A Novel Audio Power Amplifer Topology with High Efficiency and State-of-the-Art Performance 5197

Thomas Frederiksen,Henrik Bengtsson,and KarstenNielsen

Bang & Olufsen PowerHousea/s
Struer, Denmark

Presented at
the 109th Convention
2000 September 22-25
Los Angeles, California, USA
This preprint has been reproduced from the author’s advance
manuscript, without editing, corrections or consideration by the
Review Board. The AES takes no responsibility for the
contents.

Additional preprints may be obtained by sending request and

remittance to the Audio Engineering Society, 60 East 42nd St,,
New York, New York 707652520, USA.

All r@hts reserved. Reproduction of this preprint, or any portion

thereof, is not permitted without direct permission from the
Journal of the Audio Engineering Society.

AN AUDIO ENGINEERING SOCIETY PREPRINT

 A novel Audio Power Amplifier Topology
with
High Efficiency and State-of-the-art performance

Thomas Frederiksen, M.Sc.E.E

E-mail: [email protected]

Henrik Bengtsson,B.Sc.E.E (Hons)

E-mail: [email protected]

Karsten Nielsen, Ph.D

E-mail: [email protected]

Bang & Olufsen PowerHouse a/s , Denmark

Abstract

A novel high efficiency power amplifier topology for

audio reproduction is presented. The topology breaks
previous performance barriers in switching
technology, by combining an effective error
correction method, Multivariable Enhanced Cascade
Control (MECC), with a new integrated modulator
topology - Controlled Oscillation Modulation (COM).
This topological combination proves to be very
elegant. Extensive measurements are given on 250W,
500W and 1000W case implementations of the MECC/COM
topology, showing e.g. 0.0005% (-106dB) true THD
combined with state-of-the-art in power and volume
efficiency.

1
1. Introduction
Well respected audio guru Ben Duncan has stated in
his book “High Performance Audio Power Amplifiers”
just three years ago in 1997:

“ It has been said since 1960 that once the potential

shortcomings of class D have been overcome to
everyone’s satisfaction, class D amplification will
be all there is for anything over a 100 watts or so.
But it hasn’t happened yet.... While simple on paper,
the calibre of engineering design needed to produce a
class D amplifier that doesn’t radiate EMI and is
measurably and audibly on a par with equivalent
analog amplifiers is truly formidable ... “

Why?

Compared to linear power amplifier designs, the

classical analog and digital class D or Pulse
Modulation Amplifier (PMA) systems present numerous
challenges to the designer. Just to mention a few
challenges:
• The complex switching power conversion stage is
difficult to model in detail and generates a
significant amount of switching noise disturbing
the feedback error correction system.
• The reconstruction filter further complicates the
implementation of effective error correction by

2
 implementing a higher order system transfer
function.
• Feedback in the digital modulation based systems is
not possible hence complicating this alternative
approach.
• Significant EMI considerations are necessary in
amplifier design and system implementation.

Some of the problematic issues have been reported in

earlier work by Attwood and Nielsen [2], [8], [15].
Generally, the list of required competencies to
design high performance power amplifier systems based
on switching technology is long and moreover
completely different from the competencies needed to
design linear power amps. Mr. Duncan’s statement
given just 3 – 4 years ago and basically concludes 40
years of work on this challenging topic.

2. The MECC based PMA system

However, much has happened over the last few years
[4] -
[21]. The research activities in the field have been
dramatically intensified. In an earlier paper (Paper
4839 / AES105 in San Francisco), a novel error
correction topology – Multivariable Enhanced Cascade
Control (MECC) – was proposed by one of the authors
as a new contribution to the field. The topology was
devised by detailed considerations of the specific
design problems in audio power amplifier systems

3
based on switching power conversion. The MECC
topology was shown to overcome the constraints of
previously applied feedback control methods, and
realize these objectives by remarkably simple means.
MECC was verified by a first generation prototype
indicating a clean break with the limitations in
previous designs. Research has continued on MECC
including studies on suitable modulator
implementation methods and this paper proposes a
novel modulator topology called the Controlled
Oscillation Modulator (COM).

Any Pulse Modulation Amplifier (PMA) power amplifier

system using switching power conversion can be
decomposed into three fundamental blocks: (1) the
pulse modulator (analog or digital), (2) the
switching power conversion stage with a passive
demodulation filter and (3) the control block. A
general system block diagram is shown in Fig. 1. The
pulse modulation may be either analog (i.e. analog
PMA) or digital (i.e. digital PMA). Independent on
the use of analog or digital pulse modulation, the
pulse modulator output, power stage output and filter
output are inherently analog signals, and thus
sensitive to jitter, pulse amplitude distortion or
any form of non-ideal behavior. Subsequently, open
loop operation has proven to be irrational from any
point of view (performance, complexity, power supply
requirements ....)

4
[17], and the control system is thus an essential part
of any PMA system. Recently, a suite of control
methods for analog PMAs were investigated
[12].

As presented in
[14] MECC has two fundamental variants henceforth
referred to as MECC(N) and MECC(M,N). Fig. 2 shows the
extended general (N+M)-loop MECC(N,M) topology.
Fundamentally, MECC is a recursive structure of N
loops formed as an enhanced cascade from a single
feedback source. This simple extension offers some
advantages. MECC(N) is characterized by the following
distinct points:
• A single feedback source.
• A single feedback path A(s ) independent upon the
number of loops N, providing a minimal system
complexity.
• The feedback path has a low-pass characteristic, to
filter the noise from vp and compensate the
demodulation filter.
• An initializing B1 ( s ) compensator block with special
characteristics.
• A recursive structure with a set of preferably
identical forward path compensator blocks Bi (s ) .

Thus, the Enhanced Cascade refers to these special

cascade control characteristics or this dedication of

5
the cascade to the PMA control problem. Cascade
control methods have previously been applied to
linear power amplifier systems, in terms of e.g. the
well known Nested Differential Feedback Loop method
(NDFL’s) presented by Cherry
[3]. The motivation for developing MECC for PMA system
has been similar to Cherry’s for linear amps. The
characteristic effective loop transfer function is
shown in Fig. 3.

A fundamental constraint within MECC(N,M) system

design is:

M ≥1⇒ N ≥1 (1)

MECC(N) provides optimized control in dedicated

applications where filter linearity is unproblematic
and the load is known. The MECC(N,M) provides
optimized control in all general applications. Both
topologies have their place.

The MECC(N,M) topology is founded on a MECC(N) design

and should be seen as a direct extension of this
topology. The MECC(N,M) topology is characterized by:
• A MECC(N) system, that is optimized specifically
for the global enhanced cascade.
• A single feedback source vo .

• A single feedback path compensator C .

• A D1 compensator to initialize the cascade.

6
• A recursive structure with a set of preferably
identical compensator blocks Di .

The topological resemblance between MECC(N) and

MECC(N,M) also leads to similarities in the synthesis
of the two cascade structures. However, MECC(N,M) is
constituted of two closely connected enhanced
cascades, where the global enhanced cascade relies on
the compensation from the local cascade. Loop
synthesis in the MECC system is addressed in
[14],
[15]. As also shown in
[14],
[15] – a high performance level can be achieved by the
MECC(1,1) topology.

3. The novel COM modulator topology

This paper extends previous research and proposes a
new modulation techniques optimally suited to the
MECC control system. Traditional PWM techniques or
PDM techniques provides certain limitations
[15] to the MECC control system. PWM has a range of
shortcomings as e.g.:
• Precision carrier implementation is troublesome.
Errors on the carrier limit performance.
• Control system design is complicated.
• The high amplitude switching noise source limits
control system bandwidth.

7
Stable and robust control system design is difficult
• [15].

Accordingly, the primary objective in the search for

more suitable modulation techniques has been to
develop a modulation technique that overcomes some of
these fundamental problems.

The result of the extensive search for better methods

is the Controlled Oscillation Modulator (COM)1. The
basic topology is shown in Fig. 4. The COM system is a
combined modulation and control system surrounding a
central power conversion stage. As seen in the
isolated system in Fig. 4, an input reference voltage
vi is feed to a compensation unit B, also feed by the
feedback compensator A to derive and process error
information. The compensated signal vb is feed to a
comparator which is referenced to a constant voltage
v DC , preferably a DC voltage. The resulting pulse
modulated signal is power amplified in a switching
power conversion stage supplied by VS , generation the
power pulse signal VP , this pulse signal driving an
inductive load. The A block processes output
information from the output voltage VP and controls
the overall transfer function characteristics of the
system. The COM system is characterized by at least
one pole in the A and B blocks, in combination with

1
COM is proprietary technology of Bang&Olufsen PowerHouse a/s.

8
the propagation delay of the modulator and power
stage generating a sinusoidal like modulating signal
vb to be compared with v DC . Under these presumptions,
the system realizes a oscillating system at the
frequency of positive feedback. The typical
characteristic of the COM modulating signal is
illustrated in Fig. 5.

3.1 COM system example

Consider the system in fig. 5. Assuming, that a
constant gain K is desired over a certain bandwidth,
example general A and B block characteristics are:

1 1 1 τ zB s + 1 1 (1)
A( s ) = B( s) = K B
K τ 1s + 1 τ o s + 1 τ pB s + 1 τ o s + 1

In this illustrative example, the A-block having a

first order characteristic with a pole s = −τ 1 placed at
lower frequencies, generally more than a decade below
the desired oscillation frequency. The oscillation
conditions are conformably determined by τ o with the
1
two poles placed at s = −τ o = − . The requirements for a
ω0

“ Controlled Oscillation Modulation ” is:

L( jω o ) = K P A( jω o ) B ( jω o ) = 1 ∧ p L( jω o ) = 180o (2)

9
where the desired system oscillation frequency is ω0 .
Hence, in this preferred example, the condition for
controlled oscillation is:

ω oτ p1τ pB (3)
KB KP = K
2τ zB

The COM system will be forced to oscillate at ω o due

to the non-linear gain characteristic of the
comparator and power stage. The resulting COM system
is easily integrated in the MECC system. Actually,
the COM system is equivalent to a MECC(1,0) system
[15].

COM offers superior characteristics compared to

widely used carrier PWM and PDM techniques. Some of
the general advantages of COM are:
• The COM system is inherently unstable leading to
robust operation.
• Very simple implementation. No carrier generator is
needed saving components and improving quality (no
distortion, noise, jitter etc. from carrier or
clock generator).
• The power supply variable VS is eliminated from the
effective loop transfer function. The rejection to
perturbations on VS is infinite as opposed to none
in e.g. a PWM system or a limited factor in a
feedback PWM system.

10
• The bandwidth of the control system is
approximately equal to the resulting carrier
frequency.
• The modulation is clean with a comparison of a
sinusoidal signal with zero or a DC voltage.
• Controlling loop order and propagation delay can
control the switching frequency variation for
improved EMI and efficiency.

4. Evaluating the MECC/COM PMA topology

The MECC/COM PMA topology has been thoroughly
evaluated and optimized and three cases will be
investigated in the following. The case examples are
the ICE250A, ICE500A and ICE500A products which have
been implemented using selected variables for the
modulator and control system. Essential parameters
for the three case examples are shown below:

Parameter ICE250A ICE500A ICE1000A

Av. output power 250W 500W 1000W
fb 80kHz 80kHz 40kHz

fc 400kHz 400kHz 200kHz

N 1 1 1
M 1 1 1
Vp 50V 75V 110V
K 26dB 26dB 26dB

A picture of the three ICEpower modules is shown in

Fig. 6. In general, the power stage implementation is

11
very relaxed and optimized for efficiency. Open loop
THD is 1-2% worst case. Since the performance is
equivalent for the three power levels, we will focus
on ICE500A performance.

Fig. 7 illustrates the frequency response of the

system in 2.7Ω to open load. The system response is
within ±0.2dB in all loads from 2Ω to an open load
situation. This is due to the very low output
impedance of the system, which is below 25mΩ at all
frequencies.

Fig. 8 shows THD+N at various frequencies for the 250W

case module. 7kHz loading corresponds to the worst
case situation with 22kHz and 30kHz bandwidth
filtering (AP). For the higher power modules, the
performance is equivalent
[22].

Fig. 9 shows an FFT analysis of the amplifier output

at 5kHz/100mW. The analysis reveals the extreme
linearity of 0.0005% or –106dB of the MECC based PMA
system at typical output powers – even at higher
frequencies. This is quite exceptional for such a
high power PMA system and fully comparable with what
is achieved by the very best linear power amplifiers.
As shown in Fig. 8, a high level of linearity is
maintained at all frequencies and output powers.

12
Thus, THD+N maintains to be below 0.025% to the
maximum output levels in the tweeter range.

Fig. 10 illustrates the efficiency characteristics,

again for the 250W case example in an 8 ohm load.
Notice the high efficiency also at lower output
powers.

Detailed specifications for the 250W case example are

illustrated on the following page.

13
 Electrical Specifications – 250W case example

SYMBOL PARAMETER CONDITIONS TYP UNIT

Vp Power Supply Operation 50 V
Output power @ 0.05%THD+N
RL=4Ω. Vp=50V 200
PO 10Hz < f < 20kHz W
RL=8Ω. Vp=50V 110
(22kHz BW measurement)
THD+N THD + N in 4Ω f = 1kHz, PO=1W 0.005 %

Maximal THD + N in 4Ω 10Hz < f < 20kHz

THD+N 0.03 %
(22kHz BW measurement) 100mW < Po < 200W
IVp Quiescent current Vp=50V 30 mA
fo Offset switching frequency Offset carrier at idle 380 kHz
n Power stage Efficiency RL=8, PO=100W 93 %
PSRR Power Supply Rejection 70 dB
Output referenced idle A-weighted
VNo 65 µV
noise 10Hz < f < 20kHz
Output referenced offset
VOFF Terminated input 5 ±mV
(DC calibration active)
Av Nominal Voltage Gain 27.0 dB
Frequency response
F 20Hz-20kHz, All loads ±0.2 dB

Upper bandwidth limit

fu RL=8 80 kHz
(-3dB)
Lower bandwidth limit RL=8
fl 4 Hz
(-3dB)
Zo Abs. output impedance f = 1kHz 5 mΩ
D Dynamic range A-weighted 115 dB
IMD1 Intermoduation (CCIF) f=19kHz,20kHz, Po=10W 0.001 %
IMD2 Intermodulation (SMPTE) f=60Hz,7kHz(1:4), Po=10W 0.001 %
Transient intermodulation f1=3.15kHz square,
TIM 0.001 %
(TIM) f2=15kHz, Po=10W

Detail specifications for the MECC/COM based full bandwidth

PMA system.

14
5. Conclusions
The paper has presented a novel PMA topology
realizing state-of-the-art performance. A novel
modulator topology was presented – Controlled
Oscillation Modulation (COM) - which integrates well
with the previously proposed MECC control topology.
The COM modulator proves to have many advantageous
characteristics over conventional PWM or PDM
modulator topologies:
• No carrier generator is needed saving components.
• Inherently unstable – hence very robust since
damaging instability cannot occur.
• The power supply variable VS is eliminated from the
effective transfer function - > PSRR is infinite.
• The bandwidth of the control system is
approximately equal to the resulting carrier
frequency.
• The modulation is clean with a comparison of a
sinusoidal signal with zero or a DC voltage. This
improves the precision of the system.
• A controllable variable switching frequency (by
controlling loop order and propagation delay)
improves efficiency and can be used to lower EMI.

These theoretical advantages have been extensively

proved in practice by the implementation of three
case examples; 250W, 500W and 1000W. To conclude,
the PMA performance level and sound quality is now

15
fully comparable with high end linear class A/B
technology and on many parameters superior to linear
class A/B amplifiers with the presented topology.

6. Patent protection
The MECC and COM methods are protected by several
patents and patent applications and are the
proprietary rights of Bang & Olufsen PowerHouse a/s.

7. Acknowledgement
The authors are very grateful to professor
Michael.A.E.Andersen at IAE/DTU. Our fruitful
research partnership in efficient power conversion
for audio reproduction is the very foundation for the
results presented in this paper.

16
8. References

[1] Duncan, Ben

" High Performance Power Amplifiers "
Newness. Butterworth, Heineman. 1997.

[2] Attwood, B.E.

" Very high Fidelity Quartz Controlled PWM (Class D) Stereo
Amplifiers for Consumer and Professional Use"
59th Convention of the AES. March 1978. Hamburg. Paper 1331.

[3] Cherry, E.M.

" Nested Differentiating Feedback Loops in Simple Audio Power
Amplifiers "
JAES. Vol. 30, No.5, May 1982. pp. 295-305.

[4] Vanderkooy, John

“New concepts in Pulse Width Modulation “
97th Convention of the AES, November 1994. San Francisco.
Preprint 3886.

[5] Klugbauer, Josef

"A Sigma-Delta Power Amplifier for Digital Input Signals"
102nd AES Convention. Munich, March 1997. Preprint 4448.

[6] Anderskouv Niels, Nielsen, Karsten. Andersen, Michael.

"High Fidelity Pulse Width Modulation Amplifiers based on
Novel Double Loop Feedback Techniques"
100th AES Convention. Copenhagen, May 1996. Preprint 4258.

[7] Nielsen, Karsten

" Parallel Phase Shifted Carrier Pulse Width Modulation
(PSCPWM) –

17
A novel approach to switching power amplifier design "
102nd AES Convention. Munich, March 1997.
Paper 4447

[8] Nielsen, Karsten

" A Review and Comparison of Pulse Width Modulation methods
for
analog and digital input switching power amplifier systems "
102nd AES Convention. Munich, March 1997.
Paper 4446.

[9] Nielsen, Karsten

"High Fidelity PWM based Amplifier Concept for active speaker
systems with a very Low Energy Consumption"
Journal of the Audio Engineering Society. July/August 1997.
pp. 554-570.

[10] Nielsen, Karsten

" Pulse Edge Delay Error Correction (PEDEC) - A Novel Power
Stage Error Correction Principle for Power Digital-Analog
Conversion”
103rd AES Convention. New York, September 1997.
Paper 4602.

[11] McLaughlin, R. David, Stanley, Gerald R. and Wordinger,

James.
"Audio Amplifier Efficiency and Balanced Current Design – A
New
Paradigm"
103th AES Convention. New York, USA. September 1997.

[12] Nielsen, Karsten, Taul, Thomas, Andersen, Michael

" A comparison of Linear and Non-Linear Control Methods for
Power

18
Stage Error Correction in Switching Power Amplifiers ”
104th AES Convention. Amsterdam, Holland.

[13] Risbo, Lars, Mørch, Thomas

"Performance of an all digital power amplification system”
104th AES Convention. Amsterdam, Holland.

[14] Nielsen, Karsten

" MECC – A novel control method for high end switching audio
power amplification “
105th AES Convention. San Francisco, USA. Preprint 4839.

[15] Nielsen, Karsten

" Audio Power Amplifier Techniques with Energy Efficient Power
Conversion ”.
Ph.D. Thesis.
Department of Applied Electronics, DTU, Denmark. May 1998.

[16] Nielsen, Karsten

" PEDEC - A Novel Pulse Referenced Control Method for High
Quality Digital PWM Switching Power Amplification "
IEEE Power Electronics Specialist Conference (PESC).
Japan, May 1998. Conf. Proc. pp. 200-208.

[17] Nielsen, Karsten

" Linearity and Efficiency Performance of Switching Power
Amplifier
Output Stages – A fundamental analysis "
105th AES Convention. San Francisco. September, 1998.

[18] Nielsen, Karsten

"Digital Pulse Modulation Amplifier systems based on PEDEC
control”
106th AES Convention. Munich, Gernany. March, 1999. 25 pp.

19
Paper 4942

[19] Nielsen, Karsten

"Parallelled Phase Shifted Carrier Pulse Width Modulation
(PSCPWM)
schemes – A fundamental analysis ”
106th AES Convention. Munich, Gernany. March, 1999. 25 pp.
Paper 4917

[20] Johansen, Morten, Nielsen, Karsten

" A review and comparison of digital PWM methods for digital
pulse
modulation amplifier systems ”
107th AES Convention. Munich, Gernany. March, 1999.
Paper 5039

[21] Christensen, Frank, Frederiksen, Thomas, Andersen,

Michael, Nielsen,
Karsten
" Practical Implementation and Error Analysis of PSCPWM based
switching power amplifier systems “
107th AES Convention. Munich, Gernany. March, 1999.
Paper 5040

[22] Web-site: www.BangOlufsen-PowerHouse.com

20
 DC power supply

Pulse
Analog input
Modulator Power Switch Demodulation

Control

Fig. 1 General analog Pulse Modulation Amplifier topology.

vc
C(s)

va
A(s)
vS

vr
DM(s) vd M D1(s) vd 1 BN(s) vb N B1(s) vb 1 (KP) vp vo
Modulator + Power Switch Demodulation
General (N+M) - loop MECC2 Topology

Fig. 2 General (N+M) - loop MECC(N,M) topology

21
 80 L
4

L
60 3

Gain (dB)
L
40 2

L
20 1

−2 −1 0 1 2
10 10 10 10 10
Normalized frequency (f/fb)

−50
L
1
Phase (deg)

−100 L
2

−150 L
3

−200 L
4
−2 −1 0 1 2
10 10 10 10 10
Normalized frequency (f/fb)

Fig. 3 MECC(N) parametric analysis of effective loop transfer

function LN . (N = 1,2,3,4).

Controlled Oscillation Modulator

Va
A

Vs

Vb
Vi B VDC
Vp

Fig. 4 Basic idea of COM system

22
 1

0.5

Modulating signal
0

−0.5

−1
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1

0.5
Power signals

−0.5

−1

0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Normalized time

Fig. 5 Modulating signal

Fig. 6 Modules implemented with the MECC/COM topology.

Physical size is only 80x80x25mm (250W), 90x90x25mm (500W) and
100x100x25mm (1000W), respectively.

23
 +30 +175
+28 +150
+26
+125
+24
+100
+22
+75
+20
+50
+18
d +25
B +16 d
g +0 e
+14 g
A -25
+12
-50
+10
-75
+8
-100
+6
-125
+4
+2 -150

-0 -175
10 20 50 100 200 500 1k 2k 5k 10k 20k 50k 200k
Hz

Fig. 7 Frequency Response in 2.7Ω, 4Ω, 8Ω and open load. Top

– amp. Bot – Phase.

Fig. 8 THD+N at 100Hz, 1KHz and 7kHz in a 4 ohm load (22kHz

bandwith). 250W case.

24
Fig. 9 16K/16x av. FFT at 5KHz/100mW. 250W case. THD = -106dB

100

90

80

70

60

% 50

40

30

20

10

0
0 20 40 60 80 100
Watts

Fig. 10 Efficiency vs. output power. 250W case (8 ohm load).

25