Transcript aaa

Design of Class-D Audio
Amplifiers
Chun-Hsien Su
中央大學電機系 蘇純賢
[email protected]
May 17, 2006
OUTLINE
1. Introduction of Class-D Audio Amplifier
2. Large-Power Class-D Applications
3. Integrated Class-D Audio Power Stages
4. Sigma-Delta Class-D Audio Amplifiers
5. Conclusions
Chun-Hsien Su
2
Introduction of Class-D Audio Amplifier
– Traditional Class-AB Amplifier
 Class AB amplifier uses linear regulating transistors to
modulate output voltage.
 η = 30% at temp rise test condition.
Chun-Hsien Su
3
Introduction of Class-D Audio Amplifier
– Class-D Audio Amplifier
 Class D amplifier uses MOSFETs that are either ON or OFF.
 PWM technique is used to express analog audio signals with
ON or OFF states in output devices.
Chun-Hsien Su
4
Basic PWM Operation
Chun-Hsien Su
5
Audio Output Spectral Density
Fundamental
Carrier and modulated signals
LC-filtered
Chun-Hsien Su
6
Topology Comparison: Class AB vs. Class D
Chun-Hsien Su
7
Efficiency: Class AB vs. Class D
[From TI’s report]
Chun-Hsien Su
8
Measured Efficiency
[From TI’s data]
– TI-TPA2000D4
Chun-Hsien Su
9
Temperature vs. Output Power
– Class-D Advantage: Less Heat
Chun-Hsien Su
10
Battery Life
A head to head test of
TI’s class-AB and class-D
~1W mono solutions:
Class-AB - TPA731
Class-D - TPA2001D
Both set-up on the PnP
platform powered by
3x1.2V NiMH batteries
driving a typical 8-Ω
speaker used in wireless
applications.
Chun-Hsien Su
11
Battery Life
A head to head test of
TI’s class-AB and class-D
~1W mono solutions:
Class-AB - TPA731
Class-D - TPA2001D
Both set-up on the PnP
platform powered by
3x1.2V NiMH batteries
driving a typical 8-Ω
speaker used in wireless
applications.
Chun-Hsien Su
12
Analogy to Buck DC-DC Converter
Chun-Hsien Su
13
Loss in Power Device
Chun-Hsien Su
14
Full Bridge versus Half Bridge
– Full Bridge
– Full Bridge
Chun-Hsien Su
15
Full Bridge versus Half Bridge (Cont.)
Chun-Hsien Su
16
Major Cause of Imperfection
Architecture
Analog/digital
w/SDM
Switch nonidealities
Finite Ron
Body diode Recovery
(EMI)
Poor PSRR
Noise coupling
Feedback
Nonlinear LC
PWM
Audio source
Modulation error
Quantization error
Chun-Hsien Su
Gate
Driver
Dead time
Delay time
Single-bridge/Full bridge
Filterless modulation scheme
17
THD and Dead Time
Chun-Hsien Su
18
Shoot Through and Dead Time
- Shoot through charge increases rapidly as dead time gets shorter.
- Need to consider manufacturing tolerances and temperature
characteristics.
Chun-Hsien Su
19
Power Supply Pumping
- Significant at low frequency output
- Significant at low load impedance
- Significant at small bus capacitors
- Largest at duty = 25%, and 75%
Chun-Hsien Su
20
EMI consideration: Qrr in Body Diode
1. Low side drains inductor current
2. During dead time body diode of low side
conducts and keep inductor current flow
3. At the moment high side is turned ON after
dead time, the body diode is still conducting
to wipe away minority carrier charge stored
in the duration of forward conduction.
 This current generates large high frequency
current waveform and causes EMI noises.
Chun-Hsien Su
21
Gate Driver: Why is it Needed?
 Gate of MOSFET is a capacitor to be charged and
discharged. Typical effective capacitance is 2nF.
 High side needs to have a gate voltage referenced to it’s Source.
 Gate voltage must be 10-15V higher than the drain voltage.
Need to control HS and LS independently to have dead time.
Chun-Hsien Su
22
Functional Block Diagram Inside Gate Driver
 High side needs to have a gate voltage
referenced to it’s Source.
 With the addition of few components, they
provide very fast switching speeds and low
power dissipation.
 Need to control HS and LS independently to have dead time.
Chun-Hsien Su
23
Bootstrap High Side Power Supply
When Vs is pulled down to ground through the low side FET,
the bootstrap capacitor (CBOOT) charges through the
bootstrap diode (Dbs) from the Vcc supply, thus providing a
supply to Vbs.
Chun-Hsien Su
24
Power Dissipation in Gate Driver
Whenever a capacitor is charged or discharged through a
resistor, half of energy that goes into the capacitance is
dissipated in the resistor. Thus, the losses in the gate drive
resistance, internal and external to the MGD, for one complete
cycle is the following:
PG  V  f SW  QG
Chun-Hsien Su
25
Power Dissipation in Gate Driver (Cont’d)
The use of gate resistors
reduces the amount of gate
drive power that is dissipated
inside the MGD by the ratio
of the respective resistances.
These losses are not
temperature dependent.
Chun-Hsien Su
26
MOSFET Power Switch
A MOSFET is a voltage-controlled power switch.
A voltage must be applied between Gate and
Source terminals to produce a flow of current in
the Drain.
Chun-Hsien Su
27
Key Parameters of Power MOSFETs
(1). Voltage Rating, BVDSS : Drain-source breakdown voltage.
Temperature dependent.
(2). Gate Charge, Qg : Directly related to MOSFET speed
Temperature dependent.
(3). D-to-S On-Resistance, RDS(ON) : Directly related to MOSFET
conduction losses. Temperature dependent.
(4). Body Diode Reverse Recovery Characteristics, Qrr, trr , Irr , and S
factor. Influence THD, EMI, and Efficiency. Temperature dependent.
(5). Package : Power dissipation capability, current capability,
internal inductance, internal resistance, electrical isolation, and
mounting process.
Chun-Hsien Su
28
Voltage Rating, BVDSS
 This is the drain-source breakdown voltage (with VGS = 0).
BVDSS should be greater than or equal to the rated voltage
of the device, at the specified leakage current, normally
measured at Id=250uA.
 This parameter is temperature-dependent and
frequently ∆BVDSS/∆Tj (V/°C) is specified on datasheets.
 BVDSS MOSFET voltages are available from tens to
thousand volts.
Chun-Hsien Su
29
Gate Charge, Qg
 parameter is directly
related to the MOSFET speed
and is temperatureindependent.
Lower Qg results in faster
switching speeds and
consequently lower switching
losses.
 The total gate charge has two
main components: the gate
source charge, Qgs and, the
gate-drain charge, Qgd (often
called the Miller charge).
Chun-Hsien Su
30
D-to-S On-Resistance, RDS(ON)
 This is the drain-source
resistance, typically
specified on data sheet at
25°C with VGS = 10V.
RDS(ON) parameter is
temperature-dependent, and
is directly related to the
MOSFET conduction losses.
lower RDS(ON) results in
lower conduction losses.
Chun-Hsien Su
31
Body Diode Reverse Recovery Characteristics, Qrr, trr , Irr , and S Factor
 Power MOSFETs inherently have
an integral reverse body-drain
diode. This body diode exhibits
reverse recovery characteristics.
Reverse Recovery Charge Qrr,
Reverse Recovery Time trr,
Reverse Recovery Current Irr and
Softness factor (S = tb/ta), are
typically specified on data sheets
at 25°C and di/dt = 100A/us.
 Power recovery characteristics
are temperature-dependent and
lower trr, Irr and Qrr improves
THD, EMI and Efficiency η.
Chun-Hsien Su
32
Package
 MOSFETs devices are available in several packages as
SO-8,TO-220, D-Pak, I-Pak, TO-262, DirectFET™, etc.
 The selection of a MOSFET package for a specific
application depends on the package characteristics such
as dimensions, power dissipation capability, current
capability, internal inductance, internal resistance,
electrical isolation and mounting process.
Chun-Hsien Su
33
Choosing the MOSFET Voltage Rating
MOSFET voltage rating for a Class D amplifier is determined by:
– Desired POUT and load impedance (i.e. 250W on 4Ω)
– Topology (Full Bridge or Half Bridge)
– Modulation Factor M (80-90%)
Chun-Hsien Su
34
Trends in Class-D Audio Amplifiers
– Make it smaller!
 higher efficiency
 smaller package
 Half Bridge
– Make it sound better!
 THD improvement
 fully digitally processed modulator
Chun-Hsien Su
35
Trends in Class-D (1)
– Patent Blooming since Y2K
US Patent Class 330/207A – Class-D Amplifier
50+
Patent #
40
33
30
25
Chun-Hsien Su
15
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
1985
13
36
Trends in Class-D (2)
– Systems by ICs
Bipolar
CMOS
DMOS
(doublediffused
MOS)
DMOS features:
• It is a (lateral) double-diffused MOS transistor.
• The device is asymmetrical.
• A lightly doped Nwell(extended drain) region supports high drain voltages.
• The thin gate oxide allows a high device but restricts
• The Pbody is shorted to source of device in metal 1.
Chun-Hsien Su
37
Trends in Class-D (3)
– Full versus Half, Performance versus Cost
Full Bridge
Half Bridge
Chun-Hsien Su
38
Trends in Class-D (4)
– Performance Improved by Analog FB Loop
US Patent : 6300825, issued 2001/10/9
“PWM with feedback loop Integrator”
Similar: Yokoyama 4504793
1985/3/12
 This structure improves output waveform & PSRR.
 This patent can be avoided by using different architecture.
Chun-Hsien Su
39
Trends in Class-D (5)
– Filterless Modulation for Portable Applications
Gate
Driver
Information in pulse width
Information in phase
US Patent : 6262632, issued 2001/7/17
“Concept and method to enable filterless, efficient operation of Class-D amplifiers”
 Limits the development of small-power (~1W) class-D amplifier.
 Can be solved by using different modulation schemes.
Chun-Hsien Su
40
Trends in Class-D (6)
– Performance Improvement by Sigma-Delta?
PWM
Delta
Sigma
-Delta
The power efficiency and the signal-to-noise
ratio (SNR) for a amplifier based on the
above scheme are compared with the values
typical of linear and PWM amplifiers of the
same rated power. The power efficiency of
the solution is an intermediate value
between the linear and PWM amplifiers,
whereas SNR mainly depends on the
modulator type adopted. A SNR around 60–
70 dB was estimated by adopting a fourthorder modulator. However, this value can be
significantly improved by increasing the
switching frequency or the modulator order.
[Dallago, Tran. CAS-I, Aug, 1997]
Chun-Hsien Su
41
Trend in Class D Amplifiers
In any case, to extend the power range of high-frequency
modulation, future studies should be oriented in the following
direction:
(1). Reduction of power circuit parasitic capacitance by technological
improvement of both semiconductor and magnetic devices;
(2). Adoption of soft-switching schemes able to perform a larger
exploitation of converter parasitic parameters under a load range
that is as wide as possible;
(3). development of extended modulation techniques (for example of
multilevel type or nonconstant switching period type or, generally,
of hybrid type) allowing a lower switching frequency under the
same baseband performance.
Chun-Hsien Su
42
PWM versus PDM
THD
Class-D
CT SDM
Comparator Run-free
Clocked
Output
PWM
1-b PDM
Min. pulse
Width
1/fs
fs<1MHz
1/fs
fs>1MHz
Drive
H-bridge
OK
May burn out
PWM
1-bit
PDM
For second-order  M
fs=40kHzx128=5.12MHz
Open-loop, Full-bridge, low-OSR M with bit grouping
Audio source
Chun-Hsien Su
High-order
Single-bit
M
Bit flipping
(bit grouping)
(digital)
(digital)
Gate
Driver
(d & a)
43
 Class-D Amplifier
Dual noise-shaper architecture - Simple, almost all digitized
Audio
source
Digital
M
PCM
Interpolator
+
-
 Noise
Shaper
PCM
Interpolator
Chun-Hsien Su

Grouping
Gate
Driver
Waveform
Represent
m-bit
AD
1/m-bit
DAC
CT
filter
 Noise
Shaper
 Class D
Amplifier
Class AB Speaker
/Class D
Amplifier
Speaker
44
Conclusion
– Highly efficient Class D amplifiers now provide similar
performance to conventional Class AB amplifiers - If key
components are carefully selected and the layout takes into
account the subtle, yet significant impact due to parasitic
components.
– Constant innovation in semiconductor technologies helps the
growing Class D amplifiers usage due to improvements in
higher efficiency, increased power density and better audio
performance. Trends in class-D: half-bridge filterless
scheme, analog-feedback loop to improve performance.
– All-digital Sigma-delta class-D  Future work.
Chun-Hsien Su
45
References
[1]. M. Berkhout, “An integrated 200-W Class-D audio amplifer,” IEEE J.
Solid-State Circuits, vol. 38, pp. 1198–1206, July 2003.
[2]. H. Ballan, M. Declercq, and J. U. Duncombe, “12 V - Class-D
amplifier in 5 V CMOS technology,” in Proc. IEEE CICC, 1995, pp.
559–562.
[3]. E. Dallago, “Advances in high-frequency power conversion by deltasigma
modulation,” IEEE Trans. Circuits Syst. I, vol. 44, pp. 712–721,
Aug. 1997.
[4]. Marco Berkhout, “An integrated 200 W class-D audio amplifier,” IEEE J.
Solid-State Circuits, vol. 38, no. 7, pp. 1198–1206, Jul. 2003.
[5]. Jun Honda and Jorge Cerezo, IRA Technical Report.
Chun-Hsien Su
46