because there can
be up to 2.sup.N power supply levels with only N actual power
supplies. Multiple voltage levels keep the voltage across the
output stage at a minimum voltage and minimize the power
dissipation in the amplifier. Each power supply can be twice the
voltage of the preceding power supply. So if there are 4 power
supplies, their voltages will be N, 2N, 4N and 8N volts (e.g., 10V,
20V, 40V and 80V). Furthermore, two positive and two negative power
supply levels can be used.
[0093] It should be noted that to those skilled in the art, the
term switching power amplifier refers to the many different
approaches and names that are known in the switch-mode power
conversion art that include but are not limited to Class D, Class
AD, Class BD, two state amplifiers, three state amplifiers, digital
power amplifiers, pulse width modulation (PWM), pulse duration
modulation (PDM), switched amplitude amplifier, signal tracking
amplifiers, class "G", class "H", multi-bit, and switch-mode power
amplifiers.
[0094] The power amplifier embodiments of the invention as
disclosed have significant efficiency improvements and/or size and
cost reductions when compared to prior art parametric loudspeaker
power amplifiers.
[0095] It is to be understood that the above-described arrangements
are only illustrative of the application of the principles of the
present invention. Numerous modifications and alternative
arrangements may be devised by those skilled in the art without
departing from the spirit and scope of the present invention. The
appended claims are intended to cover such modifications and
arrangements. Thus, while the present invention has been shown in
the drawings and fully described above with particularity and
detail in connection with what is presently deemed to be the most
practical and preferred embodiment(s) of the invention, it will be
apparent to those of ordinary skill in the art that numerous
modifications, including, but not limited to, variations in size,
materials, shape, form, function and manner of operation, assembly
and use may be made, without departing from the principles and
concepts of the invention as set forth in the claims.
* * * * *