Could there possibly be enough inductive reactance to reduce the AC current to borderline? I really question the mfg. tolerances on GFCIs. I faulted hot to ground through my Fluke and a variable resistor on a GFCI at work. It didn't trip until about 9mA, which was more than I expected. I had tested one this way in response to a thread here where someone had suggested using a Wiggy to test a GFCI. I seem to remember the Wiggy flowing a current several times that considered lethal.
Joe
