I couldn’t say really, my experiment set-up was pretty simple and I didn’t detect any other interference anywhere. The basic configuration was a single LiPo 3S battery charged to full @ 12.4V, connected to a 40A ESC driving a powerful brushless motor at high (but not full) throttle, then just passively collecting data by running the script while I was off doing something else. The connections to everything appeared to be strong and stable, and the motor was being held in place in such a way that wouldn’t be physically interacting with the electronics / voltage divider.
Probably a stretch, but I did keep getting low voltage warnings on the Raspberry Pi because at the time I was using a USB battery pack that couldn’t provide stable 5V. I swapped that out and resumed the test with data I didn’t publish and strangely saw the cliff show up for only a few seconds then return to the normal linear voltage drop. In theory this could’ve affected the timing of when I was reading the voltage divider values resulting in noise, but this wouldn’t explain the cliff.
I can set-up another experiment and re-run soon just to see if I get consistent results. Though one of the things that gives me pause is why the voltage divider would think I’m getting 13V+ in the first place when two separate reliable measurement devices (my balance charger and voltmeter) both say it is exactly 12.4V. I also can’t understand why the readings are so sporadic even when things are working normally (e.g. around t=3000 a bunch of samples start randomly showing up in the 12.4-12.8V range even though the majority of samples are around 13.0-13.4V).
Any insights would be greatly appreciated!