Ungrounded Handy Box

[crhras]

Just spent the last two hours trying to figure out why the GFCI my electrician installed in my new building doesn't trip correctly (that includes at trip to the shop to get a new GFCI). I found out that the box it is using as a ground isn't grounded. Another box 2 feet away is grounded just fine and the GFCI housed by it trips under all tests.

What do electricians do in this case ? There's conduit to the box and I'm wondering why it wouldn't be grounded in the first place. I'm thinking of pulling a third copper wire and using that to ground the outlet but that probably isn't the best solution.

Thanks

[dpb]

If, as it sounds, you're talking a metal box and conduit, it should be grounded. About the only way it can't be if the box feeding it is is if the conduit isn't continuous or there was a plastic fitting used somewhere. My first preference would be to fix the conduit rather than pull another ground.

--

[Bobk207]

There must be something else going on.

A GFI will trip even if not grounded properly. That's why the GFI comes with stickers that say "no equipment ground".

A GFI compares the hot current to the neutral current & based on a difference; trips.

I've installed GFI's in old houses w/o grounds (& applied the sticker) ......they trip just fine.

cheers Bob

[spebby_92]

Is the power souce connected to the load terminals on the GFCI?

[crhras]

The GFCI trips just fine when using it's self test button. It doesn't trip when using the GFCI test button on the small outlet tester I am using. All other (10 or so) GFCI outlets in the building trip when using the tester on them but some of them didn't before I grounded them to the boxes they are installed in.

The tester has three lights on it - 2 yellow and one red and when an outlet is well grounded to the box the 2 yellow lights are bright and the red one is completely off. When an outlet doesn't seem to be grounded the two yellow lights are lit but not too brighly and the red one glows a bit. The instructions that came with the tester don't address what that means. I am assuming that something is not right.

So, after doing everything possible to ground the GFCI to it's box I then used a meter to measure the voltage difference between the hot lead and the box. The voltage measured only 60V whereas a different, working box measured 120V. That's why I assumed the box is not grounded.

Finally, I just hanged a copper wire from the ground screw of the non-working GFCI to a good box and it then trips correctly.

Thanks the responses, Curt

[crhras]

No, I am using the line terminals and other GFCIs in the building are connected exactly the same way but trip differently.

[crhras]

That would be my first preference also but I can't imagine how one would find where the conduit is open. I mean, it could be anywhere in the wall, right ?

[Bobk207]

Curt-

Good description of the situation.

Here's what I think may be happening..........

Your GFI tester works by generating a "true ground fault"....that is, a small current leak to ground to simulate a problem (like you or someone using power from the GFI about to get shocked).

So your tester cannot do this IF a ground does not exist.

I'm thinking that the GFI is fine but the method of test used by the tester cannot work without a ground local to the GFI receptacle.

Other way to test the GFI is to get someone to stand bare foot on a wet surface & stick a paper clip into the hot of the GFI................j/k don't do this. :)

cheers Bob

[crhras]

So, you're saying that the GFCI is probably working just fine and I can ignore the results from the handheld tester? If that's what you are saying then I think you are right but man, what good is this tester then ?

[PeterD]

Possibly it is working 'correctly', that is as designed. There is no promise it is going to provide protection

Only at your own risk.

See above! The tester serves a useful purpose, you discovered that.

IMHO, and hell, I'm an idiot (but I do have an EE degree), one should

*NEVER* rely on conduit to provide ground paths. A seperate ground wire must be installed. Relying on conduit to provide grounds leads to floating boxes, outlets, GFI's that don't work, ARC fault interrupters that don't work, and other strange and sometimes dangerous situations. Pull a ground between the two boxes... But then again, I'm just a guy on the Internet, so what do I know?

[Bobk207]

Peter-

Could you expand on your comment....

promise it is going to provide protection

[Doug Miller]

The authors of the National Electrical Code do not agree.

Not if the conduit is installed properly.

It certainly does no harm to pull a ground, and there's no doubt that doing so makes a better, more reliable installation. But it isn't necessary, and it's not enough better than properly installed conduit to warrant the extra expense.

[Doug Miller]

Give that man a cigar.

[PeterD]

I know the NEC doesn't agree with me... But I've seen too many conduits installed with shoddy workmanship, in high moisture/corrosive environments, with gaps (!!!) in the middle, and all sorts of bad stuff.

I'm perhaps a bit conservative, I don't use (or allow to be used on work that I'm involved in) push on terminals either--must use the screw terminals.

I really think the original post really reinforced my feelings: the conduit wasn't making a good ground path between the boxes. Wire is cheap, but one's life and the lives of others is not, so I always insist on a pulled ground wire in these cases.

but, thanks for your comments!

[Doug Miller]

No correction needed -- you're not wrong. There's an excellent discussion of how GFCIs work here:

[Doug Miller]

Unfortunately not everyone takes enough pride in his work to get the job done right, and done right the first time.

I heartily agree with that.

But the GFCI will operate correctly even if there is no ground at all; they simply don't need one. They operate by sensing a difference in the currents on the hot and neutral conductors. If the current going out is not the same as the current returning, it must be leaking somewhere, and so the device trips.

[crhras]

Thanks for your interest in my problem. I learned alot from this thread. To solve my problem I grounded the box. I know that I probably didn't need it for the GFCI to work but I wanted to err on the side of safety.

Curt

[PeterD]

And that was probably a very good move.