although when it moves around slowly it looks exactly like what happened to us on a batch of boards..
There was this issue where a PWM signal was filtered with a two pole Sallen Key and the guy who drew the schematic forgot to put the capacitor between the positive input of the opamp and ground on one channel.
It was ridiculous! the schematic was thight, if he had just copied the filter from another channel the infamous capacitor would have been on the title block, so he must have removed it never to add it back again
Anyway, the issue was that besides worse filtering of the signal there was a heavily distorted oscillation when the duty cycle changed between zero and about 20%, of the same kind i saw on the video. I would inspect the calibration signal circuit to see if there are cracked caps or something. every time i did self calibration i never saw a signal like that