1 GFCI trips when breaker first turned on but then is fine after reset?

[Edrrt]

I need some help troubleshooting this.

I have a GFCI outlet that occasionally trips. The only abnormal behavior I can recreate is if I cycle the breaker on and off, the gfci will trip.

A breaker was added to a swimming pool sub panel to run power to a gate 150 ft away. It goes from the breaker in conduit about 5 feet to this GFCI outlet in a box. From there, it travels 150 feet with 10 gauge wire to a final GFCI outlet. Powers a battery charger and maybe draws 1a tops. That GFCI does not have a problem.

Was working fine for a few weeks until the GFCI closest to the subpanel tripped. I reset it and it did it again a day later.

The only abnormal behavior I can re-create is if I cycle the breaker the outlet connected to the GFCI will immediately trip. If I reset the GFCI it works for a time.

I haven’t been able to track down any faults. Someone mentioned that because of a long run of wire in conduit, induction may be occurring and interfering with GFCI?

It does this whether or not anything is plugged into it.

Thoughts?

 

[Fitter30]

So there are two gfci in the same circuit? 5' then at 150'. 5' trips?

 

[Sponsor]

 

[cdherman]

Couple of points.

-- is it just a GFCI or one of the new AFCI with arc fault and ground fault combined? If its the latter they are notorious for false trips. You might want to just replace the GFCI receptacle first -- they go bad too.

The fact that it will trip with no loads really makes me think its the receptacle itself. BTW -- you can change over to a panel mounted GFCT at the main breaker box and then the whole circuit is protected. And, wired properly, you don't need multiple GFI outlets in a given circuit, assuming that all distal loads are beyond the first GFI protected outlet. In other words, a GFI outlet is designed to protect itself and down stream outlets if wired correctly....

 

[Edrrt]

So there are two gfci in the same circuit? 5' then at 150'. 5' trips?

I took 5’ apart.

There’s an odd daisy chain method.

There’s two insulated ground wires, one coming in from the sub panel and one going out to 150’. Line and load are all being connected via the outlet rather than being spliced together then having pigtail wires to the outlet. However the grounds were not connected directly on the outlet. One was connected to the outlet and the other was connected to the ground screw on the metal box holding the outlet.

I moved the ground wire bonded to the box to the receptacle with the other ground and it seems to have solved the issue.

I guess they figured the yoke and metal box should bonded via the screws holding them together (self grounding) although the box is painted which may be interfering. This also makes me wonder if the metal outlet box needs a jumper to the metal weather proof box that houses it too.

Is it ok to leave it like this? Does there need to be jumpers now connecting the grounds on the receptacle to metal box and box to enclosure as well?

Tempted to just take it all apart and parallel pig tail this in the typical fashion.

 

[Edrrt]

Guess I was wrong. Still tripping on its own. But no longer when throwing the breaker.

“
The circuit is too long for the equipment being protected by the GFCI.

This is common in GFCI circuit breakers with circuit lengths over 150’. Capacitive leakages due to the wire length of the circuitry may exceed the threshold of the GFCI.”

How realistic is this as a cause?

 

[Reach4]

Line and load are all being connected via the outlet rather than being spliced together then having pigtail wires to the outlet.

Tempted to just take it all apart and parallel pig tail this in the typical fashion.

I expect that will fix at least the GFCI outlet tripping.

 

[wwhitney]

I moved the ground wire bonded to the box to the receptacle with the other ground and it seems to have solved the issue.

A metal box must be bonded to the EGC run with the wiring method, and may not rely on a receptacle for its bond.

As to removing the GFCI protection from the wires running to the gate, you need to determine the burial depth of the wiring method used. Per NEC Table 300.5, for general locations like not under concrete, not under a driveway, etc. a 120V 20A or less circuit protected by GFCI only needs 12" of cover (burial depth measured to the top of the wiring method), regardless of wiring method. Whereas without that extra protection, direct bury cable (e.g. UF) would require 24" of cover, and PVC conduit would require 18" of cover.

Chapter 3 Wiring Methods and Materials: California Electrical Code 2022 | UpCodes

UpCodes offers a consolidated resource of construction and building code grouped by jurisdiction

up.codes

So if you have PVC conduit (likely from the mention of an insulated EGC), and it is has 12" of cover but not 18" of cover, or if you have UF cable, and it has has 12" of cover but not 24" of cover, then you may not remove GFCI protection from those conductors.

[I guess you could do it briefly for a test, but if you find the conductor length is the problem, you will have to redo the supply to the gate opener. But if conductor length alone is the problem, you should still get a trip even if you temporarily remove the receptacle at the gate opener and just cap the wires. If conductor length alone is not enough to cause excess leakage current, but the gate opener's own leakage current pushes the total over the trip threshold, then you could use a 120V-120V isolation transformer at the gate opener to supply a GFCI receptacle there, so that neither GFCI sees the combined leakage current.]

Cheers, Wayne

 

[Edrrt]

A metal box must be bonded to the EGC run with the wiring method, and may not rely on a receptacle for its bond.

As to removing the GFCI protection from the wires running to the gate, you need to determine the burial depth of the wiring method used. Per NEC Table 300.5, for general locations like not under concrete, not under a driveway, etc. a 120V 20A or less circuit protected by GFCI only needs 12" of cover (burial depth measured to the top of the wiring method), regardless of wiring method. Whereas without that extra protection, direct bury cable (e.g. UF) would require 24" of cover, and PVC conduit would require 18" of cover.

Chapter 3 Wiring Methods and Materials: California Electrical Code 2022 | UpCodes

UpCodes offers a consolidated resource of construction and building code grouped by jurisdiction

up.codes

So if you have PVC conduit (likely from the mention of an insulated EGC), and it is has 12" of cover but not 18" of cover, or if you have UF cable, and it has has 12" of cover but not 24" of cover, then you may not remove GFCI protection from those conductors.

[I guess you could do it briefly for a test, but if you find the conductor length is the problem, you will have to redo the supply to the gate opener. But if conductor length alone is the problem, you should still get a trip even if you temporarily remove the receptacle at the gate opener and just cap the wires. If conductor length alone is not enough to cause excess leakage current, but the gate opener's own leakage current pushes the total over the trip threshold, then you could use a 120V-120V isolation transformer at the gate opener to supply a GFCI receptacle there, so that neither GFCI sees the combined leakage current.]

Cheers, Wayne

That’s great information. It’s in PVC conduit with a burial depth of 12-18 inches but the train is not filled in yet.

The 5’ GFCI trips randomly even with no load on the line and nothing plugged in at the end. The final GFCI does not trip.

I’m gonna try a different GFCI first. If that doesn’t work, I will try a GFCI breaker instead.

 

[Edrrt]

I temp wired a new GFCI in a pigtail fashion. Problem solved.

Question though. I can’t find anything about a metal receptacle inside a metal enclosure. Do I need to run another ground from the 6 slot Wago to the bonding point on the enclosure as well? Or will the screws through the tabs on the bell box into the enclosure count as sufficient bonding?

Although it’s 10 gauge wire due to the length of the run, since its a 20A breaker I presume it’s OK to use 12 gauge wire for the pigtails?

The Wago’s are only rated for 12g solid, the 10 gauge wire fits but I’ll change them out.

Metal extension will be added to the cover as well. I presume the screws holding it are sufficient bonding there too as it has no ground point.

I’m learning and trying to do it right, though it wont be inspected.

 

[wwhitney]

I temp wired a new GFCI in a pigtail fashion. Problem solved.

If that means that the buried conductors running ~150' are no longer GFCI protected, then you will need to ensure that the PVC conduit has at least 18" of cover to grade (unless it runs under a non-dwelling driveway or parking area, or a road; that region would need 24" of cover). As long as you do that, this is a good solution to your issue, namely minimizing the length of conductors protected by each GFCI.

Question though. I can’t find anything about a metal receptacle inside a metal enclosure. Do I need to run another ground from the 6 slot Wago to the bonding point on the enclosure as well? Or will the screws through the tabs on the bell box into the enclosure count as sufficient bonding?

As per the first sentence of my previous post, bonding the box must not rely on the receptacle. So if you are referring to the receptacle screws through the tabs on the bell box, no, that is not sufficient bonding. Run a bonding jumper to the metal box.

Of the two bonding jumpers, one to the box and one to the receptacle, what a self-grounding receptacle lets you do is omit the jumper to the receptacle, not the jumper to the box.

Although it’s 10 gauge wire due to the length of the run, since its a 20A breaker I presume it’s OK to use 12 gauge wire for the pigtails?

The Wago’s are only rated for 12g solid, the 10 gauge wire fits but I’ll change them out.

Yes, #12 is fine for pigtails on a 20A circuit. Wago 221-4XX are only rated for #12; Wago 221-6XX are rated for #10.

Metal extension will be added to the cover as well. I presume the screws holding it are sufficient bonding there too as it has no ground point.

I believe the metal screws holding a metal box extension to a metal box are sufficient for bonding the metal box extension; but I haven't really thought about it. If the instructions for the metal box extension are silent on bonding, then nothing extra is required.

Cheers, Wayne

 

[Edrrt]

If that means that the buried conductors running ~150' are no longer GFCI protected, then you will need to ensure that the PVC conduit has at least 18" of cover to grade (unless it runs under a non-dwelling driveway or parking area, or a road; that region would need 24" of cover). As long as you do that, this is a good solution to your issue, namely minimizing the length of conductors protected by each GFCI.


As per the first sentence of my previous post, bonding the box must not rely on the receptacle. So if you are referring to the receptacle screws through the tabs on the bell box, no, that is not sufficient bonding. Run a bonding jumper to the metal box.

Of the two bonding jumpers, one to the box and one to the receptacle, what a self-grounding receptacle lets you do is omit the jumper to the receptacle, not the jumper to the box.


Yes, #12 is fine for pigtails on a 20A circuit. Wago 221-4XX are only rated for #12; Wago 221-6XX are rated for #10.


I believe the metal screws holding a metal box extension to a metal box are sufficient for bonding the metal box extension; but I haven't really thought about it. If the instructions for the metal box extension are silent on bonding, then nothing extra is required.

Cheers, Wayne

Awesome! Thank you so much for the help. It is still GCFI protected. Everything you said is clear except bonding the metal receptacle box to the metal enclosure it’s mounted onside of?

The outlet and the receptacle box are both grounded. But what about the NEC metal enclosure that houses it?

Do I need to run a bonding wire or ground to the metal enclosures bonding stud too? (What my pinky is on in the photo) Or if the metal bell box is screwed into it would that be considered sufficient?

 

[wwhitney]

Do I need to run a bonding wire or ground to the metal enclosures bonding stud too? (What my pinky is on in the photo) Or if the metal bell box is screwed into it would that be considered sufficient?

Ah, right, I didn't realize that your receptacle box is mounted inside a larger metal enclosure.

What else is in that larger metal enclosure? Certainly there's no downside to bonding it, and depending on what's in it, or its location (sufficiently close to a pool, say), it will be required to bond it. If that bonding stud is for a jumper to the door, to ensure that the door is bonded to the rest of the enclosure when the door is open, that certainly suggests that the manufacturer expects the enclosure to be bonded for the way the enclosure is usually used.

BTW, the use of solid yellow for grounding/bonding conductors is non-standard and like prohibited (I'd have to check the fine print on the NEC rule; it's certainly prohibited for an EGC.) IIRC green with a yellow stripe is fine.

Cheers, Wayne