How old is the GFI? Are you absolutely certain you got LINE and LOAD correct?

Also, WHY did you protect the lighting?

Everything is brand new.
Line is fed from circuit breaker. Load is to switches.
I protected the lighting because the fan/light is above the shower/tub. I suppose I didn't have to protect the vanity light.

How old is the house and wiring?

A GFCI does not require a ground to operate, and the test button on the device will still work.

A GFCI tester requires a ground to check GFCI operation.

Even if there is a ground in the bathroom wiring, it may not go back to the panel and actually be hooked up.

It is an older house, but new panel. Besides, I ran the new circuit and breaker. It is definitely grounded.

Does the real loads (not the tester) go off when you push the test button (either on the GFCI itself or the tester)? The little neon bulbs in these things can sometimes light just from stray coupled voltage.

When I press the TEST button on the GFCI, it kills the circuit. When I click the button on the tester, it does not - everything stays on and the lights (on the tester) indicate that there is an open ground. I can not find any open ground. What are the chances that the tester or the outlet are defective? Are there any other things that could be wrong that would make my tester fail and indicate an open ground condition?

Does this tester behave properly on another known good receptacle?

Do you have a grounding wire going to the screw on the GFCI receptacle, or are you counting on contact through its mounting yoke (not approved and perhaps not working)?

Can you rig up a light bulb tester connected to the hot (black) and ground (green) of an old power cord and see if that trips the GFCI?

Yes, tester works on other GFCI outlets (in kitchen).
Yes, the line cable's ground wire is connected to the green screw on the GFCI. The two other cables' (the ones going out from the switch to the lights) ground wires are pigtailed and connected to the ground screw on the double switch.
Can you explain in a little more detail the last paragraph. Are you saying plug in an old power cord into the outlet and then check the wires?

If you aren't comfortable with that kind of experimentation, don't try it. The idea was to find a light socket and power cord that could be wired together in a non-standard way in order to do a test here. The light bulb would be connected between hot and ground. Plug it into the GFCI receptacle. If it lights without tripping the GFCI then the GFCI is bad or something isn't wired right. If it trips, then maybe the GFCI just isn't quite as sensitive as most of them are or the push button tester doesn't draw as much leakage as it was intended to so that it takes a more sensitive GFCI to trip. If it doesn't light but the GFCI doesn't trip either, then you don't have a real ground connection through the GFCI receptacle.

I don't have a light socket handy, so I haven't tried that yet.

I did check the wires with the voltmeter. Both white and ground came back exactly the same - roughly 116 volts.

I noticed that when I press the GFCI test button on the TESTER in any outlet that is NOT GFCI, the lights indicate "open ground." So, I'm thinking that maybe the lights don't really mean anything in this case - the tester is just not recognizing this as a GFCI. As billhart said - "maybe the GFCI just isn't quite as sensitive as most of them are or the push button tester doesn't draw as much leakage as it was intended to so that it takes a more sensitive GFCI to trip."

Should I forget about it and assume everything is okay, or should I try the billhart experiment, or just buy another GFCI outlet and try that?

Or try another GFCI tester.

[Use a volt meter and measure from the hot to the ground of the receptacle.]

In a properly functioning GFCI this will cause the device to trip.

sumac - I'm really glad you mentioned that because that exact question crossed my mind. It seemed logical to me that if I was completing a circuit using the ground wire, then the GFCI should definitely trip.

I bought a new outlet and a new tester this morning. I'll be doing some more testing later on this evening.

In a properly functioning GFCI this will cause the device to trip
A voltmeter will not place enough of a load on the ciruit to trip a GFCI. A solenoid type tester (Wiggy) will.

I switched the GFCI outlet- exactly the same results, so we can rule that out.

I tried the new tester that I bought. This one worked on the GFCI outlet, but not on the outlet downstream! The outlet downstream, by the way, is an outlet that I attached for testing purposes. Later, this will be the vanity light.

I'm getting a little frustrated now. I'm not too proud to say that it sounds even to me like it must be a wiring problem, but there is so little here, I don't know what I could have done wrong.

Here is exactly what I have - this is all new wiring from the panel:

• 12/2 from panel to GFCI LINE (black, white, & ground to proper screws)

• LOAD side of outlet

  • White - pigtailed (two loads - vanity light & fan/light)

  • Black - common terminal of double switch (tab still attached)

• Two ground wires from the vanity and light/fan are pigtailed and attached to ground screw on double switch

• Two black wires (from vanity and light/fan) are attached to the two screws on the switch

So, to sum up one more time - everything appears to be working fine. The switches properly control the vanity light (I have a regular duplex outlet attached temporarily) as well as the fan/light above the tub. The outlets have power and the testers indicate that they are wired properly. The TEST button on the outlet works fine. The concerns that I have are:

2. The "old" tester that I have won't trip the outlet; however, it works on all of the other GFCI outlets that I have elsewhere in the house.

3. The "new" tester trips the GFCI outlet when plugged directly into it; however, it does not trip when I plug it into the outlet downstream (which will later be the vanity light).

Yes, I know that the vanity light does not need to be GFCI protected, but I won't feel comfortable if I don't understand why this isn't working. The shower fan/light is supposed to be GFCI protected, so how will I know if it really is?


BTW - the GFCI outlet is 20Amp and the circuit breaker is 20Amp

I can't tell from your posting.....

""12/2 from panel to GFCI LINE (black, white, & ground to proper screws) ""

""Two ground wires from the vanity and light/fan are pigtailed and attached to ground screw on double switch ""

All these grounding conductors from the GFCI and the switch need to be connected together...Are they ?

At the VanityLightTempRecpt.....check for voltage between Hot and Ground.........

From your text, it sounds as though, when you have multiple devices in the box, you've been attaching the "line" ground wire to one device and all of the "load" ground wires to another.

This is not how it's done. They ALL have to be connected together.

Hopefully I'm misunderstanding you - because, if I'm not, you now have to go back to every single box in your house that has multiple devices in it and fix them all.

The box is metal, so originally, I figured that they would all be connected via the outlet yokes and the box itself. However, to eliminate any question, I did add another wire to tie all of the ground wires together.

Is it likely, or possible, that the GFCI outlet is just not sensitive enough? Even though I bought a second one, maybe it's the way they are manufactured. Should I bother trying a differnet brand/model?

At the VanityLightTempRecpt.....check for voltage between Hot and Ground.........

Yes, I tested the hot/ground with my voltmeter and got the same reading as I did on the GFCI outlet.

12/2 from panel to GFCI LINE (black, white, & ground to proper screws)

LOAD side of outlet

White - pigtailed (two loads - vanity light & fan/light)

Black - common terminal of double switch (tab still attached)

You need the ground from the load side of the GFCI to attach with the other ground to the line side of the GFCI.

All the grounds get connected together on both sides of a GFCI receptacle.

If you are not hooking the load side grounds to the line side ground in the GFCI box you have dropped the ground.

The GFCI function downstream will operate correctly (it does not need a ground) but the GFCI tester requires a ground to leak current to trip the device.

Thanks to you all. You all steered me in the right direction. It was definitely due to the grounds not all being wired together. I was relying on the yokes and the metal box, but apparently that is not sufficient.