GFCI fault indicated, no power- no loads.

Status
Not open for further replies.
P

purduephotog

Member
Location
Rochester, NY, USA
While I'd like an answer to the specific problem I encountered I'm looking more for theory as to what is happening than specifics.

Background:
Continuing on the helping the neighbor and hooking up their disposal I installed a GFCI after a switch. I removed all of the previous electrical romex (which was DUCT TAPED), installed 3/8 shielded to the outlet, grounded the outlet. Installed a switch, grounded said switch and box, and ran the armored cable back to the previously utilized junction.

The panel breaker was installed and labeled previously as only a disposal. I haven't traced the connection to verify it. I did have to splice and nut the armored to romex inside of the previously used junction box, and I did piggy the ground to the metal box as well.

Problem:
Immediately upon turning on the switch the GFCI flashes red once, and turns off. Resetting repeats the flash once, then off. The GFCI has no power, indicating a fault state. This is at the end of the circuit, with nothing attached to the outlet, and yet... the theory to me is that there is an imbalance between line and neutral- except there is nothing plugged into the outlet!

I then grabbed another GFCI (She had an old one), and it too indicated a fault state and wouldn't stay energized. I did not go back with a third GFCI (yet).

I plugged in a circuit tester to ensure that line/neutral weren't reversed somewhere else, and all three were good- ground, line, neutral.

What am I missing here? If there is something else on this line, how is it faulting the GFCI? The junction box I used to convert the armored back to romex (previously installed romex) has a ground onto the box itself- could something be powering the earth there, that's following the metallic sheath to the GFCI? Is it possible to have a DC Offset voltage mess things up?

I've already told her her panel looks atrocious- and recommended a master electrician I've worked with in the past. The panel is full of splits including a phase-jumped one to power the stove (taking up a single slot instead of 2...). I haven't pulled open the panel as I'm quite afraid to find there. But I am really not understanding what's happening with the GFCI and how it can go off at the end of a circuit if nothing is in it.

I appreciate your time for answering this probably very basic question.
 
If you haven't already done so, measure the voltages between hot, neutral, and equipment ground with a meter. Circuit testers provide limited and sometimes misleading information.
 
A GFCI device will trip when the magnetic field generated as current passes through its phase conductor on the way to the load is not exactly equal to the magnetic field generated as current passes through its neutral conductor on the way back from the load. It is clear that that is not what is happening here. My only guess has already been touched upon by other members. It is possible (though I don't know how likely) that somewhere upstream the wiring is incorrect. Perhaps the phase conductor you attached to the outlet is not a phase conductor all the way back to the panel. The same might be conjectured regarding the neutral conductor and the ground wire. So the device might be tripping simply because it is recognizing that it was not wired correctly.

So let me pass a question to the forum's electricians: Suppose a house's wiring system was installed perfectly. Suppose you were to add a new GFCI receptacle somewhere. Suppose you wired it correctly, and it worked properly. Suppose then, just for the knowledge gained from the experiment, you temporarily removed the receptacle and reversed the black and white wires. Finally, suppose you energize the circuit, but do not plug anything into the outlet.

Question: Would the GFCI device do what the OP described in Post 1 (i.e., flash red and turn off)?
Similar Question: Would the device trip if you had reversed any two other wires?
 
charlie b said:
A GFCI device will trip when the magnetic field generated as current passes through its phase conductor on the way to the load is not exactly equal to the magnetic field generated as current passes through its neutral conductor on the way back from the load. It is clear that that is not what is happening here. My only guess has already been touched upon by other members. It is possible (though I don't know how likely) that somewhere upstream the wiring is incorrect. Perhaps the phase conductor you attached to the outlet is not a phase conductor all the way back to the panel. The same might be conjectured regarding the neutral conductor and the ground wire. So the device might be tripping simply because it is recognizing that it was not wired correctly.

So let me pass a question to the forum's electricians: Suppose a house's wiring system was installed perfectly. Suppose you were to add a new GFCI receptacle somewhere. Suppose you wired it correctly, and it worked properly. Suppose then, just for the knowledge gained from the experiment, you temporarily removed the receptacle and reversed the black and white wires. Finally, suppose you energize the circuit, but do not plug anything into the outlet.

Question: Would the GFCI device do what the OP described in Post 1 (i.e., flash red and turn off)?
Similar Question: Would the device trip if you had reversed any two other wires?
Click to expand...
Most will trip on reverse polarity as well as line connected to load.

The will also trip with a load neutral touching a ground wire.

Sent from my LML211BL using Tapatalk
 
I'd suspect there is something on the neutral downstream from the GFCI. I've had CFL and fluorescent lights on the circuit before the GFCI and they will trip the GFCI.
 
Thank you for all the answers. I'm glad to see that my understanding isn't so fundamentally flawed that you all are smacking me upside the head :)

I'll go back and re-check the circuit and lines. And while I won't say I couldn't have accidentally touched ground/neutral, I did pull the bloody thing apart 6x to figure out what was going wrong. I can't make that mistake THAT many times.

I did NOT measure the voltages across neutral/hot. I will do that.

I'll also go back to the junction box and disconnect the ground from it, in case something else is mis-wired and using the ground to return what the neutral should have.

However... there's nothing on that circuit, supposedly. I guess it's possible too I grabbed a bad GFCI twice. I've seen stupider things happen with cables in the computer before...
 
I just tested a Pass & Seymour 20A GFCI 2095-TRWR meeting the 2015 UL spec, and it reset and held fine with the line side hot and neutral reversed. It worked the same whether I had a ground return connected to the green screw or not. I think a ground return would be required if the receptacle needed to distinguish the grounded conductor from the ungrounded conductor, because there's obviously 120V AC across the receptacle terminals whether they're reversed or not.
I did not expect any continuity from the receptacle's hot or neutral line side terminals to ground, but I confirmed they were an open circuit to ground with a meter just to be sure.
 
synchro said:
I think a ground return would be required if the receptacle needed to distinguish the grounded conductor from the ungrounded conductor, because there's obviously 120V AC across the receptacle terminals whether they're reversed or not.
Click to expand...
A GFCI device does not need access to an equipment grounding conductor, and voltage is not relevant to its operation. Consider a bench experiment in which you run a white wire and a black wire from a plug that is connected to an outlet. Connect the load ends of the two wires to a light bulb. Then take a third wire and wrap it several times around the black and white wires, connect this coil to a meter, and turn on the light. The current flowing to the light along the black wire will create around itself a magnetic field, and your coil would not be able to detect it. That is because the current flowing back from the light along the white wire will create around itself a magnetic field that is exactly opposite, at every instant of time, from the field created by the black wire. The two magnetic fields cancel each other out.

That is what is happening inside a GFCI device. If the magnetic field sensed by the coil that, internal to the device, is wrapped around both the black and whiter wires, is not exactly zero at any point in time, it means that the two currents are not equal. That can only happen if "not all" the current leaving on the black wire is returning on the white wire. That means that current is returning to the source via some other path. Given the possibility that that other path could include a human, the GFCI is set to trip at a value of leakage current that is too small for the shock to be fatal.

Note that nowhere in this discussion was there mention of current flowing in the green wire or mention of voltage.
 
charlie b said:
A GFCI device does not need access to an equipment grounding conductor, and voltage is not relevant to its operation. Consider a bench experiment in which you run a white wire and a black wire from a plug that is connected to an outlet. Connect the load ends of the two wires to a light bulb. Then take a third wire and wrap it several times around the black and white wires, connect this coil to a meter, and turn on the light. The current flowing to the light along the black wire will create around itself a magnetic field, and your coil would not be able to detect it. That is because the current flowing back from the light along the white wire will create around itself a magnetic field that is exactly opposite, at every instant of time, from the field created by the black wire. The two magnetic fields cancel each other out.

That is what is happening inside a GFCI device. If the magnetic field sensed by the coil that, internal to the device, is wrapped around both the black and whiter wires, is not exactly zero at any point in time, it means that the two currents are not equal. That can only happen if "not all" the current leaving on the black wire is returning on the white wire. That means that current is returning to the source via some other path. Given the possibility that that other path could include a human, the GFCI is set to trip at a value of leakage current that is too small for the shock to be fatal.

Note that nowhere in this discussion was there mention of current flowing in the green wire or mention of voltage.
Click to expand...
synchro said:
I think a ground return would be required if the receptacle needed to distinguish the grounded conductor from the ungrounded conductor, because there's obviously 120V AC across the receptacle terminals whether they're reversed or not.
Click to expand...

Charlie, I agree 100% with what you have described that justifies a ground connection is not necessary for GFCI operation. What I was trying to convey in the context of the OP's specific problem, but which I obviously failed to articulate properly, was the following:
If the GFCI had to provide an additional detection mode to disable it when the hot/neutral line input terminals were miswired with "reversed polarity" as masterinbama mentioned in post #7 (and along the lines of the additional requirements in 2015 for disabling the GFCI with line/load miswiring, etc.) then an EGC connection would likely be needed. Otherwise I don't see how the GFCI circuitry could distinguish a grounded conductor from an ungrounded one.

I haven't seen any mention of a requirement to disable GFCIs if the line inputs are reversed in any references to the 2015 UL requirements or other sources, but I haven't looked into this in any detail.
 
suemarkp said:
Small metal device box and GFCI side screws not screwed in (load side)?
Click to expand...
Nope, but I might admit that has happened in the past :)

I've poked and prodded this bloody thing and can't come up with any reason for it to trip off. I'm halfway tempted to run an extension cord all the way back to the panel to see if it's a house issue. But I'm really concerned that their panel is totally messed -I mean... this was their disposal.

1580427384420.png
 
RoyABC123 said:
place the GFIR ahead of the switch and see what happens
Click to expand...
That, or at least measure the voltage between the hot and neutral line input terminals of the GFCI to make sure there's 120V present with the switch ON and you are trying to reset the GFCI.
If it appears the voltage is there but it still won't reset, I suggest you remove the GFCI and temporarily power it through a line cord (after taping up the terminals for safety) or install it at a known working receptacle location, so that you can verify it works properly. That's to double check in case there might be some very short voltage dropout on the switched circuit that's preventing the GFCI from resetting, maybe from a bad switch or whatever.
 
Last edited:
Status
Not open for further replies.
Top