Taylor Murphy & Remo Panella EE 333 12/12/18 Project 1 Final System Design and Performance Report Class D Amplifier Intro For this project, we designed a class D amplifier circuit. Class D amplifiers work by passing the input signal through a comparator with a triangle wave. This then generates a pulse width modulated wave that is then passed through a transistor switching stage. In this switching stage, the transistors are sized so that only one is on at a time, therefore improving the efficiency of the amplifier. This signal is then passed through a low pass filter to obtain the amplified output. We initially wanted to construct the circuit using only discrete parts, but were informed the switching stage is very difficult to design. Instead, we opted to use a class D chip to simplify the circuit. The main component of the circuit is the TPA3125D2, a Texas Instruments Class D amplifier chip. The circuit is powered by a wall transformer and drives stereo speakers. There is a mute and shutdown function that is implemented through switches. There is also gain control with four different gain settings which are 20, 26, 32, and 36 db. This report will break down the design and performance of our circuit.
Circuit Schematic Schematic Details Power Supply: The circuit is powered off a 16.5 VAC wall transformer. This is then passed through a full wave rectifier and smoothing capacitors and then an 18 V linear regulator. This 18 volts provides power to the left and right channels of the chip. It also is used with the mute, shutdown, and gain switches.
Mute and Shutdown: The Mute function is enabled when the switch is closed, connecting the mute pin to the voltage supply. The Shutdown function is enabled when the switch is closed, connecting the shutdown pin to ground. Gain: The gain if the circuit is controlled by two switches. Closing the switches increases the gain by connecting the respective gain pins to the voltage supply. The gain settings are 20, 26, 32, and 36 db. Output Filter: There are two output filters in this circuit, one for each channel. This is a low pass LC filter mainly consisting of a 33 uh inductor and a 0.22 uf capacitor. There are also resistors to bleed charge off the capacitors once the circuit is shutdown.
Design Details The main focus of this design was behind powering the chip and filtering the output. The power supply for the chip is 18 V. We obtain this by using a 16.5 VAC wall transformer. This is then passed through a full wave rectifier and smoothing capacitors. This is then passed through a 18 V linear regulator that then powers the chip. The output filter of this circuit is based on recommendations for the chip and output configuration we are using. Since we are using a single ended stereo output, Texas Instruments recommends an output filter of a 33 uh inductor and a 0.22 uf capacitor. This provides a cutoff frequency of around 59000 Hz. Below is the pin diagram for the chip. Pins 1 and 10 are the power supply for the left and right channel switching stages. They are connected to the 18 V linear regulator with bypass capacitors. This is an attempt to remove the hum from the power supply so it can t be heard at the output. Pins 2 and 3 are the
shutdown and mute pins and they are connected to switches. These switches allow these pins to be connected to ground or the power supply to provide logic low and high. Pins 4 and 5 are the left and right inputs and are connected to the input through input capacitors. These are recommended by Texas Instruments to improve the low frequency response of the circuit. Pin 6 is the bypass pin and it is connected to ground through a capacitor. The size of this capacitor affects the startup time of the amplifier. With a 1 uf capacitor, start up time is around 500 ms. Pins 7 and 8 are analog grounds that are connected straight to ground. Pin 9 is connected through a capacitor to ground. This 1 uf capacitor is required as it allows an internal regulator to clamp the gate voltage for the output transistors. Pins 11 and 20 are the power ground for the switching stage of the left and right channels. These pins are connected to ground. Pins 12 and 19 are the output for the left and right channel. This output is in the form of a pulse width modulated wave. This wave is then passed through our low pass filter to obtain the final output. Pins 13 and 18 are connected to the left and right outputs through bootstrap capacitors. These bootstrap capacitors are required to allow the switching stages to turn on correctly. Pins 14 and 15 control the gain of the amplifier. They are connected to switches that either connect them to ground, or the power supply. This allows us to vary the gain to four different settings, 20, 26, 32, and 36 db. Pins 16 and 17 are analog voltage pins. These are connected to our power supply and these pins provide power to the comparators inside the chip. Performance Details
In order to observe the performance of the circuit, we first tested it with sine waves from a signal generator. We tested with an input signal with a peak to peak voltage of 220 mv. At the left and right outputs of the chip before the filter, we observed the pulse width modulated waveform. There is some aliasing in this waveform when the output goes from low to high. Below is an image of the output before the filter. We then observed the output after the filter. We used a 2 khz, 220 mv peak to peak sine wave as the input. We looked at the output under our 4 gain settings, 20, 26, 32, and 36 db. In these outputs, we don t see a perfect reconstruction of the sine wave. The filter doesn t seem to perfectly filter out the high frequency pulse width modulated waveform, so the output looks like a sine wave modulated with the pulse width modulated waveform. This filter doesn t distort the frequency of the output signal. Below are the images of the output signals for the various gains.
Gain: 20 db
Gain: 26 db Gain: 32 db
Gain: 36 db On the highest gain setting, we begin to observe clipping. We start to approach an amplitude close to 18 V, which is our power supply voltage. This causes clipping and distortion. Below is an image on the highest gain setting where the clipping is more evident. You can see at the top of the waveform how it is beginning to be distorted. Although the output looks quite distorted, when testing the circuit with music instead of a sine wave. The output when testing with music is not distorted and sounds fine. Video of our circuit testing with music can be found in presentation 2 on the documents page of the course website. We also tested the circuits mute and shutdown capabilities. The mute effectively turns off the left and right output channels so nothing can be heard. The shutdown feature also works well and puts the chip into a low power state. It is important to use the shutdown
feature when powering the circuit. We found that starting the chip in shutdown mode, then applying the input audio signal, then turning off shutdown mode reduced popping and allowed for a better startup. We measured the input and output power in order to see how efficient our circuit was. We found the input power to be 6.75 W and the output power to be 2.67 W per channel. This leads us to an efficiency of around 80%. Overall, the circuit works well and is capable of playing music quite loud. Below is an image of the final prototype. Conclusion This circuit we designed operates as a decently efficient class D amplifier. We ran into a couple different issues when prototyping the circuit. We initially had a 24 VAC wall transformer
and a 24 V linear regulator in our power supply stage. This led to issues with the linear regulator because we were trying to drop the voltage from 34 V to 24 V. This led to a lot of heat being dissipated in the regulator. By switching to a 16.5 VAC wall transformer and a 18 V linear regulator, this heat problem was solved. We also had a hum problem when testing the circuit with speakers. The hum wasn t incredibly loud, but if your head was maybe 1 foot from the speaker, then you could hear the hum. This issue was being caused by a ground loop. By properly grounding the circuit, the hum was reduced. There is still a slight hum present, but this is caused by the 120 V, 60 Hz signal going into the wall transformer. We recently found an effective way to almost entirely remove the hum. We increased the capacitance in our power stage and properly grounded the audio inputs through 3.3 kohm resistors (These changes can be seen in the above image. large 2700 uf capacitors in the upper left, resistors from audio input to ground in bottom right). This made the hum much less noticeable. There is also the issue of the distortion from the pulse width modulated wave. We spent quite a bit of time trying to figure out how to remove this distortion. We tried various different filter components to adjust the corner frequency and remove the distortion. This slightly remedied it but the distortion was still there when testing with the signal generator. Once we tested it with speakers and an actual audio input and realized that it wasn t distorting the output, we just assumed it was how the class D chip was meant to operate. Overall, the circuit performed as desired and is a pretty cool audio device.