I am trying to use ADN8831 EVAL kit to drive a 5V TEC. I've been trying different setting (P,PI, PID control) but was unable to achieve stable temperature control. I figured out the main problem is that when the TEC current approach the crossover point (i.e., 0A), it will start to oscillate. So I can never get a curve like normal feedback behavior like overshoot 2 or 3 times and eventually settle down. In contrast, I use a commercial TEC driver to test the same TEC and everything works just fine. I have been troubleshooting this with all the clues provided from the literature (add cap across the thermistor, etc.), but cannot control TEC without oscillation using the EVAL kit. I wonder if you have any experience in addressing this problem. Please advise.
The first attached picture includes waveforms acquired from commercial TEC power driver (Thorlab TED200C). The pink curve is the voltage equivalent of its resistance, and the green curve the TEC current. The second picture include waveforms from the EVAL kit. The pink curve is the thermistor voltage (TEMPOUT) whereas the green curve is the TEC current.