GFCI tester woes

[Acorn]

Hello group,

My husband and I just spent much of today putting in GFCI outlets - 4 bathrooms and 2 kitchen circuits, with only one GFCI in sight, and it not functioning. Ah, the joys of selling your house...

All of our new GFCIs work correctly. Test/reset works, and we have one of those three-light testers that can trip them.

However, when we replaced the original GFCI (that the inspector said was not functioning), the three-light tester does not trip it. The test/reset buttons on the GFCI itself work as expected, and the GFCI cuts power to the other outlet on the circuit correctly. But when we use the three light tester, it does NOT trip the GFCI, instead it lights up as "hot neutral reverse" when its test button is pressed. (NOTE: The tester reads "Correct" when the test button is not pressed, so I don't think I actually have hot & neutral reversed.)

There is a ground wire, and it appears to be connected. I'm not sure how to test whether or not it is actually grounded.

Suggestions?

Thanks! Cathy

[Toller]

I haven't seen this before, but I would check to make sure you have the load and line connected properly.

Is there another outlet beyond the GFCI? Test that. If it trips properly, you have the GFCI wired backwards.

[me]

Best way to see if you got a ground is to take a multimeter. Set it to the OHM scale. Run a long piece of wire in and out windows, or down stairs, or whatever it takes to get it to the breaker panel Hook that piece of wire to any bare metal part of the box. Now, back at the source (outlet). Hook one wire of your multitester to the piece of wire coming from your panel, touch the outlet cover screw with the other lead. You should read ZERO ohms.

There is also a cheater method, but I wont tell you to take a socket and lightbulb and hook one lead to the box and the other to the hot lead, and if the bulb lights fully, you got a good ground. But I wont tell you that trick, because it can be dangerous, and only electricians know that trick.....

[Acorn]

There is another outlet beyond the GFCI. The portable tester can't trip the GFCI from that outlet either. (I know that this outlet is beyond the GFCI, since using the GFCI's built-in test button cuts power to this outlet also.)

More data: I read about 120V between hot and neutral, and 100 volts between hot and ground. I checked other outlets and that's normal as far as I can tell.

Still stumped.

[Chris Lewis]

According to Acorn :

The 100V between hot and ground suggests you have an open ground connection, and you're just picking up inductive coupling via a very sensitive voltmeter.

Check the upstream boxes for broken or missing ground connections. Ie: ground wires that are simply pushed into the box and not screwed down (I've seen a number of contractors think just having the bare wire bare in the box is sufficient).

While a GFCI doesn't need a ground to operate (and is legal on ground-less circuits), I believe "GFCI tester" units need them to test GFCIs. In your situation, you don't have a solid enough ground to let the tester test the GFCI.

[The GFCI outlet test button works in a different way and doesn't need the ground wire to test.]

Assuming that there is ground wire back to the panel, you do need to repair it. If there isn't, to be legal (modulo your inspector disagreeing with the NEC saying GFCI is adequate on groundless circuits), you either need to install a legal ground to it, or, disconnect all the ground wires downstream of the GFCI and label the receptacles "no ground, protected by GFCI" - you should have gotten labels for that with your GFCIs.

[Toller]

If you have copper plumbing, there is normally a connection between the cold water line and ground, and that can be a lot easier than running wires and safer than the method you didn't tell us about.

[Toller]

Chris could be right; the hot-ground should be the same voltage as the hot-neutral, and the tester shunts current to the ground so it doesn't work with an open ground. However, an open ground should show up on your tester (it does on mine anyhow). I tried my GFCI tester on a circuit that doesn't have a GFCI and it doesn't show your hot-neutral reversal. I can't see why an open ground would change that. At this point, I would bet on a defective GFCI. You can swap two and see if the problem stays with the circuit or goes with the GFCI.

[Acorn]

My tester is supposed to show an open ground as well. I'm going to create an open ground on purpose and see if the tester indicates it. (Fortunately, I have a GFCI still not in its box in the kitchen.)

The original and replacement GFCIs both don't trip in this circuit. I will put in a 'known good' one from the kitchen to verify that it is definitely not the GFCI. (I don't know what the original GFCI read with the tester, only that it was defective.)

[Acorn]

Ok, it looks like a grounding problem. I ran a line down to the copper piping (thank you Toller). I can trip the GFCI with a hot to pipe wire, and I read ~120V from hot to pipe with the voltmeter. Looks like that bathroom and the adjacent room aren't adequately grounded. (Reading ~100V hot to ground on both circuits.)

Time to read the code. My easy options seem to be:

1. Hook the ground up to that copper pipe.
2. Tag a whole bunch of 3 prong outlets as not grounded.

If I understand correctly, my best option would be to run a new ground wire all the way back to the breaker box (which is down a story and across the length of the house). That's not very appealing. :(

[Toller]

Assuming it is an open ground, it would real nice to find out why it is open. It has to be a bad connection somewhere; finding the bad connection is certainly the best alternative and might even be the easiest.

If that is not acceptable (perhaps the ground is open because it is completely missing) and assuming running a new ground is not easy, then you have to tag the outlets as not having grounds. There is no big problem with that, as long as they are protected by a GFCI; it is unusual to find anything that actually needs a ground. You might just test your GFCI more often than most people do (which is never).

Running a ground to the water pipe is a code violation. Your electrical system can only be grounded at one point. However, I would do it anyhow, in addition to tagging the outlets. The only time any current would go to the illegal new ground it when the GFCI has failed and also when there is a problem with whatever is plugged in. In that extremely unlikely occurance, I would rather take my chances with a ground loop (the reason you can only ground at one point) than with an electrical shock. Testing the GFCI regularly will keep that likelihood real close to zero.

[w_tom]

You don't have the option of grounding to water pipe. That can even be more dangerous than an ungrounded receptacle. All connections to pipe must be only to remove electricity from that pipe. Connecting a receptacle safety ground to water pipe only dumps electricity into that pipe - bad.

Find and fix the missing ground wire. It is by far the best plan. After all, a partially broken ground wire could even become a sparking hazard - a potential fire hazard.

That 20 volt difference between neutral and ground wire is a powerful tool to isolate and locate the broken ground wire. Measure each receptacle back to breaker box. That will say where the ground is broken between two receptacles. Meter is, by far, your most powerful tool for f> Ok, it looks like a grounding problem. I ran a line down to the

[Doug Miller]

NO! That's *very* dangerous. Attaching the wire to the pipe was for testing purposes *only*, to provide an easy ground for the *tester*.

[RB]

I'd appreciate it if you would explain what you mean by "removing electricity from that pipe." Is electricity like a stain that can be removed with the appropriate detergent or solvent?

RB

w_tom wrote:

[w_tom]

If a hot electric wire should touch the pipe, then where does the electricity go? Through a human hand? Not good. Idea is to connect breaker box safety ground to pipe so that electricity is removed from the pipe. By removing electricity using a good conductor, then hopefully a circuit breaker will trip - disconnecting electricity from that pipe.

One might connect safety ground of an electrical receptacle to a pipe. An attempt to dump electricity into the pipe so that electricity eventually gets back to breaker box. In this case, electricity is dumped into a pipe to connect that receptacle safety ground to breaker box. Not legal and extremely unsafe.

Old acceptable standard: earth ground was made from teleph> I'd appreciate it if you would explain what you mean by "removing

[RB]

Do you understand the concepts involved? I'd urge you to enroll in a basic physics or electrical engineering course. Please don't make it one that I teach though.

RB

w_tom wrote:

[w_tom]

My dear little man. Since it was too complex for you, then you instead attacked the messenger. Children do this.

In the meantime, it is no longer acceptable to dump electricity into pipes. Grounding to a pipe could become one of the most dangerous wiring mistakes made. All wires to pipes are only to remove electricity from those pipes. A simple human safety issue defined by code that you are having trouble grasping - maybe because code book did not tell you why code has changed.

If that is too complex, then this might work for you: Never safety ground to a pipe.

RB wrote:

[Chris Lewis]

According to RB :

w_tom is dumbing down the explanation so far it looks silly.

But he is correct. By code:

Metallic plumbing must be grounded.

However, you are no longer permitted to use metallic plumbing as a grounding conductor for electrical equipment (ie: providing ground path for ground prongs on receptacles].

[Except where buried metallic plumbing can and should be used as an adjunct to other grounding electrodes. If connected properly and in the right place. Yadda yadda yadda.]