After all of this discussion I was wondering how well reality matches the theory and did some testing and measuring with the motorhome plugged into a GFCI outlet.
GROUNDED NEUTRAL TEST(fault created inside inside the RV)
With no AC loads, GFCI trips at 11 or less ohms fault
With refrig on, GFCI trips at 73 or less ohms fault
With refrig and charger on, GFCI trips at 120 or less ohms fault
With the shore cord ground broken, GFCI does not trip at zero ohms fault
HOT CHASSIS TEST (RV ungrounded)
With main breaker off, chassis to ground voltage 1.88, short circuit current 0.15 mA
With main breaker on but no AC loads, chassis to ground voltage 6.7, short circuit current 0.28 mA
With fridge on, chassis to ground voltage 7.2, short circuit current 0.37 mA
With fridge and charger on, chassis to ground voltage 41, short circuit current 7.9 mA, AND primary frequency of the chassis to ground voltage is 20 kHz, the switching frequency of the charger
In theory the maximum current one could be shocked with due to the ungrounded RV would be 7.9 mA. One question I have is with the 7.9 mA short circuit current between hot and ground, why didn't the GFCI trip? Possible answers are (1) There's an offsetting current from neutral to ground that keeps the total of the two currents under 5 mA, (2) The GFCI doesn't respond to the 20 kHz current, (3) Something not obvious to me is in play.
Info: Measurements to ground are to the ground wire in the extension cord powering the RV. The DC resistance of the RV to ground is greater than 20 megohms. The charger is a prosine 2.0. I don't know what chip and/or technology is used in the GFCI outlet. It's about 12 years old.