Ground Leakage Current Tripping GFCI While Charging

Lately I've been charging my 2018 Prime from a different GFCI receptacle than the one I usually use. The GFCI has been tripping after about an hour of charging. Every time I reset it, it will charge for another hour or two then trip again. And in fact I have two identical (and independent) GFCIs side by side, and the same thing happens with both.

I could just replace the GFCI, but I wasn't satisfied with that. So I decided to measure the ground current from the EVSE back to the wall while charging. It's over 5mA, which is right in the 4-6mA range where a GFCI should trip. This suggests the GFCI is working normally, and the car or the EVSE has excessive leakage current to ground while charging. Interestingly the meter is showing 26.62kHz, not 60Hz as I expected. So maybe this current is from a switching converter in the car's onboard charger? Without an oscilloscope it's hard to say. Also, I'm not sure if the meter measures true RMS, but either way if it's a pulsed waveform, a reading with a meter doesn't give much info.

So why does it charge for an hour or two? Maybe the GFCI becomes more sensitive as it heats up, or maybe the leakage increases as electronics in the car heat up. I made this measurement with line and neutral going through an old lamp cord that quickly got hot, so I can't leave it for more than a minute.

Why do other GFCIs work? My only guess is they have more filtering and don't respond to this 27kHz current the same way. These GFCIs I'm currently using also trip half the time I turn on my table saw, which I haven't experienced with other GFCIs. They are some old (10 years?) GE GFCI receptacles that came with the house. I haven't had any issues with false tripping from Leviton or Eaton receptacles that I've bought new.

The only other possibly related symptom is that the car sometimes doesn't charge at public charging stations. It will either not charge at all, or stop charging randomly, with no useful error from the car or the station. I just assumed they are broken/unreliable, but maybe it's actually the car?

I think I did see a ground fault error or something like that on the display in the car once or twice in the time I've owned it, but I don't remember exactly what that said. It isn't showing that error now. The car is well isolated by its rubber tires, so there's no other path for current to ground besides the ground wire in the EVSE.

Has anyone else seen GFCI tripping while charging? I'm using the factory EVSE. It would be interesting to try a different one, but they are pretty simple so I don't think the EVSE is the problem.

The easiest solution is still to replace the GFCI or use a non-GFCI circuit, and that's eventually what I'll do. Without reverse engineering the onboard charger, I don't think I can find or fix the source of this leakage current.

 

The garage has two receptacle circuits, each with only a single GFCI, nothing else downstream or upstream, and no other receptacles on those circuits. The GFCIs are old. I know they don't last forever, but I've usually only seen them fail to trip. Haven't seen one become too sensitive until now.

I replaced one GFCI, and as expected it works fine with the new one. Doesn't change the fact that I measured greater than 5mA on the ground pin of the EVSE. I still think the car has marginally high leakage current, but it happens to work on most GFCIs.