GFCI tripping with no load when connected to power

Users who are viewing this thread

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
I have a 20A circuit which feeds a string of 3 receptacle outlets in the garage, and 3 in the house. The junction of the garage-string and the house-string fortuitously is in one of the garage outlets, so I replaced that receptacle with a Seymour/Legrand GFCI receptacle, hooking the line from the service panel and the house-string to the line side of the GFCI, and the garage-string to the load side. When I restored power, the GFCI tripped.

After some troubleshooting, I find that with the load side of the GFCI (the garage string) disconnected, and nothing plugged into the house-string receptacles, the GFCI trips when power is applied. With the breaker off, I measure infinite resistance between the hot side and ground, so I assume there's a problem between the service panel and the GFCI, but I'm at a loss to see what it might be.

I have tried 2 new GFCIs with the same result.

Any ideas?
 

Reach4

Well-Known Member
Messages
38,796
Reaction score
4,412
Points
113
Location
IL
When you disconnect the load side of the GFCI, are you disconnecting both load terminals? Those are the two terminals covered with yellow tape in this picture:
 
Last edited by a moderator:

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
When you disconnect the load side of the GFCI, are you disconnecting both load terminals?
Interesting you should ask. I had been disconnecting only the hot side. It turns out that disconnecting the neutral side (whether hot is connected or not) eliminates the GFCI tripping. Nothing is plugged in to the downstream receptacles, AFAIK. Maybe I need to know more.
 

Jadnashua

Retired Defense Industry Engineer xxx
Messages
32,770
Reaction score
1,190
Points
113
Location
New England
It doesn't make sense. The GFCI should not care about the line side...it protects things connected to the load side, and if there's nothing connected to those terminals, it should work fine. Now, are you positive you don't have the hot and neutral reversed?

I've had issues with a GFCI tripping when nothing in the load side string was attached...I finally gave up and ran a new cable between one receptacle to the next in the chain and everything has been fine since. I think there may have been a nail or staple in the wire that wasn't significant enough to trip the breaker, but did intermittently trip the GFCI (maybe wind load slightly moving things, or temperature, or who knows what!).
 

Reach4

Well-Known Member
Messages
38,796
Reaction score
4,412
Points
113
Location
IL
Interesting you should ask. I had been disconnecting only the hot side. It turns out that disconnecting the neutral side (whether hot is connected or not) eliminates the GFCI tripping. Nothing is plugged in to the downstream receptacles, AFAIK. Maybe I need to know more.
The GFCI trips when the current on one wire does not match the current on the other wire in the opposite direction.

What you want to do with your multimeter is to check that there is no voltage on the isolated neutral with respect to ground. After you see there is not a voltage, check the resistance to ground. It should be open, just as you saw on the disconnected hot.
 

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
I agree with everything you said, Jim. I'm now going to examine the downstream load circuit, which is only about 8 feet, with an eye toward finding a leak between hot and anything else. There are no staples in that circuit (don't get me started on staples), so any problems should be in a box. I'll know in a few minutes. I hope.
 

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
What you want to do with your multimeter is to check that there is no voltage on the isolated neutral with respect to ground. After you see there is not a voltage, check the resistance to ground. It should be open, just as you saw on the disconnected hot.
Did that first, with the result I'd expect. But it only takes something like 45ma to trip the GFCI, and I'm not sure if I can trust the meter (Fluke 112) to tell me what I need to know. give me a few minutes to look into the downstream boxes.
 

Reach4

Well-Known Member
Messages
38,796
Reaction score
4,412
Points
113
Location
IL
I agree with everything you said, Jim.
?
Did that first, with the result I'd expect. But it only takes something like 45ma to trip the GFCI, and I'm not sure if I can trust the meter (Fluke 112) to tell me what I need to know. give me a few minutes to look into the downstream boxes.
What result did you expect? I would expect zero to a little voltage due to some capacitive coupling, so the resistance reading should then have some meaning. I would then expect the resistance to measure over 10 megOhm.

For another test that does not use your meter, try connecting just the hot wire and leave the neutral not connected. If nothing is plugged into the downstream outlets, I predict the GFCI does not trip.

A GFCI should trip at about a 5 ma mismatch.
 

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
Well, I found it, but not in any detail just yet. But in each of the two downstream boxes, there were goofy cable clamps that look like they were intended for armored cable. This was wired with NM ("Romex"), and the clamp was cutting through the outer sheath and insulation, and into the conductors inside. All I've done so far is release those clamps; I'll figure out a way to remove them shortly.

Thanks for all your help. I still don't understand why this should cause this problem -- remember that the problem occured with the hot wire disconnected at the GFCI, so theoretically (?) I should be able to do anything with the neutral downstream and not affect the GFCI.

Reach, I think I mentioned earlier your prediction was correct. The problem occurred only when the Load neutral was connected.
 

Bluebinky

Member
Messages
588
Reaction score
16
Points
18
Location
Des Moines, WA
If there is any significant load in the circuit (even on the "line side"), a neutral to ground path on the load side can trip a GFCI outlet.
 

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
If there is any significant load in the circuit (even on the "line side"), a neutral to ground path on the load side can trip a GFCI outlet.
Hmmm. I'd better go back to GFCI 102, but I think I see why that might be; thanks. In my case, though, there was no load anywhere on the circuit, Line side or Load side. I think it's got something to do with the fact that the ground circuit is continuous back to the service panel, so if the neutral is shorted to ground, and ground really isn't at zero potential, that might cause a problem. Of course if ground really isn't at zero, there's a bigger problem, probably upstream from the panel.
 
Last edited:

Jadnashua

Retired Defense Industry Engineer xxx
Messages
32,770
Reaction score
1,190
Points
113
Location
New England
It doesn't take much to get 5ma of current, and that's, as was said, all it takes to unbalance a GFCI and cause it to trip. Bottom line, the GFCI expects what comes in on the hot goes out on the neutral. Any differences means that it is going somewhere it shouldn't, and the GFCI is just doing what it is supposed to do. Keep in mind that the GFCI is protecting not only itself, but anything downstream, assuming it has a load side and it is being used.
 

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
This might suggest a quick test of the basic system: Wire up a GFCI receptacle with the Line side going to a standard NEMA 5-15P plug, and the Load side neutral wired to ground, with no connection to the Load side hot. Plug 'er in and see what happens. If it trips, as I expect it will in my case, it must mean something is wrong, but what?
 

Reach4

Well-Known Member
Messages
38,796
Reaction score
4,412
Points
113
Location
IL
This might suggest a quick test of the basic system: Wire up a GFCI receptacle with the Line side going to a standard NEMA 5-15P plug, and the Load side neutral wired to ground, with no connection to the Load side hot. Plug 'er in and see what happens. If it trips, as I expect it will in my case, it must mean something is wrong, but what?
Let's assume for the moment that you are calling the green wire ground. The green and white wires are connected together in the main breaker panel.

It says there is a small voltage differential between the ground and neutral. That can easily happen if the neutral is not on a home run to the breaker panel. If that neutral is shared with another circuit (MWBC), there will be current flowing in the shared wire. If there is current flowing, there will be a voltage drop across that wire.

If we used a water pipe ground, other earth ground, or even wood in a humid environment instead, there could be another voltage differential, even if the neutral wire was a home run to the main breaker. The neutral coming in is connected via wires to other houses. They use current. So even though there is a ground rod connected your incoming neutral, that is not a perfect connection. There can be a small voltage difference. A small voltage difference can cause a small current, and that can be detected by the GFCI.

So I am expecting you can read a small voltage from neutral to ground. That can generate a current.
 
Last edited:

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
I was coming to that line of reasoning. The neutral and grounds are both home runs from the GFCI to the service panel. But a quick experiment proves the basic assertion: If I connect the neutral to ground in any protected receptacle downstream from the GFCI, the GFCI trips. I tried this on several other GFCI-protected receptacles in the home (some protected by a GFCI breaker in the service panel) with the same results. I tried measuring the neutral-to-ground voltage; my meter (Fluke 112) reads randomly from ~7mv to 40mv. I'd love to pursue this, but I'm moving on Thursday...
 

Reach4

Well-Known Member
Messages
38,796
Reaction score
4,412
Points
113
Location
IL
Last edited:

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
I've always been suspicious of the service entrance ground. Improving that was on my list, but now I'm starting a new list on a new home. Seymour-Legrand's tech support answer was "So if the neutral goes to ground you are creating the imbalance."
 

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
GFCI weirdness redux...

I responded to neighbor's call for help with several electrical problems. A fried receptacle was easy -- a backstab failed -- but a general "flaky outlets" complaint was more interesting.

The flaky outlets were a pair of duplex receptacles in a single 4" box in the kitchen. I eventually discovered that they were fed from an upstream GFCI a few feet away, and now and then when anything was plugged into, or unplugged from, any of the "bad" receptacles, the GFCI would trip. The plugged-in/out device didn't have to be turned on. The receptacles looked brand new, but she had purchased new ones anyway, so I took the box apart planning to swap out the old for the new.

The two duplex receptacles were wired funny. The hot from the GFCI went to the hot side of the right-hand receptacle, and the neutral went to the neutral side of the left-hand receptacle. They were then cross-wired to complete the circuits, and the leftover terminals were used to feed another downstream circuit whose path and destination were unknown The receptacles were not backstabbed, all wiring was #12 Cu. The Test and Reset functions on the GFCI worked normally.

Electrically, it should have worked fine. But since it didn't, I rewired it more conventionally, retaining the old receptacles, with the incoming pair wired to one of the receptacles, that 1st one crosswired to the 2nd, and the remaining terminals on the 2nd used to feed the unknown downstream circuit. After rewiring, the original problem went away.

If anyone has any idea why it failed originally, I'd love to hear it. The only real difference is the length of the hot and neutral legs from the GFCI to either receptacle was slightly different in the "failing" configuration, but only by 6" or so, so there would be less than a nanosecond difference in the time for the hot and neutral currents to travel the circuit (if my math is right).
 

Jadnashua

Retired Defense Industry Engineer xxx
Messages
32,770
Reaction score
1,190
Points
113
Location
New England
Probably not the issue, but variations in how tight the connection was torqued down, or the actual length of the loop under the screw, might have an infinitesimal difference in resistance. If the receptacles had overheated previously, that would indicate a poor connection. When I had a similar situation where I got intermittent GFCI tripping, it was the wiring...somewhere in the wall, there was either a pinched wire, or a nail, screw, or staple that wasn't enough to overload the breaker or start a fire, but enough to trip the thing once in awhile...you didn't have to have anything plugged in for it to happen. I ended up replacing the wiring between two receptacles, and it's been working fine now for several years.
 

Mikey

Aspiring Old Fart, EE, computer & networking geek
Messages
3,024
Reaction score
17
Points
38
Location
Hansville, Washington
Interesting to see how sensitive these things are.

I may have some Schluter questions for you soon; replacing 2 tubs with roll-in showers.
 
Top
Hey, wait a minute.

This is awkward, but...

It looks like you're using an ad blocker. We get it, but (1) terrylove.com can't live without ads, and (2) ad blockers can cause issues with videos and comments. If you'd like to support the site, please allow ads.

If any particular ad is your REASON for blocking ads, please let us know. We might be able to do something about it. Thanks.
I've Disabled AdBlock    No Thanks