Sometimes, the only rational way to isolate the problem is to start disconnecting things until the remaining circuit works properly. Once you've isolate a branch, continue with disconnecting things or bypassing individual items until you're only left with one. While a GFCI device can fail, if it turns on and trips when you hit the test button, while not an absolutely 100% check, it's close...swapping it with a new one where the same thing happens is pretty much 100%. Then, it's back to trying to isolate the problem. First thing is to carefully open and do a very thorough visual inspection of everything that's accessible. Understanding how the things works helps. You'd never want the neutral and ground tied together after the panel.
Important note - I'm not a pro
Retired Defense Industry Engineer; Schluter 2.5-day Workshop Completed 2013, 2014
Most real electricians should have a o-scope and a megometer available to them. A VOM does not tell all.
Handheld scopes are cheap and used for working on 3 phase systems and looking at line noise. Meggers are great for testing wire runs and true ground faults. You can also use a Ground Fault tester.
Many "residential electricians" are wire runners and do not have the proper tools for testing, but they can follow color codes and wiring drawings. No need to even read or speak English.
quote; The neutral is not shared.
Are you absolutely certain that there is NO OTHER circuit using the neutral wire? Does the neutral go DIRECTLY to the panel box so you can be sure of it.
Licensed residential and commercial plumber
It sounds like there was a Neutral connected to a Ground somewhere, or a wire was damaged like JW mentioned. Or the GFCI was bad.
A ground to Neutral connection would show a problem if a Load was enough to drop the voltage and the GFCI could detect it.
With no or little load the GFCI can not see the difference because the Ground and Neutral are bonded at the panel, and the potentials are the same with no load or voltage drop on the wire run.
I think the problem is known and has been fixed, this was just a test.
P.S. The neutral may have been shared, with a ground.
Last edited by DonL; 03-22-2013 at 11:08 AM.
In my limited experience with residential wiring (I am not a pro in that area), it's not uncommon for neutrals from different circuits to be intentionally tied together (obvious DIY hacks), but in that case, a GFCI breaker *should* always trip with a load on either circuit, and in theory possibly even without any load at all -- depending on the design of breaker.
I still think what you describe sounds like a flaky breaker. Probably a manufacturing defect...
As I mentioned before, this was an issue that I resolved some time ago. The reason for my post was to gain insight on how others would go about diagnosing such a problem.
It was a neutral to ground short, and while my ohmeter showed infinity, the megger did not.
I started at the panel and spit the circuit in half twice before being confident that it was in the lights. Unfortunately, I had to cut apart all 4 recessed cans to pinpoint the fault. The lights were split wired with 12-3 from 2 switches, and the short was located where the last recessed can junction box NM cable clamp was tightened on the cable.
After replacing the lights, I loaded the circuit over 20 amps and the GFCI breaker never as much as whimpered.
Last edited by cacher_chick; 03-22-2013 at 06:35 PM.