Another GFCI problem!

Factual Questions
1

Disclaimer: I know that electricity is dangerous - I’ve done some home wiring before. I know that advice from a message board is just that - advice from strangers on a message board. I realize that if I eff this up it’s my own fault, and that I may electrocute myself or burn the house down. I take all precautions - turn off appropriate breakers, test for current before touching anything etc. etc.

That said: I’m having a problem installing a new GFCI receptacle. I’m pretty sure I’ve wired it correctly - I read the other thread on here, and I’ve followed the instructions. I want to wire it to protect the downstream devices, so roughly my topology is this:

[Breakers] --> GFCI --> other outlets

I’ve identified which wires are line & load (using a tester) and connected them to the appropriate terminals on the GFCI, as shown in the instructions. Line hot wire connected to line hot terminal on GFCI, line neutral connected to line neutral terminal on GFCI etc.

However, as soon as I turn the power back on, and hit the Reset button, the GFCI trips, and refuses to be reset.

Here’s what I know:

  • I checked that neutral & live are not reversed by using my tester between the metal body of the electrical box and the live wire - current flowed.

  • I tried other GFCIs - same result, so it’s not a faulty unit.

  • I have one of those little yellow plug-in testers (like this). When I reset the GFCI, the tester momentarily flashes its “happy” gang sign - two orange lights, then of course goes dark immediately when the GFCI resets.

  • If I don’t reset the GFCI, the plug-in tester shows a single red light on the left. This case isn’t covered in the legend on the tester, but it seems to be close to “hot/grd reverse” or “hot/neu reverse”, minus the extra yellow light. This is strange because as I mentioned above, I verified that (as far as I can see) they AREN’T reversed.

I’m pretty strong on electrical theory and electronics in general, but I’m somewhat new to home wiring.

Any ideas? What can I do to troubleshoot this?

2

In case it helps, here are some tests I did with my neon tester. I touched one probe to the grounded part of the electrical box ( it’s metal counduit so no ground wire ) and tried touching the other probe to the various wires.
Ground -> Hot “line” : bright glow
Ground -> neutral “line” : none
Ground -> hot “load” : dim glow
Ground -> neutral “load”: bright glow

3

That’s the part that always scares me when I do wiring in my home. I had pretty much the same situation that you have recently. My solution was to ask a friend to check out what I did. Turns out I had everything wired correctly, but I was a little sloppy. When I shoved the outlet into the box, things moved around just enough to create a short.

Any burning odor? Any “soot” on the inside of the box?

4

That ain’t right. This test was done with the GFCI in the tripped condition? If so, they your load lines are energized, which probably means they’re actually on another circuit. Are you installing this circuit from scratch or are you trying to modify an existing circuit. What is connected downstream from the GFCI? If there are outlets, plug a small load, such as a lamp into them and see if it comes on with GFCI tripped. Report back.

5

I had the same thing happen. Did all the tests, but when I wired in the outlets down stream, it tripped every time. I also unplugged everything on the outlets downstream and it still tripped. I disconnected the outlets down stream and it was fine.

I had the line and load perfect. I though I was losing my mind.

My next test is to wire just one outlet down stream…a test outlet, not even one in the house to see what happens.

This has driven me nuts, too.

6

Before you pull out the rest of your hair, try disconnecting the load outlets from the GFCI and measuring the resistance from the neutral to ground (use a voltage tester first to be sure there really is no voltage present!) If it’s anything but near infinite, you’ve probably got a pinched neutral wire in one of the outlet boxes, which would cause the GFCI to trip.

7

No burning smell or soot, no sparks - GFCI just trips and stays tripped.

Yes, the tests were done with the GFCI tripped.

This is part of a kitchen remodel, so I’m modifying existing circuits.

I’m still trying to ascertain what’s downstream from the GFCI; I think it’s a couple of other outlets, but everything is concealed behind drywall with no easy access, so I’m having to go through a lot of guesswork.

One thing I did notice - if I only wire in the “line” wires to the GFCI, it’s happy; it’s when I add the “load” wires that it starts tripping.

Is it possible that the load wires are something other than load? To figure out which was which, I disconnected the wires from the receptacle, turned on the breaker to the circuit, and, using the tester, found that one set of wires carried current, whereas the other didn’t. I deduced that the current-carrying ones must be the “line”, and the non-current-carrying ones the “load”. Is there an alternative explanation? I’m at the point where I just want to call in an electrician and let him rip up all the drywall.

I have to verify this, but it seemed to me that I had to turn off TWO breakers to disable the current in all the wires; is it possible the existing wiring has two different circuits coming into the same box?!

8

Might be a pinched wire in your setup causing the problem, too. You need to identify all the downstream outlets, switches and fixtures and, with the power off, inspect the wiring in each box. I’m not so good at diagnosing electrical problems by description alone, though–if I could SEE the setup, it would be much easier. I could be overlooking something. But it seems that in both your and Philster’s situations, the problem is in the load circuit after the GFCI.

9

YOU MUST HAVE A GROUND WIRE CONNECTED TO THE GFI!
It wont work with

Are you trying to install a GFI in a house wired with two conductors?
Run a ground wire back to the breaker box which should be connected to a driven ground rod! Also connected to power drop messenger, BARE GALVANIZED, cable to power pole and the utility services ground.

10

This is false. A GFCI does not require a ground in order to function correctly. It only senses the imbalance between the current “in” through the hot line and the current “out” through the neutral caused by some of the current being diverted, say, through you and trips if this imbalance exceeds 5 mA. If you install a GFCI on a circuit with no ground, you must apply the “NO EQUIPMENT GROUND” sticker to it and each downstream outlet connected to it.

11

Well, this is distinctly odd. I’d double check that before moving forward.

You say that the GFCI is happy with just line connected, I assume that also means the plug-in tester was happy too, and that all the outlets you expected to be dead were, in fact, dead. I’d probably connect up a plain outlet instead of the GFCI and plug-in test the other outlets, maybe one of them is wired wrong.

Or, start at one end and move the GFCI from outlet to outlet until it works. Lots of work, but you’ll get to isolate the problem.

12

If you can, borrow a tester from somebody that will give you actual AC voltage readings for this.

That last one is absolutely wrong – and unsafe. The ground-nuetral connection should NOT show electrical flow!

What does a Hot-to-nuetral test show? bright or super-bright glow?

I wonder if you have 2 different hot lines here, one of which you think is neutral.

13

Good call on using a meter.

Luckily, I have my handy-dandy Radio shack meter that can measure AC up to 600V.

I think t-bonham may have nailed it - what’s going on here is that there are two circuits in the same box, both live. Here are some measurements I took with the meter. First, my symbols:

G, H, N = ground, hot, neutral
H[sub]li[/sub] - hot for supposed line (etc.)
N[sub]load[/sub] - neutral for supposed load (etc.)

It seems that there are indeed 2 circuits, controlled by two separate breakers - #1 and #5.

Here’s the results with #1 & #5 ON:

G->H[sub]li[/sub]: 115V
G->N[sub]li[/sub]: 0V
G->H[sub]load[/sub]: 60V
G->N[sub]load[/sub]: 114V

H[sub]li[/sub]->N[sub]li[/sub]: 115V
H[sub]li[/sub]->H[sub]load[/sub]: 166V
H[sub]li[/sub]->N[sub]load[/sub]: 236V

N[sub]li[/sub]->H[sub]load[/sub]: 61V
N[sub]li[/sub]->N[sub]load[/sub]: 117V

H[sub]load[/sub]->N[sub]load[/sub]: 40V

With breaker #1 OFF, #5 ON:

Line H-N: 0V
Load H-N: 40V
G->H[sub]load[/sub]:60V

With breakers #1 and #5 both OFF:

All combinations 0V.

One thing I didn’t mention, because I’d forgotten about it until I saw it again: there is a red wire that seems to be “just passing through” the box; it didn’t connect to any terminals or anything else in the box; just a red wire coming in, wire-nutted to a red wire leaving.

This outlet used to be behind an oven; I wonder if it was meant to be for 220V? I didn’t think this would be the case since it’s a gas oven - it was my belief that only electric ovens require 220V. If I understand correctly, 220V means two hot wires coming in, each carrying 110V. I don’t remember seeing the “special” 220V outlet shape there when I removed it, though.

Note: in case anyone’s concerned that I don’t know what I’m doing, you’re probably right. I have applied for a permit and the city inspector will be inspecting my work when it’s done.

14

This does sound like it was once used for a 220 V installation, or junction box for a 220 V circuit running elsewhere in the house. You’re correct, a residential 220 V circuit comprises two 110 V circuits with a neutral between them. The red wire is one hot leg, and the black wire is the other.

Something about those numbers is very wrong, however. Are you quite sure you’re measuring the right lines? Red and black are both hot, neutral is ALWAYS white. You should not measure any voltage between ground and neutral, unless it’s an open end of an energized circuit–which probably means there’s something connected to it. Are there any lamps or appliance on that circuit?

15

Yeah, one of my load oulets (outlets down stream in my house), must have a pinched nuetral wire or something. I ran a test outlet as a load and it was fine. Dang problem is that some load are light fixtures in the ceiling. Dang.

I was adding a recepatacle (in place of a switch) in my garage shop, making it GFCI as required in a garage. The light that the switch would have thrown I had added a motion sensor to, so taking the switch and turning it to an outlet was perfect. I have power for my tools and the light comes on automatically.

In my case, I can ‘pigtail’ the connection, meaning downline the load recepts and fixtures aren’t GFCI. They weren’t to start and they won’t be now.

16

One thing that occurred to me - I could circumvent this whole problem by just wiring in a regular outlet, rather than a GFCI. It’s in the kitchen but it’s nowhere near a water source; NEC only requires GFCIs near water sources, right? It was a regular outlet before, but then the house was built in 1985 to the then existent code.

17

Builders often provide both fixtures for both gas & electric stoves in a kitchen; it’s cheap to do when building and adds to the resale value.

But if it was a 220V oven outlet, the wires would be significantly bigger than the normal #12 or #14 that you will see in house wiring. Were they the bigger?

(It could still be a 220V outlet, but at only 15 or 20Amps. Though I don’t know any common use for that in a standard kitchen.)

18

All the wires are #12 from what I can see, including the “through” red wire.

19

My WAG - What you believe to be the load side neutral is really the hot side of a second circuit. Check the downstream outlets for a hot-neutral reversed. The other possibility is that you’ve got something plugged in on the second circuit (like a clock) and what you are reading on the load neutral is simply the voltage backfeeding through the device. Either way, I’d use the outlet tester to check every single plug and if that doesn’t show a problem, I’d start looking at light fixtures. It’s entirely possible that a previous owner thought he was an electrician.

Another thought - are the wires connected to your outlets under the screw terminals or are they stabbed in to the back of the device? On several occasions I have seen stab-in connections get loose over time and cause some fairly odd behavior. A loose or missing neutral can give some unusual voltage readings.

20

In addition to your above notation, those receptacles downstream must also be marked “GFCI Protected” per 406.3 (D)(3)(c) 2005 NEC. Darn few folks comply with these requirements anyway. :wink: