I have a GFCI located in my garage which feeds into my bathroom and then goes to an external junction box and finally underground to my deck. Upon doing some routine testing, I noticed by deck receptacle showed "Open ground". Upon initial investigation, I noticed that the wiring from the bathroom outlet to the external junction box was missing a ground wire. I assumed that was the problem so I replaced that span with brand new wires, inclusive of a ground.

I reconnected everything the same. The only difference is that the new ground from the bathroom now connects to the ground wire coming from the deck, which meets in the external junction box. I assumed all was good, but no chance. The GFCI kept tripping.

Long story short, I can confirm that the short happens when I connect the bathroom ground and the deck ground in the external junction box. If I do not connect them, then everything works, but I still get the "open ground" at my deck.

Why is this happening? I thought the ground wire can't trip a GFCI? Is water, likely in the underground cable/conduit, causing the ground to short things out?

Not to complicate things, but the end of the underground cable at the external junction box is 12 AWG; however, the opposite end at the deck is 14 AWG. I do not see another external junction box. Is it possible they connected two different cables underground and that's what's causing this?

Appreciate any feedback. I am a novice at electricity but I love learning.


  • Is there a load on the GFCI when it trips, or does the GFCI trip with no load whatsoever present? Oct 11 '20 at 4:54
  • 1
    the latter - no load whatsoever.
    – dabronx02
    Oct 11 '20 at 5:00
  • What sort of wiring method was used for the run to the deck then? Oct 11 '20 at 5:14
  • I'm guessing you have ground and neutral swapped somewhere after the GFCI that trips
    – Jasen
    Oct 11 '20 at 5:31
  • ThreePhaseEel - at the deck end, it was comes out of a conduit and is labeled as bury rated 14 AWG cable. The other other coming thru to the junction box is 12 AWG
    – dabronx02
    Oct 11 '20 at 5:44

Let's consider Joe. Joe is living on a boat, using a diesel fired heater and using the engine to keep the batteries charged. Someone suggests fitting a carbon monoxide detector. Joe fits one in the galley, and the detector goes off. So Joe moves it to the bedroom, still goes off. Back to the store it goes, the replacement also goes off. Back it goes, Joe gets a more expensive one. It also goes off. Joe is going crazy - can't anyone make a reliable detector???

Laugh if you will, but this is how most people react to GFCI trips.

The evidence is leading you to a conclusion, so let's try it on for size: "attaching the ground wire makes the circuit less safe". Does that make any sense? Of course not. The ground should always be connected, always.

What's happening is there's a second problem - which was always there all along. And hooking up the ground wire improved the GFCI's ability to detect it.

Power is supposed to flow out the hot wire, through the appliance and back the neutral. Any current moving on a third path is a ground fault. The reason we started installing safety grounds in the 60s was to provide a reliable third path, instead of allowing Fate to choose its own, typically through a dying human. So it makes perfect sense that installing/repairing a ground wire would increase a GFCI's sensitivity to ground faults.

Disconnecting the ground, thus, became a "red herring". It showed correlation, but not causality.

Now, you need to admit that you probably have a ground fault, and then just work standard failure-tree analysis until you find it. As far as the grounds, there might be diagnostic value in temporarily unhooking one, but don't see it as a path to a solution. You must finish with all grounds connected.

ThreePhaseEel was quizzing you about cable type because the usual reason for a ground fault on an outdoor run is use of NM-B cable, which is not rated for outdoor/wet usage, and the insulation will fail when wet.

  • I agree the GFCI tripping was a good thing. I never questioned its reliability. My conclusion thus far is that the cable is likely shorting at some point where the THTN cables transitions to the UF-B. Unfortunately, there is no other JB visible so I am concluding the bad connection must be happening underground. Unfortunately, I can't think of an easy way to find it. Any idea if residential electricians are equipped to find something like this, that appears to take place underground? I've also read that I should do a continuity test, which is next on my list.
    – dabronx02
    Oct 12 '20 at 6:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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