Inductive whistle problem in digital phase-shifted full-bridge circuit, please help

菜鸟003

Inductive whistle problem in digital phase-shifted full-bridge circuit, please help [Copy link]

Today, I did an open-loop experiment. The sampling data processing and PID loop calculation were placed in the AD interrupt service function. Only the interrupt processing data was performed. PI only calculated but did not change the value of the phase register. By online debugging and then changing the register value, the phase shift angle was changed. When the load increased, the phase shift angle was changed to make the output stable at 110V. There was no inductor howling. However, after turning on the phase register assignment statement for the PI loop calculation (that is, when the closed-loop control was realized), there was no problem with the startup. However, when the load was added to 4.8A, inductor howling occurred. The output voltage jitter was shown in Figure 3. The output voltage jittered with a frequency of 20ms, which made it impossible to add the load. Figure 2 is the waveform when the load is 7A in the above open-loop case. Yellow and pink are the drive waveforms of Q1 and Q4, green is the output voltage, and blue is the output filter inductor current waveform. Figure 4 shows that a level of feedforward control has been added to the sampling port, which is equivalent to adding a zero point to the system and improving the phase margin. Is this the reason why my output voltage is jittering? The digital PI controller is equivalent to adding a zero-frequency pole and a zero point to the system. When both zero points exist, the phase and amplitude will be improved. In this way, after compensation, the crossing frequency will be affected, or it may not be crossed at -20dB, so I don't understand this part. Figure 5 is the flow chart of the program I wrote. The previous timer interrupt is the soft start part. The PI calculation and AD sampling are all executed in the AD interrupt. I hope the power supply master can give me some advice. Thank you. My circuit parameters and topology are the topology of the primary side series DC blocking capacitor and the lagging arm series blocking diode, as shown in Figure 1. The output rated power is 110,20A, and the front stage is a three-phase uncontrolled rectifier.

2020.6.2 Change the PI parameters and continue the experiment. First, manually change the sampling value without powering on (that is, use an adjustable voltage regulator to make the sampling port change around 2.5V to observe the change in the driving fluctuation phase shift angle). The test KP and ki ranges are shown in Figure 6. Then power on the test and test within the PI parameter change range to obtain the inductor howling range shown in Figure 6. The smaller the integral, the smaller the range of the phase shift angle displayed by the oscilloscope. If the integral increases, the range of the phase shift angle in the steady state becomes larger, and the driving wave also begins to shake. Have you encountered this problem again?

maychang

"The output voltage jitters, and the jitter frequency is 20ms"

The frequency of jitter cannot be time, we can only say that the period of jitter is 20ms.

That is to say, the frequency of the jitter is 50Hz, which is exactly the power frequency. But the original poster said that the "pre-stage is a three-phase uncontrolled rectifier", and the ripple frequency after rectification is 300Hz. So it is unlikely to be the influence of "feedforward control".

maychang

My guess, just a guess, is that the output jitter is caused by inappropriate calculation parameters of the OP's PID loop.

菜鸟003

maychang published on 2020-6-2 15:47 "The output voltage jitters, and the jitter frequency is 20ms." The jitter frequency cannot be time, it can only be said that the jitter period is 20ms. This...

Feedforward control has no effect, so what is the reason? Can you help analyze it?

菜鸟003

maychang posted on 2020-6-2 15:49 I guess, just guess, the output jitter is caused by the inappropriate calculation parameters of the PID loop of the original poster.

I also suspected that it was a problem with the PI parameters at first. In fact, my KP was 0.003 and ki was 0.001, which seemed too small. I didn't dare to adjust it in a large range. When I tested the phase-shift drive wave again, if KP was very large, the drive wave would vibrate when testing the microcontroller. I thought that the drive wave would start to vibrate if I didn't run the circuit but just changed the parameters. If I changed the parameters significantly when the power was on, would the drive wave vibrate more severely? There is no problem using 28335 for open-loop experiments. It can reach 20A, but the teacher can't add a closed-loop load.