Any load trips a new GFCI breaker:


  • Brand new construction.

  • New 75' 4wire feed to new subpanel from a long working and reliable main subpanel.

  • One GFCI breaker feeding 4 outlets from new subpanel.

  • ANY load on any outlet trips a new GFCI or new Arc-Fault GFI breaker,including one new and one old vacuum cleaner, and even a simple 40 WATT LIGHT BULB!

None of the other 10 individually GFI fitted outlets anywhere else on the property has tripped for any reason except a water splash on the actual outlet.

  • I have lifted each new outlet one at a time from the GFCI breaker and tested circuits individually.

  • I have removed all outlet wires from the GFCI breaker and directly wired just one new outlet hanging out of the new panel.

  • I measured a little voltage from ground to neutral (typically this is normal on a sub-panel) so I tried bonding the neutral to ground at the new panel.

  • I have used a Megger on all branch and neutral wires looking for shorts, and everything measures good. I even used the Megger on the lamp and vacuums and no problems found.

  • I have read troubleshooting sugestions from at least 30 different sites (all say essentially the same things) and none of the typical wiring or equipment faults that could cause a false trip are present in this case, or simply do not apply.

Nothing stops the breaker from tripping as soon as any load is applied.

I am an electronic engineer and have a lot of experience as an apprentice electrician. I have brought out all my troubleshooting skills and test gear to no avail ... so this one is twisting my brain.

Should I tear into the main panel, megger the new subpanel feed wires, check the main house ground, Sink a local ground wire at the new subpanel?

Any ideas out there?

  • 1
    1) Megger make a lot of different test equipment so it isn't clear to me what you mean by "megger the new subpanel feed wires"? 2) It sounds like you've swapped out the GFCI to rule out a faulty GFCI - is that right?. Mar 7, 2016 at 9:31
  • 1
    A megger is a high voltage ohm meter @ RedGrittyBrick. they are used to verify the insulation is good.
    – Ed Beal
    Mar 7, 2016 at 14:00
  • 1
    @Khirsch so both "GFCI or new Arc-Fault GFI breaker" that you have tried failed? what do you mean by a little voltage neutral to ground. At 75' you have exceeded the distance for #12 wire on a arc fault breaker (14 awg =50', 12awg = 70'). NEC 210.12.A.3.b
    – Ed Beal
    Mar 7, 2016 at 14:10
  • If you had the breaker hooked up while meggering the insulation -- you probably fried the breaker's electronics in the process! Jun 18, 2016 at 21:06

6 Answers 6


Either the breaker is faulty, or you wired it wrong. Check the documentation for the breaker to determine the correct wiring.

Typically you connect the ungrounded (hot), and grounded (neutral) branch circuit conductors to the appropriate breaker terminals. Then you connect the breakers grounded (neutral) pigtail to the neutral bus bar.

  • 2
    I agree @Tester101, Usually I find the neutral from the breaker not connected correctly to the neutral in the panel for this kind of problem.
    – Ed Beal
    Mar 7, 2016 at 14:14
  • 1
    @EdBeal I've seen where folks connect the branch circuit neutral to the pigtail. And where they connect the branch circuit neutral to the neutral bus, instead of the breaker terminal (common when updating to GFCI).
    – Tester101
    Mar 7, 2016 at 14:32

Sounds like a lot of effort for a possibly bad breaker. If you've triple checked the wiring as it seems and everything is clear, then just return the breaker.

Also, the ground in this case has nothing to do with the breaker. The breaker is checking for a difference between hot and neutral.

Oh and disconnect your bond between the ground and neutral bars. They should only be bonded together in the Main panel. Not the 'main subpanel' either, just the first panel that gets power in the house. All subpanel grounds should be isolated from the neutrals. Tieing them together creates a parallel path between the panels for power to run on. See if the breaker works after this.


You most likely have an AFCI (Arc Fault Circuit Interrupter) not a GFCI (Ground Fault Circuit Interrupter) on this circuit. Although it's not impossible that the breaker is bad, I think this is unlikely. It's more likely that there is an arc fault somewhere in the circuit. These can be hard to find, but worth the effort.

I had a similar experience wiring a new receptacle onto a circuit protected with an AFCI, in a relatively new house. After the new receptacle was installed, every time I switched on a steamer plugged into any receptacle on this circuit the AFCI would trip, even with no other loads on the circuit. This nearly drove me crazy until I discovered some paint on the ground conductor under the screw terminal of the wire on the new receptacle. I cleaned the paint off, and the AFCI ceased tripping. That's how sensitive (three cheers for this) AFCIs are to serial and parallel "arc faults", which have been shown to be the cause of the majority of electrical home fires.

So start by removing your loads one by one, and seeing if the nuisance tripping ceases. When it does, you've identified the culprit appliance, which probably has a bad/corroded/oxidized connection in it, or in the cable or plug.

If you remove all the loads, and the tripping still occurs, then check every wire connection to every receptacle and every wire nut on that circuit. Make sure the wires are clean, and the terminal screws/wire nuts are tight.

If you still get nuisance tripping, then you may have a problem in the actual concealed building wiring, like a squirrel or other rodent having chewed through the wire insulation in the attic or inside the walls, and the conductors starting to arc. This is a serious fire hazard, and your AFCI is gonna save your life if this is the case. Call an electrician to diagnose.

Or, if you feel like you're a competent debugger, start disconnecting portions of your circuit, piece by piece, starting from the most distant (from an electrical circuit perspective) outlets. If the nuisance tripping stops after disconnecting a segment, you know the fault is in the just-disconnected segment. Fixing it could involve running a new cable for that segment, finding the arc fault (if rodent-chewed cable in an open attic, for example) and repairing with a new junction box, or pulling a new cable to replace the damaged one.

Even though the AFCI has by now caused you hours of diagnosis and frustration (if you've gone through all the steps above), thank your AFCI for potentially saving your (and your family's) life, not to mention your house.

  • Using a megger with good results says all the wiring is good. Meggers usually have much higher voltages than 120v to ground as this is the max voltage on a split phase system. I have had 4 Meggers only 1 went below 250v I test at 500v & 1000v. At these voltages any damaged insulation causing a ground fault would be quickly identified.
    – Ed Beal
    Sep 20, 2018 at 17:05
  • While a megger might "out" any parallel arc faults in wires barely touching, they won't detect a serial arc fault, such as the one on my ground wire connection in the above example. The sequential disconnect process I articulated above is, I think, still the only fool-proof way to isolate the culprit section of a circuit.
    – AndyW
    Sep 24, 2018 at 1:41

The first rule of GFCIs

You have connected the GFCI's neutral pigtail to the neutral bar correctly.

The only thing that can explain several GFCIs failing in every iteration of your rather exhaustive testing - is also one thing you did not mention.

The first rule is: All hot and neutral current MUST come through the GFCI device. Because, the basic function of a GFCI is to compare current on all hots + neutral, and see that (flow direction included) they sum up to equal/zero. This accounts for all the current, assuring none is running an unintended path.

So in your installation, you have the circuit coming in on a Romex presumably... you took the hot to the GFCI... but you took the neutral to the panel's neutral bar just like every other circuit, ever.

So, the GFCI saw current moving on hot. It did not see equal current moving on neutral. That is an imbalance, so it tripped. It didn't see neutral current because neutral current was bypassing the GFCI.

The same problem can crop up when single-pole GFCIs are used on multi-wire branch circuits. When a circuit has 2 or 3 hots (and optionally a neutral), all hots and neutrals must go through the GFCI*.

A 'trivial' Code requirement that actually helps

Nobody likes hearing "Read the Fine Manual". And granted -- NEC 110.3(B) doesn't require you to read the instructions. But it does require that you follow the instructions, and it's hard to do that without reading them, eh?

Aside from telling you about surprises, the instructions also define the usage for which UL has tested and certified the device as safe. 110.3(B) effectively forbids "off-label" usage contrary to instructions. That application of that device has never been tested.

Note also that UL will never approve instructions which contradict the electrical code, unless they are certifying that as safe. Example: no-neutral smart switches which "bootleg" ground in a protected way.

In my experience, EE's have it much worse

A very common trope around here is people with vast professional experience with electrical science, who are absolutely flummoxed when it comes to AC mains electrical work. It's just junction boxes and wires, it ought to be a walk in the park, right?

There's a reason for that: All the science in mains electrical work has already been nuggetized and wrapped up in a can. You never get near it. It's all ritual and procedure. You never get to use Ohm's Law, but you gotta use a redhead when terminating that MC cable.

That 180 degree pivot from "reason" to "ritual" is especially hard for EEs, who presumed themselves king of the domain, only to find they don't know what a redhead is.

And this also tends to misinform their sense of whether the instructions have anything new to tell them.


Take a look at differential to common mode generation. GFI's can trip on common mode currents. If there is coupling ( magnetic or electric ) on either the line or neutral the common mode current will be seen by the GFI transformer. I think it is the lack of symmetry on the line and/or neutral wiring.

  • Do you mean common-mode to differential conversion? Sep 20, 2018 at 1:02

Just to be sure... On the new sub panel, is the the new circuit in question a Multiwire branch circuits (common neutral). Single pole GFCI breakers will not work on a MWBC. Doesn't sound like your problem, but just wanted to cover this base.

  • Covering a base is usually done by leaving a comment.
    – JACK
    Sep 20, 2021 at 21:49
  • Yeah, I know. However, I don't have enough points to leave comments except to my own questions Sep 22, 2021 at 14:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.