1

I have two pool lights connected to a 15a Square D GFCI breaker. The breaker started tripping recently. After a bit of troubleshooting I've found that each light works individually but trips the breaker when both are connected.

  • I removed both light enclosures from the water and verified the metal case is grounded.
  • The inside appears to be dry, no sloshing water.
  • checked resistance between each wire running to the fixture, resistance is zero
  • Enclosures have LED lights installed about 2 years ago, clamp meter shows each draws about 36W
  • The panel with breakers is outside, but dry and looks good inside.

Ideally, I'd like to open the fixtures and remove the bulbs but that would require me to buy new gaskets. I was trying to find a way to see which component may be bad without throwing money at it.

Some History:

The lights had originally been setup with a standard 20a breaker, GFCI receptacle/switch and 500w lights. A few years ago, I wanted to add a smart switch for the lights so I replaced the breaker with the 15a GFCI, swapped the outlet/switch for the smart switch and added LED lights.

The breaker isn't that old, I don't think that's the issue, but I could return to the 20a and add a GFCI receptacle I have on hand and test that.

Any ideas or should I just order the gaskets and a new bulb?

Edit:

I removed the GFCI breaker and added a standard 20a, tested with both lights and they both came on. In this doomsday configuration, the clamp measured no current on ground or with the hot and neutral clamped together. As suggested, the clamp is not sensitive enough.

I also added a GFCI receptacle in between and it also trips immediately.

I've restored the original GFCI and single light and will probably order a pair of gaskets so I can open the light housings.

Improve this question
4
  • 1
    Can you disconnect neutral from the breaker and then measure neutral wire (going to light but disconnected from panel) to ground (in the panel) resistance? It should be infinite, indicating no connection. If instead it shows high resistance then (based on Ohm's law) some current when running will go to ground, presumably enough on each that the two together trip the GFCI. That's my theory. This may be what others have referred to as the use for a "megger" tester, which I don't have (and have never used), so it is possible an ordinary tester won't spot it. Hopefully one of the experts will
    – manassehkatz-Moving 2 Codidact
    Jun 29 at 15:39
  • either confirm my theory (and indicate necessary tester) or deny it (in which case I'll delete these comments).
    – manassehkatz-Moving 2 Codidact
    Jun 29 at 15:40
  • 2
    "checked resistance between each wire running to the fixture, resistance is zero" It's the leakage current from hot to ground that needs to be measured, and should be zero, too low to check with a clamp-on meter. Likely, condensation inside both lamps has increased leakage current, since either alone does not trip the GFCI.
    – DrMoishe Pippik
    Jun 29 at 17:00
  • 1
    I removed the neutral from the breaker and tested the resistance to ground, it was infinite. I've added a few more details to my question.
    – Gary Bak
    Jun 30 at 1:57

1 Answer 1

Reset to default
2

You can measure the leakage current of the lights by putting the clamp meter around live and neutral (but not ground) then check the current while the lights are operating.

If one light's leakage current is significantly worse than the other then replace it.

Improve this answer
2
  • 3
    GFCI trip level is nominally 6ma. Each light is OK (so less than 6ma) but together they trip (so more than 6ma). Which means if they are the same then each is leaking between 3ma and 5ma. Can your clamp meter either: read/display that little of a difference between the hot reading and neutral reading? or detect/display that little if you put it on ground? I suspect "no" to both.
    – manassehkatz-Moving 2 Codidact
    Jun 29 at 16:02
  • Thanks, I tried this and it measured 0, I don't think my clamp is sensitive enough.
    – Gary Bak
    Jun 30 at 1:58

Your Answer

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

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