We bought a brand new refrigerator and connected the water line and plugged it into the wall in a GFCI outlet. When attempting to push the refrigerator back into the space against the wall, the water line touched a plate on the back of the refrigerator and sent sparks flying. The fuse in the main fuse box flipped and the GFCI has to be reset when the fuse was turned back on. The water line is copper pipe entering through the floor from a crawl space. What could possibly cause this spark and how do we fix it? Thanks in advance!

  • 1
    Clarify the "tripping a fuse" aspect of your post. Are you finding that the GFI pops and has to be reset or that the breaker in the panel is the one that has to be reset? It's unlikely to change the correct answer, but it also an important distinction, generally speaking. As an addendum to this comment, please consider to not touch any exposed metal on this appliance. The copper pipe may be ground and leakage from the appliance is traveling an inappropriate and dangerous path, causing either the panel breaker or GFI to pop. If you can manage to do without, unplug it until it is resolved.
    – fred_dot_u
    Commented Feb 6, 2017 at 19:39
  • It's highly likely the copper water pipe is grounded, so it's also likely that the cause is a short between the metal chassis of the fridge and the copper water pipe. A relatively simple test you can try is 1) Unplug the fridge, 2) Use a multimeter to measure the resistance between the ground (round) on the plug and an exposed metal area of the fridge, 3) Report back the value it reads. It should be a low value. A high value would indicate improper equipment grounding, and the sparks could be a symptom of a line-to-ground fault bypassing the GFCI through the water supply.
    – Hari
    Commented Feb 6, 2017 at 21:05
  • Also, please upload a picture of the area. I'm curious where the fridge touched the water line.
    – Hari
    Commented Feb 6, 2017 at 21:07
  • 3
    Or a line-ground fault may have simply caused an overcurrent trip due to a bolted fault. The fridge's chassis is clearly electrified at 120V. That sounds like a wiring fault to me, it would be real easy to swap hot and neutral on a GFCI with no ground, and tempting to bootleg a ground off the "neutral" oh wait, that's the hot. Commented Feb 7, 2017 at 0:02
  • 2
    Just for the sake of my OCD, check to see that the copper water line really is grounded rather than hot. I have known two cases of electrified water pipes, where equipment was grounded to a water pipe + equipment was faulty + pipe was not conducting all the way to the earth. Commented Feb 7, 2017 at 20:29

1 Answer 1



What you likely have with that poor, maligned GFCI is a situation called a "reverse polarity bootleg ground", where neutral and ground are both connected to the incoming hot and hot is connected to the incoming neutral. This situation is extremely dangerous, even on a GFCI protected circuit, as the miswired ground puts the chassis of whatever's plugged into it at 120V in a way that cannot be disconnected by the GFCI receptacle.

What you can do to test for this is use a non-contact voltage detector -- place its business end next to the receptacle but not in the slots. If the receptacle ground is live -- this means that the entire receptacle mounting yoke is hot at 120VAC and will set off the voltage detector. On a normal receptacle, this won't happen -- the only way to get the detector to go off is to stick the tip in the hot slot on the receptacle.

If that's the case, then you'll need to have the bootlegged "ground" removed (it may be upstream of the receptacle and quite hard to find) and the receptacle polarity corrected as well. If the receptacle comes up clean, then I'd use the non-contact voltage detector to check the copper water pipe -- if that's hot, then there's a bonding problem and an equipment fault, which is a combination you'll want a qualified electrician to fix. (Not the least because trying to fix that DIY puts you at even more risk of the 60 cycle shuffle than you're already at!)

  • Great answer, too bad Katie was a one-time SE user. One question - would a "plug in" outlet tester uncover this problem, or is that unsafe to try and thus why you stressed the non-contact type? Also - just a remark for others (I had to Google this) but by "entire receptacle mounting yoke" the answerer means "the metal wing mounting tabs of the outlet". Commented Feb 24, 2019 at 16:44
  • @Crossfit_and_Beer you need a non-contact tester for this, and the receptacle yoke is indeed the "frame" of the wiring device, yes Commented Feb 24, 2019 at 16:45
  • Thanks. Generally speaking, the reason a plug-in tester is not appropriate is because it may not uncover the problem? Rephrased, using a plug-in tester in this situation is not inherently dangerous, correct? And thanks for all your posts... I've read quite a few. Commented Feb 24, 2019 at 17:21
  • 1
    @Crossfit_and_Beer -- yeah, a plug-in tester cannot uncover a reverse polarity bootleg ground issue Commented Feb 24, 2019 at 17:29

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.