GFCI on ungrounded system (system grounding, not equipment grounding)

[kwired]

That's an interesting point... Do you mean that in the sense that it can provide a return path to the source and actually create a path for the unbalanced current to flow? Seems though to a certain degree this would provide a more "reliable" path for the unbalanced current than the current that would otherwise flow from capacitive charging.

This is all interesting to me. Bonding is critical to have a low-resistance path to the source to clear ground-faults on grounded systems. We want high levels of current to flow to trip breakers. But now it seems grounding, though it does nothing to trip a breaker on a fault, could be beneficial to trip a GFCI considering the little bit of current it requires... Granted choosing a grounded system solely for this reason is unlikely.



Which leads me to this...

By what standard? NEC? Or are you saying only recognized as such by the manufacturer?

The rod to earth resistance is nearly always too high to give a reliable path to facilitate operation of overcurrent devices or even to carry normal circuit loading. 10 ohms would be considered a good ground rod install, yet it will limit a 120 volt fault to 12 amps - never going to trip a 15 amp breaker.

GFCI - it certainly will make them trip almost immediately in most cases of "ground fault" if you have a grounded system.

If you don't have a grounded system - you need a second ground fault or there is no fault current, but remember the purpose of GFCI is preventing electrocution - there is no shock hazard on first ground fault in an ungrounded system - all that first fault does is establish a ground reference to the system.