Chasing electrical gremlins - GFCI outlet puzzle

[Empty_Nest__Soo]

On our trip at the end of last month, a table lamp quit working. Changing the bulb did not help, but a voltage detector showed voltage present at the base of the socket. I figured the switch in the lamp socket died.

I finally got around to buying a new lamp socket and I installed it today. No joy.

Using an outlet tester on the receptacle showed hot and ground reversed! I checked the breaker panel, and things were wired correctly on that end of the circuit. Some other outlets on that circuit checked out okay so I began to work upstream from the problem.

Long story short, I found the circuit routed through a GFCI outlet in the basement. That GFCI outlet and everything downstream tested “hot and ground reversed.” I pulled the line side of that outlet and found that the power coming in was correct, black = hot, white = neutral, bare = ground. The GFCI outlet was wired correctly, line side and load side. The problem has to be in the GFCI outlet.

After removing the GFCI outlet and connecting the appropriate wires with Scotch locks / wire nuts, everything downstream works okay.

The GFCI outlet does not seem to want to reset. I will replace it.

Also, I think I will re-route some wiring so that the interior outlets that are not near water, etc., are not routed through the GFCI receptacle, as I see no need for that kind of protection.

My question: I expected that if the GFCI outlet tripped, it would “interrupt” the circuit, i.e., break it. Could it have shunted the hot to ground, thereby activating the ground wires downstream, yet leaving the hot wire dead downstream?

I Googled to see how the GFCI outlets work, but came up empty.

Wayne

[myredracer]

The difference between an analog and digital voltmeter is the impedance. An analog voltmeter is low impedance and a digital one is high impedance. Fluke makes a digital multi-meter with dual impedance.

Excellent explanation from Fluke here including ghost voltages: analog vs digital and high vs low impedance voltmeters

One nice thing about an analog meter is that it doesn't need a battery like digital ones do. I bought a digital one from HF for our TT and I've forgotten to turn it off a few times and killed the battery. 😞

I have an old neon bulb voltage tester similar to the one in the photo. As mentioned in the Fluke article, it could be a ghost voltage but can lead you to conclude that you have a solid 120 volt circuit. Dad gave it to me decades ago and I've never used it for that reason. Same situation can happen with a NCVT.

[Harvey51]

I think an analog voltmeter typically has a full scale current draw of 50 micro amperes. All the current goes through the meter movement. If it says "20 000 ohms per volt" on the meter face, that means 50 micro amps for sure.

That's more than a digital meter but still it leaves room for a misunderstanding with a high resistance short.

[YC_1]

bukzin wrote:

YC 1 wrote:
The breaker did not trip because there was not enough of a load. I know it is hard to visualize seeing a voltmeter read 120 volts and not have enough power to run a light. I cannot tell you how many times I have seen techs caught by trusting a digital voltmeter.

My understanding is that an analog meter avoids this false reading?

Does this also apply to little cheap meters or good ones like a Simpson?

Even a cheap Radio Shack analog meter will keep you out of trouble in a case like this. The meter movement actually requires a minute amount of current to magnetically move the needle.

[Empty_Nest__Soo]

YC 1 wrote:
I would bet a beer on a new one fixing the problem. I have seen plenty of them go up in smoke.

Do try the voltmeter across a battery trick for a real learning experience. Not enough current to even feel but enough to drive a digital voltmeter.

I have an old Simpson 260 analog meter that solves that issue when any doubts come up.

I've not had much experience with GFCI outlets failing. This is only the second one I can remember. I've never done this sort of thing for a living. I have no doubt that a new GFCI outlet will make things right.

I have a Square D analog voltmeter that is probably 50 years old (belonged to my father, who was an electrician). But anymore I reach for one of the digital multi-meters. Never thought of trying the analog model.

Wayne

[bukzin]

YC 1 wrote:
The breaker did not trip because there was not enough of a load. I know it is hard to visualize seeing a voltmeter read 120 volts and not have enough power to run a light. I cannot tell you how many times I have seen techs caught by trusting a digital voltmeter.

My understanding is that an analog meter avoids this false reading?

Does this also apply to little cheap meters or good ones like a Simpson?

[YC_1]

I would bet a beer on a new one fixing the problem. I have seen plenty of them go up in smoke.

Do try the voltmeter across a battery trick for a real learning experience. Not enough current to even feel but enough to drive a digital voltmeter.

I have an old Simpson 260 analog meter that solves that issue when any doubts come up.

[Empty_Nest__Soo]

myredracer wrote:
Whatever you find the cause to be, please post it!

OP here.

I now know that the GFCI outlet did not simply trip, it failed. I say this because when I got curious and opened up the GFCI outlet to see what was happening inside, I found a bit of black soot and a couple of parts burned loose from the little printed circuit.

I can only conclude that the way the GFCI failed caused it to leak voltage to the ground wire, enough for the GFCI and other outlets down the line to show hot and ground reversed and for my digital multi-meter to show 120 volts. As others have said, probably a negligible amount of amps could have passed through if I had applied more of a load to the ground-neutral wires down the line.

This would also seem to explain why outlets and devices in the circuit prior to the failed GFCI were still working normally.

I have not yet replaced the GFCI outlet to verify that a new one will do the job, but I now see no reason to think that there may be another problem. Without the failed GFCI in the circuit, everything works and tests normally. In the unlikely event that problems continue after I install a new GFCI outlet, I will update this thread.

Wayne

[myredracer]

Whatever you find the cause to be, please post it!

[wa8yxm]

Here is something I will describe.. You have a "Water use reducer" washer this is a small washer with a tiny hole (lets make it a pin hole) in it in a line, When you shut off the valves a pressure gauge on the down-line side of the thing shows full line pressure.

but when you open a valve the pressure drops to zero and water basically drips out of the faucet.

I have actually seen this in electronics, the DIGITAL Volt meter showed normal voltage because it used less current (water) than the fault (Washer) would pass.

But a test lamp found it and fast.. (And the fault (washer) was thus removed and full flow restored).

[YC_1]

The breaker did not trip because there was not enough of a load. I know it is hard to visualize seeing a voltmeter read 120 volts and not have enough power to run a light. I cannot tell you how many times I have seen techs caught by trusting a digital voltmeter.

[CA_Traveler]

Everything you described suggests a higher resistance short which would not trip the CB vs a direct short. Voltmeters, testers, etc typically draw so little current that the results can be misleading but testing with a load will often show different results.

[Empty_Nest__Soo]

After checking everything else on that circuit, I decided to take a look inside the GFCI outlet.

There was black residue inside the GFCI outlet. Some of the electronics on the little printed circuit had fried. It did not trip, it failed.

I can only conclude that the GFCI outlet failed in such a way that it activated the ground when it failed.

One thing I haven’t quite figured is, if this shorted to ground, and assuming a good ground to earth, why did the breaker not trip? Perhaps it could pass very little current, not enough to trip the breaker.

At this point, I will replace the GFCI outlet and call it good.

Wayne

[Wayne_Dohnal]

I'll jump on the bandwagon of saying there's more to this than meets the eye. The way the GFCI outlets are mass produced as identical devices, I think the odds of one being able to reverse the hot and ground are astronomical. Coupled with the previous observation that the ground goes 'to', rather than 'through' the outlet, the odds are astronomical-squared.

The outlet tester gives a possible interpretation of the pattern of lights, not a definitive answer. With 2 lights lit in the 'hot-ground reversed' pattern, it's indicating voltage between ground-neutral and ground-hot. Another cause of this indication would be 'neutral connected to hot', which I believe would be consistent with your measurements and observations. While I still find it unlikely that the outlet would do this, the odds are IMO somewhat less than astronomical. There are probably other causes of the outlet tester's indication, too.

[myredracer]

I too would want to know what caused that.

I would think the very first thing you want to make certain of is that the ground inside your MH isn't at 120 volts to earth (as at ground point in your house panel or pedestal) otherwise you'd have a hot skin which is dangerous. Use a non-contact voltage tester to determine that. Google "No-shock Zone".

It's hard to diagnose something like this without being there, but some other thoughts:

Once you know the shore power recept. you're plugged into isn't mis-wired, you might try using the ground bus behind the converter panel as the ground reference point and use a long wire to get to your voltmeter to test hot, neutral and ground to the ground reference point at various receptacles throughout the MH.

Maybe you could also measure resistance between H, N & G in receptacles and at the panel? With shore power disconnected of course. You should have full continuity between the panel and H, N & G in all farthest recept. branch circuits.

If the table lamp is intermittent, that could indicate a faulty connection somewhere. Intermittent connections can be difficult to find. You may want to pull out every receptacle in that branch circuit and inspect the terminations as well as at the panel.

If you have an open wire somewhere, a voltage tester or high impedance voltmeter can give you a false reading. Or a tester or voltmeter may show a circuit is okay with no load, but a bad connection won't support a load.

Not sure if these will help, but have a look anyway.
The circuit detective
Home Improvement help

In this Youtube vid a master electrician says not to trust a plug in tester if it says hot and ground reversed.