GFCI weirdness

[atgcpaul]

I was trying to install a photocell on an outdoor lamp post but when I finalized the connection and flipped the light switch, the GFCI outlet would trip. The GFCI is the first outlet in the series. I also had another GFCI downstream of that (which I installed) that then fed power to the light switch going to the photocell to the lamp post. So I noticed that with the 2nd GFCI tripped but the first one on, I didn't have any current on any of the individual wires at the lamp post according to my non-contact voltage tester. However, with my black and ground in close proximity, the tester would indicate voltage.

So I thought maybe the 2nd GFCI (which I installed years ago) was improperly installed. I swapped it out for a normal outlet and tried again. This time, the original GFCI did not trip but the light still didn't come on. There was current flowing to the photocell, but none leaving to my lamp post light--even with my finger covering the photocell. But the weird thing was with the light switch in the off position, my tester was still beeping as before when the black and ground wires were close together at the pole.

I returned the photocell, but I think there's a bigger issue with the GFCI or wiring, but don't know how to proceed. Think the UF cable may be nicked or something? I don't know why this wouldn't trip the GFCI in its "normal" state. The lamp post switches on and of like normal and doesn't trip the GFCI, though.

Thanks,
Paull

[A Squared]

atgcpaul said:

---

Think the UF cable may be nicked or something?

---

I wouldn't think so. If there were a ground fault (cut in the insulation allowing current leaking to the ground) in your wiring, a GFCI should be tripping all the time, even without the photocell installed.

Quote:

---

So I noticed that with the 2nd GFCI tripped but the first one on, I didn't have any current on any of the individual wires at the lamp post according to my non-contact voltage tester.

---

Most non-contact testers detect whether a wire is energized ie: whether voltage is present in the wire. The simple ones do not indicate whether or not current is flowing. A wire with current also has voltage, but a wire with voltage doesn't necessarily have current. it depends if there a complete circuit.

Quote:

---

But the weird thing was with the light switch in the off position, my tester was still beeping as before when the black and ground wires were close together at the pole.

---

Not 100 percent sure I'm picturing this correctly, but if you're using your non-contact tester on the wires downstream from the switch, and it's indicating energized wires, it sounds like you may have the switch on the neutral leg. ie: with the switch off, the hot is energized all the way out to the lamp and back along the neutral to the switch. The switch should be on the black wire, not the white. With a correctly installed in-line switch, your non-contact tester shouldn't be detecting any voltage beyond the switch.* Even if the switch is on the black wire, it's possible that black and white have been switched on hot and neutral at some previous connection in the circuit. I would start by checking that black is hot and white is neutral at all connections along that circuit, starting at the first outlet. I'd use a voltage meter with contact probes instead of a non-contact tester. At every point along your circuit with all switches off and GFCIs reset, you should have 115-120 volts between the black wire and the ground (bare wire) and 0 volts between the white white and the ground. If you find 115 +/- volts between a white wire and ground, then you have hot and neutral switched someplace.

*all that is assuming that the switch to your lamp is an in-line switch, if switch controlling the lamp is on a switch leg, then you could have voltage out to the lamp.

[atgcpaul]

A Squared said:

---

Not 100 percent sure I'm picturing this correctly, but if you're using your non-contact tester on the wires downstream from the switch, and it's indicating energized wires, it sounds like you may have the switch on the neutral leg. ie: with the switch off, the hot is energized all the way out to the lamp and back along the neutral to the switch. The switch should be on the black wire, not the white. With a correctly installed in-line switch, your non-contact tester shouldn't be detecting any voltage beyond the switch.* Even if the switch is on the black wire, it's possible that black and white have been switched on hot and neutral at some previous connection in the circuit. I would start by checking that black is hot and white is neutral at all connections along that circuit, starting at the first outlet. I'd use a voltage meter with contact probes instead of a non-contact tester. At every point along your circuit with all switches off and GFCIs reset, you should have 115-120 volts between the black wire and the ground (bare wire) and 0 volts between the white white and the ground. If you find 115 +/- volts between a white wire and ground, then you have hot and neutral switched someplace.

*all that is assuming that the switch to your lamp is an in-line switch, if switch controlling the lamp is on a switch leg, then you could have voltage out to the lamp.

---

Fortunately all the wires associated with this circuit are in this box in the my basement ceiling. Wire cluster 1 contains 1 black wire fed from my upstream GFCI outlet in my basement, 1 black wire I assume is going to my GFCI outdoor outlets, and 1 white wire which I think goes up to the light switch that controls the lamp post. At my switch, that white wire is not marked as hot so I will do that. I double checked at this box this morning and the black wires at 2 are not energized until the switch is flipped to On.

I am going to replace my switch with a dawn/dusk timer switch. Even if the photocell had worked, it would have been a challenge to get it down into the pole and out the little hole. I'm still positive there was voltage detected out at the lamp post even when there shouldn't have been, but that may have to remain a mystery for the ages.

Thanks,
Paul