Analog Pulse Width Modulation Amplifier

Schematic Rev 3
PCB Rev B
R. Balog
September 2001
Revision 3/19/2003
 ECE 369 Lab Procedure:

BE SURE TO TURN OFF THE POWER WHEN MAKING ANY CONNECTIONS TO

THE PWM AMP!!!
1. Set the power supply for approximately 14 V, with the current limit set to 1.5 A. Connect the
power supply to the PWM AMP via a wire harness with a red MTA push-on header.
Observe proper polarity. The PWM AMP is not extensively protected and reverse polarity
could damage the amp.

2. Use the differential probe to connect the oscilloscope to the output of the bridge. Test points
TP8 and TP9 provide access to the square wave bridge output. TP10 and TP11 provide
access to the output. Observe polarity.

3. Connect a standard oscilloscope probe to TP1, TP2, TP3 to observe the modulating function,
the carrier function, and the PWM output respectively. Turn on the supply and ensure the
PWM AMP is functioning. You should see waveforms similar to those below. Plot your
waveforms.

Top trace is the triangle function TP2 and vs. the unconnected input TP1.
Middle trace is the output of the bridge measured by the differential probe TP8 and TP9.
Bottom trace is the signal at TP3, the PWM out.

Revision 3/19/2003
4. Adjust the triangle wave for a peak-to-peak voltage of about 2V (resistor R23). Adjust the
duty ratio of the bridge output to 50% (resistor R5). Adjust the switching frequency to
approximately 130 kHz (resistor R3).

5. Connect a resistive load, approximately 8 to the output of the PWM AMP, and add a
current probe to monitor the output current. Turn on the supply. You should see waveforms
similar to those below. Plot your waveforms.

Top trace is the triangle function TP2 and vs. the unconnected input TP1.
Middle trace is the output of the bridge measured by the differential probe TP8 and TP9.
Bottom trace is the current out of the red banana jack.

6. Measure the average input current, and both the average and “ac” load voltage (after the
filter). What would you expect the average voltage to be?

Modulating function input:

7. Set the function generator to a sine wave. Set the amplitude to match the peak to peak
amplitude of the triangle function on your PWM AMP, and set the frequency to 1 kHz. Use
the function generator as the input to the amplifier.

8. Observe the output at various values of the “volume” attenuator (resistor R6), including 0%
and 100%.

Revision 3/19/2003
9. When R6 is at about 75%, plot the modulating function, the carrier function, the bridge
output (square pulse), and the current into the load together. Record the input average
current, and both the average and “ac” load voltage (after the filter).

Audio source:

10. Connect an audio source to the amplifier input, with the volume at 0%. Connect a
loudspeaker to the output jacks. Turn on the amplifier power supply.

11. Verify that the output of the output of the bridge is still a 50%-duty square wave. Set the
volume at a comfortable level. Plot the modulating function, the carrier function, the bridge
output (square pulse), and the current into the load together.

Dead-time gate drive signals:

12. Connect a standard oscilloscope probe to each of the four gate drive signals TP4-TP7. Set
the scope to display about 2 periods for each waveform. Identify the dead-time action.
Measure and record the dead-time.

AC motor drive

13. Reconnect the amplifier for 19 V dc input, function generator 60 Hz ac input, and motor
output. See the figure below. Initially, substitute a 1 k resistor for the motor itself.

14. Turn on the power supply and adjust for effective operation at 60 Hz, with an output voltage
close to the maximum available. CAUTION: The output voltage can get high enough to
cause a shock hazard!

15. Turn off power and connect the motor in place of the resistor. Restore power. Observe the
voltage of the secondary of the transformer, the motor current into terminal 4, and the motor
operation as the “volume” setting and drive frequency are adjusted. Explore the range from
about 30 Hz to about 180 Hz.

+ 4
-
1 2 .6 / V s e c o n d a ry
117 Vac
V o u tp u t +
3
2

- 1kF

Revision 3/19/2003
 Theory of Operation:
Pulse width modulation, when used as the basis for an amplifier, is termed a “class D” or
sometimes “class S” circuit. The principle is that the switch duty ratios can be made to follow
any desired waveform, provided only that switching is fast. The duty ratio signal can be
recovered with a simple low-pass filter step. The next few pages describe the configurations of a
specific bridge PWM inverter intended for use as a class-D amplifier.

Power Supply:
The amplifier receives DC power through the 4 pin header J2. Pins are labeled as appropriate
(see PCB plots attached). For general lab experimentation, the only voltage that you need to
supply to the PWM AMP is VCC (and ground). Depending on the desired amplitude of the
output, VCC can be selected within the range of 12 < VCC < 20. Anything less than 12V will not
be enough to power the ICs. Voltages above 20V will damage the FET driver ICs.

The PWM amplifier is designed both electrically and mechanically to interface with a small 12 V
power supply. A piece of sheet steel may be needed as a barrier between the PWM AMP and the
power supply. A solid ground connection between the PWM AMP circuit common and the
power supply ground should be made. However, a lab power supply can be substituted for
instructional purposes. Two series regulators provide regulated 12V and 5V for internal use
within the amplifier circuit.

Analog input:
Analog input is supplied through the 3.5mm stereo headphone jack. Internally, the left and right
channels are summed into a mono signal. The attenuator POT R6 is a 50K linear variable
resistor that attenuates the applied input signal prior to the comparator. The input is ac coupled
into the comparator stage through C2. R5 sets the dc bias (offset) on the analog input. Use this
to adjust the input offset to compensate for any drift in the amplifier and to achieve a 50% output
waveform for a 0V input. Turning R5 CW increases the DC bias.

Carrier and PWM Generation:

The triangle carrier function is generated by the VCO labeled U1 as seen on page 1 of the
schematic. The frequency of the triangle carrier is set by C1 and R3. Turning R3 CW (clock

Revision 3/19/2003
wise) increases the frequency. R23 sets the peak to peak amplitude of the triangle function.
Turning R23 CW increases the amplitude.

A general purpose comparator labeled U2 is used to create the PWM waveform by comparing
the modulating function (analog input) with the carrier function (triangle waveform).

Dead-time circuit:
The PWM waveform resulting from the comparator stage is passed into the dead-time circuit
comprised of U3 and U4 as seen on page 2 of the schematic. The result is two gate drive signals
and their complement. These four gate drive signals ensure that one set of switches completely
turns off before another set turns on. This break before make feature ensures that both switches
in one leg of the H bridge output stage are not both on, eliminating the possibility for shoot
through current and FET failure. The four gate signals are available on the orange test points
TP4-TP7.

Soft-start circuitry (R15, C11, C22,C23) provides approximately a 200ms startup period to allow
the power supply to stabilize before the bridge is allowed to run.

OUTPUT:
The output is derived from an “H bridge,” based on the fact that the four FET switches are used
in a geometry that resembles the letter H. Switches M1 and M4 operate as a one pair and M2
and M3 operate as the second pair. When M1 and M4 are on they provide a current path in the
positive reference direction. When M2 and M3 are on, they provide a current path in the
negative voltage reference. Thus the H bridge can supply both positive and negative output
voltages from a single supply.

Low Pass Filtering:

The output square wave from the bridge is low-pass filtered by L1, L2, C19, and C20. The
frequency response has a –3dB point at about 37.5 kHz and is characteristic of a 2 pole second
order filter. L1 and L2 are made by winding 20 turns onto a T050-26 core. See attached plot for
the calculated frequency response of the output filter. For carrier frequencies above 100 kHz, the
low pass filter should yield adequate performance and low standby ripple current.

Revision 3/19/2003
Output Filter for Analog PWM AMP -6 -9
L := 20 ´ 10 C := 900 × 10
RLoad := 8
1

L L L× C 4
V out C R Load C -V out f -3db= = 3.751 ´ 10
2p
RLoad

L 4
Full Bridge: KCL & KVL Equations = 6.366 ´ 10
2p
Vi - V1
=
V1
+
V1 - V2
(1)
V1 - V2
=
V2
+
V2 - -Vi ( ) (2) Vout = V1 - V2 (3)
j × w× L 1 R Load R Load 1 j × w× L
j × w× C j × w× C

Substitute (3) Substitute (3) Solve for V2

Vi - V1
=
V1
+
Vout Vout
=
V2
+
V2 - -Vi ( ) V2 = -Vout + V1
j × w× L 1 RLoad RLoad 1 j × w× L
j × w× C j × w× C

æ V × w× L + i× w2× C× R ö
Load × L× Vout - i × R Load× Vout + i × RLoad × Vi ø
V1 = -i × è
Vout out
-2
Vi
=
L× C × w -
2 2 × i × w× L
-1
( 2
RLoad × w × C× L - 1 )
RLoad

A( w) := 20× logé ù H( w) := 20× logé ù

2 1
Half Ckt
ê 2 2 × j × w× L ú transfer function
ê 2 2 × j × w× L ú
ê 1 + ( j w) × L× C + R ú ê 1 + ( j × w) × L× C + R ú
ë Load û ë Load û

Frequency Response of Output Filter

10
5
0
5
10
15
20
Amplitude [dB]

25
30
35
40
45
50
55
60
65
70
1 .10 1 .10 1 .10 1 .10
3 4 5 6
10 100
Frequency [Hz]
Full Bridge
Half Bridge
 arg é ù arg é ù
180 2 180 1
AP( w) := Half Ckt HP( w) :=
p ê 2 2 × j × w× L ú transfer function p ê 2 2 × j × w× L ú
ê 1 + ( j w) × L× C + R ú ê 1 + ( j × w) × L× C + R ú
ë Load û ë Load û

Frequency Response of Output Filter

45

45
Phase [deg]

90

135

180

225

270
1 .10 1 .10 1 .10 1 .10
3 4 5 6
10 100
Frequency [Hz]
Full Bridge
Half Bridge
 A
B
C
D

5
5

2 GND
GND_POWER 1
TCAP 7
4 TRWOUT TRES 6
3 SQWOUT MOD 5
VCC 8
4
4

3 1
3 1
3 1
1

3
3

4 5
1 6
8
1 3

2
2

1
1

A
B
C
D
 A
B
C
D

5
5

2 GND 7 14
4
4

7 14
7 14
7 14

3
3

7 14
7 14
14 7
7 14

2
2

7 14

1
1

A
B
C
D
 A
B
C
D

5
5

4
4

11 11
3 2 2 3
1 2 1 2
1

3
3

1
1

2
2

1
1 2 1 2
3 2 2 3

1
1

A
B
C
D