Design Project 3 - The Class-D Amplifier

Part One

Click here to read the material on the class-D amplifier that I wrote for the book I use in Audio Engineering. It explains the theory of operation of the class-D amplifier and gives some of the design equtions. The object of this project is to design a mock class-D amplifier with op amps, comparators, and complementary MOS transistors. Do not use the 741 op amp for any of the circuits. It has too low a slew rate to be used in this project.

In this part of the lab, you are to assemble the amplifier in Figure 1 of the class-D article. You will have to design the triangle wave generator that dirves the + input of the op amp, the CMOS inverter output stage, and an active filter in place of the passive LC filter shown in the class-D article.

The MOS transistors used for the output stage are in the CD4007 chip that is used in the ECE 3042 lab. Only two of the transistors in this chip are to be used, one an n-channel and the other a p-channel. The maximum voltage rating for this chip is 15 V. For this reason, the circuits for the design project are to be operated at power supply voltages of + and - 7.5 V. The op amps should operated at this reduced voltage with no problems.

The first step is to assemble the triangle wave generator using an op amp in place of the comparator. The design of this circuit is covered in the course textbook and in the class-D amplifier article. Design the circuit for an initial 50 kHz triangle-wave frequency. The peak voltage of the triangle wave is to be 6 V. This voltage sets the clipping voltage of the class-D amplifier at 6 V. If you cannot get the triangle wave generator to work properly at 50 kHz, it may be necessary to lower its frequency. However, the class-D amplifier requires this frequency to be as high as possible.

The second step is to connect the triangle wave and a 1 kHz test signal from a function generator to the two inputs of an op amp that is used as a comparator. You should see an output that looks like a square wave with a modulation of the widths of its positive and negative peaks. If the op-amp comparator will not switch at 50 kHz, it may be necessary to reduce the triangle wave frequency.

Connect the CD4007 chip to the circuit. Note that it is connected as a CMOS inverter. Put a 1 kΩ load resistor from the output of the inverter to ground. Use the oscilloscope to observe the signals in the circuit. The output of the CMOS inverter should be an inverted replica of the output of the comparator.

The class-D amplifier article shows a passive LC circuit between the amplifier and the loudspeaker load. Although a passive filter must be used if the amplifier is to drive a loudspeaker, we can replace it with an active filter for purposes of this lab. The filter is to be a 3rd-order Chebyshev filter with a dB ripple of 0.5 dB. The design of this filter is described starting on page 30 here. You will probably not be able to find exact capacitor values, but you should be able to come close. For a "first cut" design, a cutoff frequency of 1/10 of the triangle wave frequency is suggested. For example, if you were able to make the pulse-width modulator work properly with a triangle wave frequency of 50 kHz, you might choose the filter cutoff frequency to be 5 kHz. The cutoff frequency can easily be changed later if desired to optimize the amplifier performance. The filter circuit diagram shows only one op amp in the circuit. A second op amp should be added to the filter input that is operated as a non-inverting, unity-gain buffer. Measure and document the frequency response of the filter before connecting it to the inverter output. You should measure and document the frequency response of the filter before connecting it to the circuit.

When you connect the filter to the output of the inverter, you should leave the 1 kΩ inverter load resistor in the circuit. If everything is working, you should see the audio output signal at the output of the filter and no 50 kHz switching signal.

Part Two

The last part of the experiment is to add negative feedback to the circuit. This is done by implementing an inverting integrator as an input stage to the amplifier as shown in figure 6 of the document on the class-D amplifier. In order for the feedback to be negative, you should adhere to the op-amp polarities shown in Figure 5. The amplifier should exhibit an overall closed-loop gain of 10 (20 dB). This might be achieved with R1 = 10 kΩ and RF = 100 kΩ. See page 3 of the Class-D Amplifier document for suggestions on picking a value for CF. If this capacitor is too small, the circuit could oscillate. If it is too large, the bandwidth will be limited.

You should document the operation of your circuit by capturing time waveforms from various points in the system using sine-wave and square-wave input signals. Because the Butterworth filter has complex poles, you should observe some ringing at its output with a square wave in. With a sine-wave input signal, the distortion can be measured using the FFT facility in the oscilloscope. The FFT can also be used to obtain the spectrum of the v′o output, i.e. the signal before the low-pass filter. Your report should include all calculations, derivations, and measurements. It does not have to be a formal report.