Finding a ground fault leak

Thread Starter

MrSoftware

Joined Oct 29, 2013
2,027
Looking for ideas on how to follow a wire behind a wall. We just had about 80 foot of exterior wall replaced, that horrible T1-11 stuff replaced with Hardie board. Coincidentally, the GFCI breaker in my power panel, on the opposite side of that wall, will not reset, it trips instantly. I replaced the breaker, problem still exists. The power was off while they were working on the wall, so I have no idea what section of wall they were working on when the problem popped up. The hot and neutral are not shorted at all, infinite resistance according to my Fluke 87V. With only the neutral connected to the GFCI breaker, it still trips. So this leads me to believe the guys replacing the wall probably put a screw through the neutral somewhere behind the wall, and it's leaking enough current into the wall (S. FL everything is humid) to trip the GFCI. My question is, is there a tool I can connect to the neutral in the breaker panel, then go to the exterior wall and use it like sonar to follow that particular wire along the wall, potentially through conduit?
 

MisterBill2

Joined Jan 23, 2018
11,879
I have a tracer that seems to work quite well, but it was not cheap at all, about $240. It is a "Triplet" "Fox and Hound" tracer that has a number of possible methods of tracing. But first, have you disconnected everything from the other end of that cable segment? And a short between conductors will not trip a gfci, but it will destroy it. So the connection is not between conductors on the cable, but between one conductor and something else. And so the very first step is to disconnect the far end and the near ends of that cable, unless you have already done that. Then check the resistance to "ground "of each conductor. The leaky conductor will have a lower resistance to ground. Once you have found which conductor is the culprit, and it is not a direct short to ground, the tracking can begin. You did not mention if this cable includes a third "ground" wire, like most newer cables do.You may have a connection between the white wire and the green wire, if a screw has penetrated the cable.
One more question, is this a cable going to a single outlet or appliance, or just a regular home wiring circuit. Is it a 120 volt circuit, or a 220/208/240 circuit?
If the fault is found to be a shorted circuit between the neutral conductor and the safety ground conductor on a 15 amp circuit, another choice is to move the ground fault protection to the far end, which does not introduce a shock hazard because nobody connects to the cable in the wall. You would need to change to a normal circuit breaker in the panel, though.
But no matter what sort of tracer device you use, you must be able to disconnect all of the conductors in the cable at each end.

I am not familiar with the material "Hardie Board", but I am guessing that it is attached with screws into the studs. And probably one screw was longer than it needed to be. One test process will be to energize the cable bypassing the GFI and powering all of the conductors. You would then follow along the presumed path with a voltmeter having one lead connected to a good ground with a long wire, and probe each screw head with a pointed probe. Hopefully the screw heads can be accessed. You may find more that one screw has damaged the cable. This method is cheap and simple and should locate the damaged point so that a repair can be made. Hopefully the screws are at a consistent height and easily accessed for probing.
 

MisterBill2

Joined Jan 23, 2018
11,879
With the power off, you could check resistance between screw/nail heads (if they are exposed) and the neutral.
If the fault is a connection between the white wire,(Neutral), and the ground wire, it will be difficult to detect the difference in resistance unless all of the connections at both ends are disconnected. This is because neutral and ground are often tied at unexpected locations. Not intentionally, except in the main breaker panel.
 

AlbertHall

Joined Jun 4, 2014
12,020
If the fault is a connection between the white wire,(Neutral), and the ground wire, it will be difficult to detect the difference in resistance unless all of the connections at both ends are disconnected. This is because neutral and ground are often tied at unexpected locations. Not intentionally, except in the main breaker panel.
But none of the nails/screws should be connected to any wire.
 

Thread Starter

MrSoftware

Joined Oct 29, 2013
2,027
Thanks guys for the replies. The wire leaves the breaker box into the attic where it branches off to power 3 bathrooms plus a handful of outdoors outlets. This is common for houses built in the 80's here, GFCI breakers must have been a million dollars back then because they put everything that requires GFCI on the one single GFCI breaker in the panel. Anyway, by disconnecting branching circuits I think I've narrowed it down to a single 50-ish foot run of wire. I cut the wall out where the wires leave the breaker panel as that's where I thought one may have been damaged. I was correct that at least 3 screws made it into that space, but none hit any wires so the actual problem is elsewhere. So after cutting the wall out I'm able to follow where the wire goes and I'm going to brute force following it through the attic to see where the issue is. Perhaps the wall guys didn't damage it and it's just coincidence that the problem popped up at the same time. For now I've disconnected the problem section and I'll get back into it next weekend.
 

MisterBill2

Joined Jan 23, 2018
11,879
One alternative is to alter the scheme and install a GFCI at the end of each run. It is never a good choice to protect several different runs with just one GFCI.
Since I have been installing them the price has dropped but then started increasing again, the last purchase was $54 for a box of four 15 amp devices. Or $16.95 each As the job required four of them my choice was easy.
 

crutschow

Joined Mar 14, 2008
29,785
If the GFCI trips then there is current between the hot and ground, not between hot and neutral (which is normal current).
 

michael8

Joined Jan 11, 2015
287
Some GFCIs will trip on a short between neutral and ground (on the load side of the GFCI).

The previous to this ICs wheren't as "smart" (no end-of-life and self test) but
did detect neutral ground faults...

https://www.ti.com/product/AFE3010
https://www.ti.com/lit/ds/symlink/afe3010.pdf?ts=1650845792760

AFE3010 Ground Fault Circuit Interrupter (GFCI) With Self-Test and Neutral-Ground Fault Detection

In addition to a ground fault leakage detection, the AFE3010 can detect a grounded neutral condition
 

MisterBill2

Joined Jan 23, 2018
11,879
Thr GFCI trips when there is a difference in current between the two conductors through the sense coil . It is only that difference that trips it. It does not matter which one or where the current flows to.
 

crutschow

Joined Mar 14, 2008
29,785
Thr GFCI trips when there is a difference in current between the two conductors through the sense coil . It is only that difference that trips it. It does not matter which one or where the current flows to.
True, that's how the GFCI works internally.
But how else would it occur in typical mains wiring other than from hot to earth ground?
 

MaxHeadRoom

Joined Jul 18, 2013
25,206
One alternative is to alter the scheme and install a GFCI at the end of each run. It is never a good choice to protect several different runs with just one GFCI.
That is how it is, or at least was, wired in the UK, with one RCD (GFI) mounted in the service panel for all outlets.
 

MisterBill2

Joined Jan 23, 2018
11,879
That is how it is, or at least was, wired in the UK, with one RCD (GFI) mounted in the service panel for all outlets.
I am aware that it was done that way at one time, and also aware that it is a VERY POOR CHOICE because then one fault takes out the whole system. I came across that in one large house where I was adding some wiring, and tripping one GFCI put the whole basement in darkness. SOME IDEAS ARE JUST PLAIN STUPID, and that is one of them. The cost saving in a house selling for $120,000, to cheap on the number of GFCI devices at $12 each in a quantity purchase is simply not justified.
 

michael8

Joined Jan 11, 2015
287
# The GFCI trips when there is a difference in current between the two conductors through the sense coil . It is only that difference # that trips it. It does not matter which one or where the current flows to.

True, that's how the GFCI works internally
.

False, or not completely true. Look at the diagram at the bottom of page 1:
https://www.ti.com/lit/ds/symlink/afe3010.pdf?ts=1650845792760

There are two transformers around the hot/neutral wires. One is driven by the GFCI circuit the other
is sensed. The two transformers are only coupled if one or both the hot or neutral are connected via
an outside connection between the source and load sides of the GFCI. At the source the normal
ground neutral bonding is a connection between ground and the neutral so a neutral ground fault
on the load side of the GFCI will couple the GFCI transformers triggering it. Thus a GFCI can detect
neutral to ground faults in addition to hot to ground faults.
 

Thread Starter

MrSoftware

Joined Oct 29, 2013
2,027
Regarding where the current leak is happening: I disconnected all of the hots but left all the neutrals connected and it still trips. Whatever the problem is, it's right on the edge. With everything connected properly, sometimes the breaker will stay "on" for several minutes before tripping, then it will get in a mood where it trips instantly when I turn it on, over and over. The new breaker behaves the same as the old one so it's not a bad breaker. Regarding just swapping it with a regular breaker and putting the GFCI per-outlet: technically this might solve the problem of the breaker tripping, but would still leave me uneasy about having an actual short somewhere. I'm not sure what it's shorted to, maybe a neutral on a different circuit, but a short nonetheless.
 

crutschow

Joined Mar 14, 2008
29,785
Regarding where the current leak is happening: I disconnected all of the hots but left all the neutrals connected and it still trips
Have you opened the breaker box circuit neutral and checked the resistance between the neutral and ground wires?
 
Last edited:

MisterBill2

Joined Jan 23, 2018
11,879
If you go farther into that data sheet then you learn that indeed, that second winding is for doing a a neutral to ground check, which makes it unclear as to what benefit that check delivers. If a portion of the load current is passing into the ground circuit instead of the neutral return, the result is that the trip threshold is reduced a bit. Or does the sensor controller somehow compensate for that??
It would provide a useful bit of information to temporarily disconnect the ground wire from that cable run, at least in the breaker panel. If the tripping stopped completely then you could think that it was a neutral to ground leakage problem. In that case, installing a GFCI device at each outlet and bypassing the one in the breaker panel would be the way to go. And a bit safer as well.
 

Thread Starter

MrSoftware

Joined Oct 29, 2013
2,027
Have you opened the breaker box circuit neutral and checked the resistance between the neutral and ground wires?
I could try that. Remove the neutral for that branch and compare between that neutral and the breaker box ground, which would be the same as the other neutrals.
 
Top