VERSATILE AUDIO POWER AMPLIFIER
BACKGROUND
[0001] This disclosure relates to a versatile audio power amplifier.
[0002] Switching amplifiers, also called Class D amplifiers, amplify an input signal
by modulating that signal into a series of pulses that drive a complementary pair of
transistors operated in the switching mode. The transistors alternately couple positive
and negative power sources to the output, which in total produce an amplified
representation of the input signal.
SUMMARY
[0003] In general, in some aspects, an audio power amplifier includes a first and a
second amplification unit. Each amplification unit includes a switching voltage
amplifier having a command signal input and an amplified signal output, an output
filter between the amplified signal output and a load terminal, a current compensator
with a current-compensated command signal output coupled to the command signal
input of the voltage amplifier, an inner current feedback loop feeding a measurement
of current measured at the output inductor back to a summing input of the current
compensator, a voltage compensator with a voltage-compensated control signal output
coupled to the summing input of the current compensator, and an outer voltage
feedback loop feeding voltage at the load terminal back to a summing input of the
voltage compensator. A first controlled signal path provides the voltage-compensated
control signal output of the voltage compensator of the first amplification unit to the
summing input of the current compensator of the second amplification unit. The
second amplification unit uses the voltage-compensated control signal of the first
amplification unit as input to the current compensator of the second amplification unit
in place of the voltage-compensated command signal of the second amplification unit
when the first controlled signal path is activated. Control electronics provide signal
inputs to the first and second amplification units and control the first controlled signal
path such that the first and second amplification units are operable with separate
loads, in parallel driving a common load, or across a bridge-tied-load.
[0004] Implementations may include one or more of the following features. The first
and second amplification units may be operable with separate loads by each
amplifying separate signals and providing the amplified signals on their separate
output terminals. The first and second amplification units may be operable in parallel
driving a common load by each amplifying the same signal, provided from the first
amplification unit to the second amplification unit via the first controlled signal path,
and providing identical amplified signals on their separate output terminals, which are
to be coupled to a common input terminal of the load. The first and second
amplification units may be operable across a bridge-tied-load by amplifying a first
signal in the first amplification unit and amplifying an inverted copy of the first signal
in the second amplification unit, and providing their respective amplified signals on
their separate output terminals, which are to be coupled to separate input terminals of
the load.
[0005] A third and a fourth amplification unit identical to the first and second
amplification units, and a second controlled signal path from the voltage-compensated
control signal output of the voltage compensator of the third amplification unit to the
summing input of the current compensator of the fourth amplification unit may be
included, the control electronics further providing signal inputs to the third and fourth
amplification units and controlling the second controlled signal path such that the
third and fourth amplification units are operable with separate loads, in parallel
driving a common load, or across a bridge-tied-load, and all four of the amplification
units are operable together with the first and second units in parallel driving a first
side of a bridge-tied-load, and the third and fourth units in parallel driving a second
side of the bridge-tied-load.
[0006] The four amplification units may be operable together by amplifying a first
signal in each of the first and second amplification units, provided from the first
amplification unit to the second amplification unit via the first controlled signal path,
and providing identical amplified first signals on the separate output terminals of the
first amplification unit and the second amplification unit, amplifying an inverted copy
of the first signal in each of the third and fourth amplification units, provided from the
third amplification unit to the fourth amplification unit via the second controlled
signal path, and providing identical amplified inverted first signals on the separate
output terminals of the third amplification unit and the fourth amplification unit, the
output terminals of the first and second amplification units are to be coupled to a first
input of the load, and the output terminals of the third and fourth amplification units
are to be coupled to a second input of the load.
[0007] The amplifier may use a four-quadrant power supply having a synchronous
output rectifier. The synchronous output rectifier may include a MOSFET. The first
controlled signal path may include a switch controlled by the control electronics. The
switching voltage amplifiers may each include a modulator, a gate driver, a pair of
transistors, and a pair of diodes coupled between the source and drain terminals of the
transistors. The transistors may include MOSFETS, the diodes being intrinsic to the
MOSFETS. The output filter may include an output inductor and the measured
current may be the current through the output inductor.
[0008] In general, in some aspects, amplifying audio-frequency signals includes, in
each of a first and a second amplification unit, amplifying a current-compensated
command signal in a switching voltage amplifier to provide an amplified signal
output, measuring current through an output filter between the amplified signal output
of the voltage amplifier and a load terminal to produce a current measurement,
feeding back the current measurement to a summing input of a current compensator
via an inner current feedback loop, at the current compensator, comparing the current
measurement to a voltage-compensated command signal and providing the currentcompensated
command signal to the voltage amplifier, feeding back voltage at a load
terminal of the amplification unit to a summing input of a voltage compensator via an
outer voltage feedback loop, and at the voltage compensator, comparing the feedback
voltage to an input command signal and providing the voltage-compensated command
signal to the summing input of the current compensator. A first controlled signal path
from an output of the voltage compensator of the first amplification unit to the
summing input of the current compensator of the second amplification unit is
controlled to selectively provide the voltage-compensated command signal of the first
amplification unit to the summing input of the second amplification unit in place of
the voltage-compensated command signal of the second amplification unit. Signal
inputs are provided to the first and second amplification units and the first controlled
signal path is controlled to selectively operate the first and second amplification units
with separate loads, in parallel driving a common load, or across a bridge-tied-load.
[0009] Advantages include comprehensive configurability with high efficiency. The
amplifier can serve a wide variety of connection topologies, load impedances, and
power levels without hardware modification. Being able to drive loudspeakers at a
wide range of impedances from a single amplifier allows the amplifier to support a
diverse range of audio system configurations without requiring a diverse set of
amplifier products.
[0010] Other features and advantages will be apparent from the description and the
claims.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] Figures 1A through ID show block diagrams of amplifier-speaker topologies.
[0012] Figure 2 shows a circuit diagram of a single amplifier stage.
[0013] Figure 3 shows a circuit diagram of two amplifier stages in a bridged
configuration.
[0014] Figure 4 shows a circuit diagram of two amplifier stages in a parallel
configuration.
[0015] Figure 5 shows a circuit diagram of four amplifier stages in a parallel-bridged
configuration.
[0016] Figure 6 shows a circuit diagram of a configurable amplifier system.
[0017] Figure 7 shows a circuit diagram of an isolation converter.
[0018] Figures 8A through 8C show power flow through the isolation converter of
figure 7.
[0019] Figure 9A shows a circuit diagram of a representative class D amplifier.
[0020] Figures 9B and 9C show energy flow through the amplifier of figure 9A.
DESCRIPTION
[0021] Power amplifiers may be connected to their loads in a number of topologies,
depending on the nature and intended use of the loads. Figures 1A through ID show
four topologies for connecting power amplifiers to loudspeakers. In figure 1A, a
single amplifier 10 drives a loudspeaker 20 by providing power to one terminal of the
loudspeaker, while the other terminal is grounded. This is a typical configuration
sometimes referred to as a "half-bridge." In figure IB, a "full-bridge" configuration is
shown, where two amplifiers 10 and 12 are used, one connected to each terminal of
the loudspeaker 22. The second amplifier 12 is driven with an inverse of the signal to
the first amplifier 10, so the total voltage across the loudspeaker 22 is doubled, while
the current remains the same as that in the half-bridge configuration. By providing
twice the voltage at the same current, this configuration can drive loudspeakers with
larger impedances than can be driven by the half-bridge. It can provide more power
total, or the same power with less dissipation per amplifier. This mode of operation is
ubiquitous in the audio amplifier field, and is often referred to as 'BTL' (bridge tied
load) configuration.
[0022] In figure 1C, the two amplifiers 10 and 12 are connected in parallel to a
common terminal on the loudspeaker 24, while the other terminal is grounded. This is
referred to as a "parallel" configuration. The parallel configuration delivers twice the
current at the same voltage as the half-bridge configuration, which is useful for
driving smaller impedances at the same power level as the BTL configuration. For
example, if the half-bridge is optimized to provide 500 W to a 4 W load, the current
required to provide the same power into a 2 W load or the voltage required to provide
the same power to an 8 W load may be at or beyond the limits of the amplifier. The
parallel configuration can drive 500 W into a 2 W load with half the dissipation of the
single half-bridge, or drive a full 1 kW if each half-bridge can handle the current. The
BTL configuration, on the other hand, can drive 500 W into an 8 W load without
approaching its voltage limits, or drive a full 1 kW if the voltages are available.
[0023] Finally, in figure ID, four amplifiers 10, 12, 14, and 16 are used, with a
parallel pair of amplifiers connected to each terminal of the loudspeaker 26. One pair,
10 and 12, is driven with the opposite signal of the other pair, 14 and 16. This is
referred to as a parallel-bridged configuration, and delivers twice the voltage and
twice the current as a single half-bridge, for four times the power. Using the same
example as above, if each half-bridge is optimized for 500 W at 4 W, the parallelbridged
configuration can deliver 2 kW to a 4 W load, with the same voltage and
current per amplifier stage.
[0024] With a class D amplifier, management of current- sharing between two
parallel amplifiers is more important than in linear, or class AB, amplifiers, because
the dissipated power in a switching device is proportional to I , rather than to I, as it is
in a linear amplifier. Sharing current between two identical devices will decrease
conduction loss by roughly a factor of two in the switching amplifier. The decrease
will actually be a bit more than a factor of two because there are further gains due to
the temperature coefficient of resistance of the FET—decreasing the current also
decreases the temperature, which in turn decreases the intrinsic resistance of the
device. If the current is not controlled, however, it is likely that one of the two devices
will deliver substantially more current than the other, losing the benefits of parallel
operation and possibly damaging the amplifiers.
[0025] To provide efficient current sharing in topologies, a feedback loop may be
added to the amplifier. For example, as shown in figure 2, a unit cell 100 provides one
half-bridge class D audio amplifier, shown connected to an arbitrary loudspeaker 120.
The core of the amplifier includes a modulator and gate driver 102, a switching
power-output stage made up of transistors 104 and 106, and an output filter including
an inductor 108 and capacitor 110. The control system includes an outer voltage loop
112 that feeds back the output voltage at the load 120 to a summer 114 and voltageloop
compensator 116. The summer 114 also receives the input voltage command V
in. The modulator and gate driver 102, in combination with the transistors 104 and
106, constitute a voltage amplifier.
[0026] To allow current sharing when two of these unit cells 100 are connected in
parallel, an inner current loop 130 is provided. The inner current loop 130 feeds back
a measurement of output current, from a current sense 130a, to another summer 132
and current-loop compensator 134. The inner current loop controls the output current
of the amplifier, so that two amplifiers operating in parallel will each provide half the
total current; neither will attempt to deliver all the current and lose the advantages of
parallel operation. In this configuration, the current loop around the core voltage
amplifier turns the system into a current amplifier, and the outer voltage loop turns the
entire unit cell 100 back into a voltage amplifier.
[0027] The inner current loop provides some additional advantages. The current loop
naturally provides current limiting within the unit cell. That is, the feedback 130 to the
current-loop compensator 134 prevents the command into the modulator 102 from
causing a current gain in excess of the maximum current command from the voltageloop
compensator 116, even if the load is shorted. Additionally, because the inner
current loop provides control, the voltage measurement used for the outer voltage
loop can be moved outside the output filter, closer to the load (as shown). (An output
filter typically imposes a 180° phase shift, around which a control loop could not be
closed.) Moving the voltage loop to after the output filter allows the amplifier to
support a greater variety of loads while the inner current loop maintains stability. The
inner current loop 130 can also be used to provide pulse width error correction in the
modulator 102, as described in U.S. Patent application , titled
"Reducing Pulse Error Distortion," and filed the same day as this application, the
entire contents of which are incorporated here by reference.
[0028] The summers 114 and 132 are not necessarily discrete components, but may
be, for example, summing inputs of the compensators 116 and 134. The compensators
preferably are built from standard circuit components, i.e., op-amps and associated
circuitry. The current-sensing element 130a may be any standard current-sensing
technology, such as discrete Hall-effect sensors. The output inductor 108 may be
formed using planar windings on a printed circuit board, and part of the current sense
is provided by a current sense winding integrated into the output inductor 108, as
described in U.S. Patent application , titled "Planar Amplifier Output
Inductor with Current Sense," and filed the same day as this application, the entire
contents of which are incorporated here by reference. As explained in that patent
application, forming the current- sense winding as part of the PCB windings of the
output inductor advantageously shields the current- sense signal from noise within the
inductor. The inductor's current-sensing winding indicates an AC component of the
current, while a Hall-effect sensor may also be used to indicate the DC component of
the current.
[0029] In some examples, these unit cells are combined in groups of 4. Each group
has appropriate interconnections within the control system, allowing a single set of
four unit cells to provide any of the topologies shown in figures 1A through ID. The
user, through control software, for example, may specify the particular topology
needed. To detect connection problems and confirm that the topology of connected
loudspeakers matches the configuration of the amplifier components, an amplifier
product may include circuitry for detecting the type and topology of the connected
loudspeakers. One such system is described in U.S. Patent application 12/114,265,
titled "Detecting a Loudspeaker Configuration," filed May 2, 2008, the entire contents
of which are incorporated here by reference. Such a system may also be used to
discover the topology of the connected loudspeakers and automatically configure the
amplifier accordingly.
[0030] For independent half-bridge operation of multiple channels, each unit cell is
connected to one loudspeaker as shown in figure 2, and separate signals are provided
to each unit cell.
[0031] As shown in figure 3, two half bridges 100a and 100b may be combined into
a full bridge configured to drive a single loudspeaker 120b as a bridge-tied-load to
provide double the voltage of a single unit cell. In this configuration, the amplifiers in
the unit cells are substantially independent, and are simply given input commands V
i 180° out of phase, i.e., +Vc_i and - Vc -i . No modification to the control of either
unit cell is needed for BTL operation, though a product intending to support this mode
may handle inverting the V _i input, rather than relying on the user to provide both
the original and inverted signal. Inverting of the V _i signal may be done in the
control electronics (not shown), or by an additional inverting amplifier (not shown),
either under the control of control electronics or directly controlled by a physical
switch available to the user.
[0032] As shown in figure 4, two half bridges 100a and 100b may be combined as a
parallel pair to provide the same voltage and twice the current of a single unit cell. In
this configuration, one outer voltage loop 112a is configured to feed commands,
through the first unit's outer voltage loop summer 114a and compensator 116a, to the
inner current loops and amplifying stages of both unit cells through a cross-cell
connection 202 controlled by a switch 204. In this configuration the second outer
voltage loop 112b and its summer 114b and compensator 116b are not used—they
may be entirely deactivated, or the signal path from the compensator 116b to the
summer 132b may be interrupted. Both half-bridge outputs are coupled to a common
input of the loudspeaker 120c, with the other input grounded. The current loops 130a
and 130b, summers 132a and 132b, and compensators 134a and 134b control current
sharing between the half bridges by stabilizing each at the target current, as discussed
above. The switch 204 may be controlled in various ways, including, for example, by
control electronics, by passive circuitry, or by a physical switch available to the user.
[0033] As shown in figure 5, four half-bridges 100a, 100b, 100c, and lOOd may be
used together to provide a bridged-parallel configuration delivering double the voltage
and double the current of a single unit cell to a single loudspeaker 120d. Half-bridges
100a and 100b are configured as a first parallel pair with a cross-cell connection 202a
and switch 204a, and coupled to a first input of the loudspeaker 120d. Half-bridges
100c and lOOd are configured as a second parallel pair with a cross-cell connection
202b and switch 204b, and coupled to the second loudspeaker input. The second pair
lOOc/lOOd are given an inverted input signal -Vc-i as in the BTL configuration of
figure 3.
[0034] In some embodiments, the control circuitry of each of the half-bridge unit
cells is independent, such that when two cells are used in the BTL or parallel
configuration, the other two may be used as independent half-bridges, in the same
two-cell configuration as the first two cells, or in the other two-cell configuration. In
some examples, pairs or all four of the amplifier stages (modulator and gate drive) are
provided in a single integrated circuit package, such as the TDA8932 from NXP
Semiconductors, in Eindhoven, The Netherlands, or the TAS5103 from Texas
Instruments in Dallas, TX, while the transistors, control loops, and output filters are
added to complete the amplifiers and enable the configurability described above. In
some examples, the control loops and cross-cell connections are included in the
amplifier IC.
[0035] As shown in figure 6, groups of unit cells can be combined, up to the limits of
the power supply, to form highly configurable systems. In figure 6, a control module
210 is shown coupled to a number of switches identified below. Dotted lines show
control signal paths while solid lines show audio signal paths. Two sets of four halfbridge
unit cells are shown, numbered 100a through lOOh. The connections in the
second set lOOe-lOOh are identical to those in the first set lOOa-lOOd, though they are
shown with the switches in different positions. Eight inputs A through H are
available, but not all are used. The switch positions in figure 6 are set to show the first
two unit cells 100a and 100b each providing their respective inputs A and B to
separate loudspeakers 120a, unit cells 100c and lOOd providing input C to a single
loudspeaker 120b in a BTL configuration, and unit cells lOOe, lOOf, lOOg, and lOOh
together providing input E to a loudspeaker 120d in a parallel -bridged configuration,
with each pair lOOe/lOOf and lOOg/lOOh powering one input of the loudspeaker in
parallel.
[0036] For the two unit cells 100a and 100b being operated as independent halfbridges,
a first switch 204a controls the signal path between the two unit cells 100a
and 100b, for providing common current control when operating in parallel, as
discussed above. In the example of figure 6, switch 204a is open, because unit cells
100a and 100b are acting separately. Another switch 212 controls which signal is
input to the unit cell 100b. For half-bridge operation, as shown, switch 212 couples
the signal input "B" to the unit cell 100b.
[0037] For BTL operation, an inverter 214 is available to couple an inverted copy of
input signal "A" to the unit cell 100b, where the switch 212 would provide that signal
rather than the input "B" used for half-bridge operation. This is the case in the second
pair of unit cells 100c and lOOd, where the switch 204b is open, and a switch 216 is
coupling an inverter 218 to the input of unit cell lOOd, providing an inverted copy of
input "C". Another inverter 220 and input switch 222 control the input to unit cell
100c, for use in parallel-bridged configurations, discussed below with reference to the
second set of four unit cells. In figure 6, the input switch 222 is coupling the input
"C" to the unit cell 100c.
[0038] Unit cells lOOe and lOOf are shown configured for parallel operation. In this
mode, the input switch 224 on unit cell lOOf does not provide any signal input to the
second unit cell, as the closed switch 204c provides the current command signal from
inside the unit cell lOOe to the current feedback loop comparator in unit cell lOOf,
skipping the voltage command input and voltage feedback loop comparator of unit
cell lOOf. In some configurations, the input of unit cell lOOf is disconnected internally
when used in parallel operation, so the input switch 224 may remain coupled to one of
its inputs. The inverter 226 is available to provide an inverted version of input "E" for
use in BTL mode. Unit cells lOOg and lOOh are also shown configured for parallel
operation, and in particular, they are configured for use in a parallel-bridged mode
with unit cells lOOe and lOOf. Switch 204d provides the current command signal to
unit cell lOOh from inside unit cell lOOg, while switch 228 is open and inverter 230 is
unused. At unit cell lOOg, a switch 232 couples an inverted copy of input "E" from an
inverter 234 to the input of unit cell lOOg, so that the two sets of parallel half -bridges
lOOe/lOOf and lOOg/lOOh receive E and -E, respectively, as inputs.
[0039] Although the switches, inverters, and signal sources are shown external to the
control module 210, some or all of the switches and inverters may be integrated into
the control module, and may be configured in hardware, firmware, or software,
depending on the technology used. If the switches and inverters are integrated to the
control module, the inputs A through H would also pass through the control module.
The control module may be any suitable device, such as a programmed
microprocessor, an application-specific integrated circuit, or a collection of discrete
devices. The control module may also apply digital signal processing to the inputs A
through H, in addition to the switching and inverting used to configure the amplifier
topology. There may be a separate control module for each set of four unit cells. In
some cases, the control module may be configured to use all eight unit cells in a
cascading pattern of bridged and parallel groupings to deliver the entire capacity of
the power supply to one loudspeaker (additional connections between amplifiers
and/or signal paths would be needed, though these may all be provided by a suitable
control module).
[0040] Such a configuration is capable of driving numerous configurations of
loudspeakers. For example, an 4000 W amplifier system containing two groups of
four unit cells, at 500 W each, can drive many different combinations of speakers, as
shown in table 1 (with the actual wattage depending on the impedance of the
particular loudspeakers used).
[0041] One or more of these groups of four unit cells are combined with an
appropriate power supply, such as that described in co-pending U.S. Patent
application , titled "Power Supply Transient Response Improving,"
and filed the same day as this application, the entire contents of which are
incorporated here by reference. In some examples, an advantageous feature provided
within the power supply is a synchronous output rectifier, which solves the 'bus
pumping' problem generally associated with class D half-bridges, and supports
efficient operation of the highly-configurable amplifier stages described above (the
bus-pumping problem is explained for reference below). A schematic representation
of this synchronous output rectifier 300 is shown in figure 7. A primary winding 302
is coupled to switches 304, 306, 308, and 310, and a secondary winding 312 is
coupled to switches 314, 316, 318, and 320. The windings are separated by an
isolation barrier 322 and convert the +400 V supply voltage to the lower voltage used
by the amplifiers, +/-80 V in this example. Such a rectifier powers the amplifier
stages described above with the +80 V and -80 V ports coupled to the +V and - V
rails of the amplifiers.
[0042] At the voltage levels typically used in an audio power amplifier, e.g., +/- 80V
or higher on the secondary windings, the secondary side switches 314-320 in this
topology would typically be simple rectifiers. Instead, MOSFETs are used with their
intrinsic diodes to provide a synchronous rectifier. This allows power to flow in either
direction at any of the three ports of this power converter (+400 V primary, +80 V
secondary, -80 V secondary), providing for full four-quadrant operation, as explained
in figures 8A through 8C.
[0043] Figure 8A shows one half-cycle of normal operation, in which power is
transferred from the primary winding 302 to the secondary winding 312. In this cycle,
transistors 304 and 310 are on, so that current (shown by dotted arrows) flows from
the +400 V rail 322a to the primary-side ground 324a by going through the primary
winding 302 in a first direction, illustrated as downward. This induces current to flow
in the opposite direction, illustrated as upward, in the secondary winding 312, due to
the direction of the windings. That current flows from the -80 V rail 330a to the
secondary-side ground 326 and from the secondary-side ground 326 to the +80 V rail
328a through the transistors 314 and 320. As noted above, the use of MOSFETs
provides intrinsic diodes across the source and drain of the transistors.
[0044] Figure 8B shows the other normal half-cycle, in which the transistors 306 and
308 are on, so that current flows from the +400 V rail 322b to the primary-side
ground 324b by going through the primary winding 302 in the opposite direction, i.e.,
upward in the figure. The induced current in the secondary winding 312 then goes
downward, and through the other set of secondary-side transistors 316 and 318, to
again flow from the -80 V rail 330b to ground 326 and from ground 326 to the +80 V
rail 328b, producing the same net power flow as in figure 8A.
[0045] Figure 8C shows a different case, in which energy is being sourced from the
+80 V rail, and flows into both the -80 V and +400 V rails. Note that the only
difference between this and the case in figure 8B is the direction of current from the
+80 V and +400 V rails. At transistors 306, 308, and 316, current is flowing in the
direction of the intrinsic diodes. At the +80 V 328b rail, however, current is flowing
against the direction of transistor 318, so that transistor must be switched on. In a
standard power converter with a simple rectifier between the +80 V rails and the
transformer secondary, energy could not flow in this direction. The use of MOSFETs
in the secondary allows current flow in this direction. We have not drawn all possible
cases, but it can be seen that the use of synchronous rectifiers on the secondary side
allows energy to flow between any port and any other port of this system, allowing
four-quadrant operation.
[0046] At lower output voltages, when the voltage drop across the rectifiers becomes
a significant source of inefficiency, synchronous rectifiers are in common use today.
However, they are not commonly used at the high voltages of audio power amplifiers,
as the efficiency increase is very marginal, and there are significant technical hurdles
to overcome to make the MOSFETs robust with reverse current flow and parasitic
diode recovery. Using the technique in the system described above allows energy
flow from reactive loads back into the system to be evenly distributed between all of
the storage capacitance in the system, and it also solves the bus pumping problem
associated with half-bridge class D stages.
[0047] Bus pumping, also known as off-side charging or rail pumping, is explained
with reference to figures 9A-9C. As shown in figure 9A, a half-bridge class D
amplifier 400 operates by switching the output inductor 416 between two voltage rails
402 and 404, labeled here is generic +V and -V. The modulator and gate drive are not
shown, and the filter capacitors 422 and 424 are the only part of the power supply
shown. The switching of transistors 412 and 414 occurs at a rate significantly higher
than the signals being reproduced, so that at any given time we can treat the current in
the output inductor 416 as relatively constant. Thus, for the state when current is
flowing out of the inductor 416 and into the load 420, there are two switching states.
[0048] In the first state, as shown in figure 9B, the top transistor 412 is on, and
current (shown by the dotted arrow) flows from the +V rail 402. In this case, energy is
flowing from the power supply to the output filter and the load.
[0049] In the second state, as shown in figure 9C, the bottom transistor 414 is on,
and current flows from the - V rail 404. In this case, energy is actually flowing back
from the output filter and the load, and into the power supply. If the power supply
filter capacitances 422 and 424 are large enough, they can absorb this energy for the
duration of the event, but this would require unusually large capacitances.
[0050] The expression for average current regenerated into the rail over a half sine
wave is as follows:
where ' ' is average power out, V is rail voltage, and V is load resistance. For
example, a 20 Hz sine wave into a 4 W load, of sufficient amplitude to give 500 W
average power output, in an amplifier with +80 V rails, will regenerate an average of
3A over a half cycle. With 10 000 m of capacitance on that rail, the system will
experience a bus pumping of 7.5 V, which is almost 10% of the nominal. By using
synchronous rectification, lower capacitances can be used and the system will still
experience significantly less perturbation. In one example, with the above-mentioned
voltages, a system with only 3500 m capacitance per channel per rail experienced
significantly less pumping of the bus than that just described.
[0051] Other implementations are within the scope of the following claims and other
claims to which the applicant may be entitled.
P-09-013-WO
WHATIS CLAIMED IS:
1. An audio power amplifier comprising:
a first and a second amplification unit, each amplification unit comprising:
a switching voltage amplifier having a command signal input and an amplified
signal output,
an output filter between the amplified signal output and a load terminal,
a current compensator with a current-compensated command signal output
coupled to the command signal input of the voltage amplifier,
an inner current feedback loop feeding a measurement of current measured at
the output inductor back to a summing input of the current compensator,
a voltage compensator with a voltage-compensated command signal output
coupled to the summing input of the current compensator, and
an outer voltage feedback loop feeding voltage at the load terminal back to a
summing input of the voltage compensator;
a first controlled signal path from the voltage-compensated control signal output
of the voltage compensator of the first amplification unit to the summing input
of the current compensator of the second amplification unit,
the second amplification unit configured to use the voltage-compensated
command signal of the first amplification unit as input to the current
compensator of the second amplification unit in place of the voltagecompensated
command signal of the second amplification unit when the
first controlled signal path is activated;
control electronics providing signal inputs to the first and second amplification
units and controlling the first controlled signal path such that the first and
second amplification units are operable with separate loads, in parallel driving
a common load, or across a bridge-tied-load.
2. The audio power amplifier of claim 1 wherein the first and second amplification
units are operable with separate loads by each amplifying separate signals and
providing the amplified signals on their separate output terminals.
P-09-013-WO
3. The audio power amplifier of claim 1 wherein the first and second amplification
units are operable in parallel driving a common load by each amplifying the
same signal, provided from the first amplification unit to the second
amplification unit via the first controlled signal path, and providing identical
amplified signals on their separate output terminals, which are to be coupled
to a common input terminal of the load.
4. The audio power amplifier of claim 1 wherein the first and second amplification
units are operable across a bridge-tied-load by amplifying a first signal in the
first amplification unit and amplifying an inverted copy of the first signal in
the second amplification unit, and providing their respective amplified signals
on their separate output terminals, which are to be coupled to separate input
terminals of the load.
5. The audio power amplifier of claim 1 further comprising:
a third and a fourth amplification unit identical to the first and second
amplification units, and
a second controlled signal path from the voltage-compensated command signal
output of the voltage compensator of the third amplification unit to the
summing input of the current compensator of the fourth amplification unit,
the control electronics further providing signal inputs to the third and fourth
amplification units and controlling the second controlled signal path such that:
the third and fourth amplification units are operable with separate loads, in
parallel driving a common load, or across a bridge-tied-load, and
all four of the amplification units are operable together with the first and
second units in parallel driving a first side of a bridge-tied-load, and the
third and fourth units in parallel driving a second side of the bridge-tiedload
6. The audio power amplifier of claim 5 wherein the four amplification units are
operable together by:
amplifying a first signal in each of the first and second amplification units,
provided from the first amplification unit to the second amplification unit via
the first controlled signal path, and providing identical amplified first signals
on the separate output terminals of the first amplification unit and the second
amplification unit,
amplifying an inverted copy of the first signal in each of the third and fourth
amplification units, provided from the third amplification unit to the fourth
amplification unit via the second controlled signal path, and providing
identical amplified inverted first signals on the separate output terminals of
the third amplification unit and the fourth amplification unit,
wherein the output terminals of the first and second amplification units are to be
coupled to a first input of the load, and the output terminals of the third and
fourth amplification units are to be coupled to a second input of the load.
The audio power amplifier of claim 5 further comprising a four-quadrant power
supply having a synchronous output rectifier.
The audio power amplifier of claim 7 wherein the synchronous output rectifier
comprises a MOSFET.
The audio power amplifier of claim 1 wherein the first controlled signal path
comprises a switch controlled by the control electronics.
The audio power amplifier of claim 1 wherein the switching voltage amplifiers
each comprise a modulator, a gate driver, a pair of transistors, and a pair of
diodes coupled between the source and drain terminals of the transistors.
The audio power amplifier of claim 10 wherein the transistors comprise
MOSFETS, the diodes being intrinsic to the MOSFETS.
The audio power amplifier of claim 1 wherein the output filter comprises an
output inductor and the measured current is the current through the output
inductor.
P-09-013-WO
A method of amplifying audio-frequency signals comprising:
in each of a first and a second amplification unit,
amplifying a current-compensated command signal in a switching voltage
amplifier, providing an amplified signal output,
measuring current through an output filter between the amplified signal output
of the voltage amplifier and a load terminal, producing a current
measurement,
feeding back the current measurement to a summing input of a current
compensator via an inner current feedback loop,
at the current compensator, comparing the current measurement to a voltagecompensated
command signal and providing the current-compensated
command signal to the voltage amplifier,
feeding back voltage at a load terminal of the amplification unit to a summing
input of a voltage compensator via an outer voltage feedback loop, and
at the voltage compensator, comparing the feedback voltage to an input
command signal and providing the voltage-compensated command signal
to the summing input of the current compensator;
controlling a first controlled signal path from an output of the voltage
compensator of the first amplification unit to the summing input of the current
compensator of the second amplification unit, to selectively provide the
voltage-compensated command signal of the first amplification unit to the
summing input of the second amplification unit in place of the voltagecompensated
command signal of the second amplification unit; and
providing signal inputs to the first and second amplification units and controlling
the first controlled signal path to selectively operate the first and second
amplification units with separate loads, in parallel driving a common load, or
across a bridge-tied-load.
14. An audio power amplifier comprising:
WO
a first, a second, a third, and a fourth amplification unit, each amplification unit
comprising:
a switching voltage amplifier having a command signal input and an amplified
signal output,
an output filter between the amplified signal output and a load terminal,
a current compensator with a current-compensated command signal output
coupled to the command signal input of the voltage amplifier,
an inner current feedback loop feeding a measurement of current measured at
the output inductor back to a summing input of the current compensator,
a voltage compensator with a voltage-compensated command signal output
coupled to the summing input of the current compensator, and
an outer voltage feedback loop feeding voltage at the load terminal back to a
summing input of the voltage compensator;
the first and a second amplification units having a first controlled signal path from
the voltage-compensated control signal output of the voltage compensator of
the first amplification unit to the summing input of the current compensator of
the second amplification unit,
the second amplification unit configured to use the voltage-compensated
command signal of the first amplification unit as input to the current
compensator of the second amplification unit in place of the voltagecompensated
command signal of the second amplification unit when the
first controlled signal path is activated;
the third and a fourth amplification units having a second controlled signal path
from the voltage-compensated control signal output of the voltage
compensator of the third amplification unit to the summing input of the
current compensator of the fourth amplification unit,
the fourth amplification unit configured to use the voltage-compensated
command signal of the third amplification unit as input to the current
compensator of the fourth amplification unit in place of the voltage!-
WO
compensated command signal of the fourth amplification unit when the
second controlled signal path is activated;
control electronics providing signal inputs to each of the amplification units and
controlling the first and second controlled signal paths such that:
the first and second of the amplification units are operable with separate loads,
in parallel driving a common load, or across a bridge-tied-load;
the third and fourth of the amplification units are operable with separate loads,
in parallel driving a common load, or across a bridge-tied-load; and
all four of the amplification units are operable together with the first and
second units in parallel driving a first side of a bridge-tied-load, and the
third and fourth units in parallel driving a second side of the bridge-tiedload;
and
a four-quadrant power supply having a synchronous output rectifier.