"Exactly the same thing" appears to mean (educated guess about the failure situation) that you've had a nearly full battery, then started regen braking at a high enough current long enough to fill the battery and cause it's BMS to shut off, which cuts off the current, which means the voltage goes up extremely rapidly (almost instantly).kdog said:
The electronics in the controller have an HVC beyond which it will disable regen, but there is a delay, even if it's small. And the motor itself still has current flowing thru it's coils even when the regen cuts off, and if it was enough current it could spike the voltage higher than the FETs can handle, and POOF.
So maybe changing the controller's regen HVC to be lower, so it reacts sooner. Preferably lower than the HVC of the battery, so it never causes the battery to reach HVC during a regen event, so it won't cutoff.
I'm going to guess that this is how the systems you're using are setup, so they don't actually cause the battery's BMS to cut off, and instead your controller is turning regen off before then.
It is possible that the non-failed controllers do have a different hardware design that prevents the problems, or simply differetn FETs that have a higher voltage tolerance than the ones in the failed controllers.
If it's the latter, you can fix the problem by replacing the FETs with higher-voltage rated ones.
