I've installed a 240V 50 A GFCI breaker for a to-be hot tub in a subpanel and just done a preliminary test. If I power it on, a non-contact voltage tester shows its load cable is live. However the TEST button on the breaker does nothing. I've pushed it fairly hard, I don't think its just stiff. And most test buttons I've used before are pretty sensitive.

I've not independently tested the two legs vs ground or neutral, but could do so.

I think there are two likely possibilities: the breaker is bad, or my wiring is bad. As far as I can tell the wiring is fine. (I've wired numerous 120 and 240 V breakers before.) Any other possibilities?

I'm looking for advice on diagnosing this. I assume that if the wiring is in fact OK, then the breaker must be defective?

FYI at the moment the breaker is OFF and the cable run to it is terminated at the other end, not attached to any equipment yet. So I think this is safe for the time being.

Actually, assuming I do have to replace the breaker, is this the best state to leave it in the meantime? I could pull the breaker & cap off the wires in the subpanel but that would leave the front cover with a big hole.

Update - I did replace it with an identical breaker - no change to wiring - and the new one worked fine. This seemed conclusive that the breaker was bad.

  • If you turn the breaker off feeding the sub that will be the safest thing to do. Do you have the neutral properly connected to the buss? I would cycle the GFCI breaker off and back on , some require a test then a reset to work and you may be seeing leakage current on the load terminals but if properly functioning and wired the test should cause a trip with no load or a load. What brand GFCI is being used and the panel type there are newer GFCI’s that require a metal rail if you have plastic it may not work.
    – Ed Beal
    Aug 22, 2020 at 18:51
  • Can you post photos of how it's hooked up? Aug 22, 2020 at 21:37
  • Non-contact testers are a good first stage testing device, but they sometimes give false positives. Any time something seems a bit hinkey it's time to put away the non-contact tester and get out a meter or a solenoid tester to test feed and output of the breaker before assuming the breaker is bad. Aug 23, 2020 at 0:21
  • @NoSparksPlease I agree with that about the NCVT and if I were to handle anything that absolutely must not be live I would use a voltmeter or similar instead (and/or multiple different checks). In this case however I was only using it to verify that the cable became live / not live as I switched the breaker on / off. So I think there was very little likelihood of a false reading in this case because the tester was several feet from the breaker itself. Aug 23, 2020 at 17:23

1 Answer 1


Well, I would replace that device .

What you could also do is get hold of an electrician who has the PAT testing unit and get it properly tested. That can tell if it functions at all and if it does, if it is within the specified response time.

  • PAT units are absolutely unheard of over in the USA... Aug 22, 2020 at 21:36
  • @ThreePhaseEel so how do you test a GCFI to make sure it is operating with the specified time range? Or if it takes a day to turn off that is ok??? Not. Really surprised that is not tested on commercial and business properties for everyone’s safety. It is an annual inspection in the UK.
    – Solar Mike
    Aug 22, 2020 at 22:40
  • Generally, North American practice is to use the TEST button on the GFCI as a "go/no-go" test -- modern GFCIs also have some degree of integral self-testing in them as well. Aug 22, 2020 at 23:06
  • @ThreePhaseEel so if it operates slowly ie acts in 65mS instead of the 25mS required? We test it and keep the users safer.
    – Solar Mike
    Aug 23, 2020 at 5:26

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.