GFCI tripping with no load when connected to power

[Jadnashua]

They're designed to trip when they see about a 5ma (0.005A) or more delta between the hot and neutral...enough to potentially feel, but not necessarily hurt you under the right circumstances. In a properly working circuit, all the current that goes out on the hot should return on the neutral, thus, the loads are even. AN old motor which has a lot of dust from the brushes, or some florescent lights, and other things can leak but not trip a breaker. They can trip a GFCI. That's why some people say they'd never use a GFCI protected circuit on a refrigerator or freezer...the GFCI could trip. But, if they're working properly, it never should. If they've worn enough where they do trip one, they're a safety hazard that should either be repaired or replaced. But, a lot of old ones get relegated to the basement or garage, and eventually will have issues. Things don't last forever, and the spoiled food is cheaper than a funeral.

 

[Mikey]

I just discovered this (in http://www.ecmweb.com/content/think-gfci):

"...the UL standard requires that GFCIs trip with a 6mA ground fault even when the neutral and ground are connected. To meet this requirement, GFCIs trip when the load side neutral and equipment grounding conductors are joined, even if there is no ground fault."

So my first question is answered, second Q not so much. Lots of other GFCI detail in that reference I've got to chew on for a while, but this next paragraph got me wondering if I might have had the problem described - there were bare grounds in there on the neutral side of the box, but I don't recall them being in a position to contact any neutral screws. Still, it was just the mechanical action of plugging/unplugging a plug in a socket that tripped the GFCI, so an infinitesimal jarring of a close-tolerance situation might have been the problem:

"At one time, it wasn't uncommon for electrical tradesmen to be instructed to reduce the possibility of accidentally shorting out the hot conductor to the bare grounding conductor in the box by placing the bare ground on the side of the receptacle with the neutral screws. When the receptacle is placed in the box, the bare ground is less likely to accidentally short out against the hot screws. This may have prevented shorting, but when GFCIs came along with the grounded neutral detection circuitry, the bare ground touching the neutral terminal screws caused an automatic trip of the GFCI. The solution here is to dress the wires carefully to assure that the bare ground does not contact any of the terminal screws."

 

[Sponsor]

 

[Jadnashua]

If there ends up being current in the ground lead, the GFCI will become unbalanced and should trip. People tend to forget or don't know that while the neutral might be bonded to ground back at the panel, it is a current carrying conductor. The ground lead should never be a current carrying conductor, it's there if there's a problem and provides a path to ensure the CB or fuse trips when it should. If ground and neutral are connected downstream, there could be current on the neutral bus induced or actual that doesn't match that on the GFCI's inlet hot side. Because of the varying paths, and that the neutral is rarely balanced between L1 and L2's loads, a bit of current could flow between the neutral and ground if they're tied together at other than the one place back at the panel.