Self-healing reversed polarity.

[Richard Moore]

This "amused" me for a while today. I tripped an exterior receptacle with a GFCI tester and heard the click coming from the nearby interior. OK so far. Later, when I'm in that room, I find the GFCI contaminated with paint and had trouble resetting it. With a basic 3-light tester in place, I finally manage to get it to turn on, but find it indicating reverse polarity. I take a picture (for my notes).

Click to Enlarge

16.07?KB

I then turn to my client who happens to be in the room and turn back to point out the offending GFCI receptacle, that mostly needs replacing because of the paint. And, as if it had had a quick trip to Lourdes while I had my back turned, it healed itself!

Click to Enlarge

13.43?KB

Subsequent trips and resets resulted in more momentary flashes of reverse polarity (less than a second) and then, voila, fine again. I suspect this has something to do with the paint slowing down the buttons, but I have never seen one behave like this. Something in the internal circuitry causing this?

[John Kogel]

So I wasn't crazy after all. I had one like that on an outdoor plug a few years ago. []

They say the 3-light tester is crap, and it is, but there are no mysteries in the circuitry. A light comes on only when there is voltage across the resistor in series with it.

My guess is there is a loose connection in the circuit board of the GFI, or some little integrated circuit chip is suffering from dementia.

Click to Enlarge

55.56 KB

Click to Enlarge

39.3 KB

[Marc]

In John's picture of the autopsied tester, if the bared conductors which cross each other were to touch, voltage to one led would become divided between two of them and cause the middle led to lose voltage. LED's are funny, they'll illuminate with just half as much voltage.

I think Richard's receptacle does indeed have a reversed polarity condition. He needs a new tester too, or open it up like John did and separate the two crossed conductors.

Marc

[rkenney]

LED's are funny, they'll illuminate with just half as much voltage.

Marc

They are neon lamps, not LEDs. I don't believe I've ever seen one with LEDs, although it should be possible, all the newer christmas lights are LED.

[Garet]

I see this once or twice a year. It's not "ordinary" reverse polarity between hot and neutral, instead it's reversed between hot and ground (at least that's what my 3-light tester tells me). That means any metal-cased appliance is now energized to the touch.

I write them up as defective. They always re-trip when I push the test button again, and I've always been able (eventually) to get them to re-set in the normal position. Sometimes I have to re-test/re-set more than once.

[John Kogel]

In John's picture of the autopsied tester, if the bared conductors which cross each other were to touch, voltage to one led would become divided between two of them and cause the middle led to lose voltage. LED's are funny, they'll illuminate with just half as much voltage.

I think Richard's receptacle does indeed have a reversed polarity condition. He needs a new tester too, or open it up like John did and separate the two crossed conductors.

Marc

Good point, Marc. Arcing between the wires in the tester could definitely produce screwed up readings or flickering lights.

I found with my variac test of last week that the neons won't light with less than 80 volts. So with shorted wiring, you may indeed have only one of the lights coming on, the wrong one.

BTW, I found those pic on the net. The GB testers I buy now are a tiny bit better with little circuit boards in them, I think.

In my bag I have a DMM for double-checking goofy readings.

[rkenney]

The two bare wires at the top of John's 'borrowed' picture connect to the neutral terminal and the ground of the tester.

Close inspection of the picture will reveal that it is upside down. The closest terminal to the lens is the ground, that makes the upper terminal neutral and the lowest terminal hot. This conclusion is supported by the fact that the uppermost neon bulb is fresh (rarely lit) and the bottom two bulbs have done the bulk of work over this testers existence.

Shorting the neutral to ground at this point will make no difference whatsoever. None of these tester even have light patterns for neutral to ground reversal, as far as the tester is concerned they are one and the same.

The description of the occurrence in the OP doesn't say anything about the tester being physically 'jarred' to correct the assumed short (which if it was hot to ground would burn through the wiring), so how was it self corrected?

Look at it again.

[Richard Moore]

Tester worked normally in all other receptacles and I had similar results with a different tester in the same GFCI (just not the initial long interval between readings).

[John Kogel]

The two bare wires at the top of John's 'borrowed' picture connect to the neutral terminal and the ground of the tester.

Close inspection of the picture will reveal that it is upside down. The closest terminal to the lens is the ground, that makes the upper terminal neutral and the lowest terminal hot. This conclusion is supported by the fact that the uppermost neon bulb is fresh (rarely lit) and the bottom two bulbs have done the bulk of work over this testers existence.

Shorting the neutral to ground at this point will make no difference whatsoever. None of these tester even have light patterns for neutral to ground reversal, as far as the tester is concerned they are one and the same.

The description of the occurrence in the OP doesn't say anything about the tester being physically 'jarred' to correct the assumed short (which if it was hot to ground would burn through the wiring), so how was it self corrected?

Look at it again.

You're probably right, but could be wrong. []

If the wire from the red lamp (top of the pic) to the ground pin was broken off at the ground pin, the tester would appear to work normally on properly wired outlets. Then if that stray wire crossed over to one of the hot wires of the middle bulb during a test, some voltage would be drained off to the red light, maybe enough to light it while robbing the middle bulb. The resistors are there to prevent a dead short condition from burning the wire.

Just saying it is possible. Richard properly proved that the GFCI is screwed up, by trying another tester.

Why do we use these crappy testers in the first place? Well, gremlins tend to steal them, they are easily replaced, and handy to use.

[Scottpat]

Tester worked normally in all other receptacles and I had similar results with a different tester in the same GFCI (just not the initial long interval between readings).

I'm thinking Indian burial grounds or the tester was once owned by Jimmy Swaggart..[]

[rkenney]

Suretest and a 'wiggy'

[Jim Katen]

This "amused" me for a while today. I tripped an exterior receptacle with a GFCI tester and heard the click coming from the nearby interior. OK so far. Later, when I'm in that room, I find the GFCI contaminated with paint and had trouble resetting it. With a basic 3-light tester in place, I finally manage to get it to turn on, but find it indicating reverse polarity. . . .

This used to be quite common with older GFCI devices. We've had discussions about it in the past. With some of them, you could induce this condition by tripping the GFCI and then resetting it by pushing on only one side of the button or by only pushing the button partway in.

It seems like I used to get questions about it every few months. Now, with the newer GFCIs, it hardly ever seems to happen anymore.

[inspector57]

Ditto what Katen said. Certain vintage GFCI do this with surprising frequency in my experience.