The post light and strange happenings

[mark32]

So, my customer has five post lights on their property, and one stopped working. I remove the fixture in question and read 98 volts from the ungrounded conductor to the egc and the grounded conductor. I see that there are leds in the fixture, so I grab a pigtail socket and an incandescent bulb from my stash, and it lights up as I figured it would. (I will mention here that all of these post lights are controlled by a single relay inside the main house) At this point, I remove the incandescent and screw in an led, as I want to run into the house and look at the relay cabinet where this circuit originates, and then poke my head out of a window and see if the led lights up. However, the led now lights up as I screw it into the pigtail?! I remove the pigtail and now I have 120v??!! At the relay cab, it looks like two - two wire circuits are wire nutted together off of the relay. The basement is finished, so I cannot view how these cables are routed. I will note that the cables are romex inside the cabinet, but change over to uf at some point. Anyway, it's getting late, so I put the, now functioning light, back in place and clean up my tools. The owner offered me a burger, which I readily accepted However, as I go to leave, what do you think happened? You got it, the dang light is now off again. At this point, I may or may not bother troubleshooting any further, as the homeowner will soon be getting all new landscaping done in this area. So, I may just have his guy trench about 70' away to the nearest post light, and I'll just refeed the light in that manner. But I am curious what could be taking place here. What could cause this circuit to behave like this?

Happy fourth!!

 

[LarryFine]

I suggest an extension cord plugged into a non-GFCI-protected receptacle to measure and compare your circuit's conductors against.

 

[ptonsparky]

Splices can be intermittent, then add moisture, dirt, rust, corrosion, etc. and you get what you describe.

Usually those show again about time you get home and in bed.

 

[Sea Nile]

I would swap the LED with an LED from a post light that is working normally

It could be that all of the post lamps are experiencing an intermittent 98v, and that one LED bulb just happens to have different tolerances. Maybe that bulb is closer to EOL or is a different brand name.

If the problem moves, then you will know.

Also, you can put incandescent lights in all 5, that way there will be a clear indication if they all start dimming at the same time or just that one post light fixture.

 

[hillbilly1]

As @ptonsparky said, it’s probably a bad connection, the incandescent draws much more power, heating up the connection making it expand, decreasing its resistance. Once it cooled, it contracted, increasing its resistance again.

 

[James L]

If 98 volts shows from hot to both neutral and ground, then you're looking at a problem before neutral and ground become seperated.

 

[Joethemechanic]

I had one like that when LED replincadesant acement bulbs first came on the market. Outside post light wouldn't work with LED. Screwed in a 100 Watt incandescent, and it lit, put the led back in and it now worked. A couple days later all the post lights are out. Screw the 100 watt in one fixture and now everything comes back on.

Bad connection in a box covered by 30 some years of landscaping the leds weren't drawing enough amperage

 

[James L]

I wonder if you go to the panel, will you find that the neutral and ground are under the same lug screw, and that maybe it's loose.

 

[Sea Nile]

I wonder if you go to the panel, will you find that the neutral and ground are under the same lug screw, and that maybe it's loose.

Since the voltage is identical from the ungrounded conductor to both the grounded conductor and ground wire. I would bet it's a loose ungrounded conductor. What are the chances that two different conductors have the exact same loose connection?

 

[mark32]

Thank you so much for the quick replies! Wow, it looks like ptonsparky, hillbilly1, and joethemechanic are all on the same page through either experience or in theory. And special thanks to Hillbilly for the explanation, it makes sense, and it could very well be what is taking place here. As far as I can tell, these posts are jumped in normal fashion, so, if I go to the post closer to the house, I'll check out the splice there.

 

[James L]

Since the voltage is identical from the ungrounded conductor to both the grounded conductor and ground wire. I would bet it's a loose ungrounded conductor. What are the chances that two different conductors have the exact same loose connection?

If they're under the same lug screw, really good chance.

If there's a good ground but bad neutral, I would expect the h-g reading to be good.

 

[winnie]

Since the voltage is identical from the ungrounded conductor to both the grounded conductor and ground wire. I would bet it's a loose ungrounded conductor. What are the chances that two different conductors have the exact same loose connection?

I think @James L brings up an interesting possibility.

The same voltage measurement H-G and H-N suggests a failure common to both paths, which strongly points to a problem in the hot conductor. But if both G and N are (improperly) landed under the same screw in the panel, this is another potential single point that would impact both paths.

Most panels permit two G conductors to be landed in a single terminal hole under one screw. However it is generally prohibited to land two N conductors or a G and N conductor. This is because the N conductors carry current and are thus subject to cyclic heating and cooling. Because two conductors _fit_ in the hole, and both G and N are going to the same bus (in a main panel), having G and N from a given circuit share a hole is a very common installation error.

-Jon

 

[James L]

which strongly points to a problem in the hot conductor.

That's something I didn't consider.
Very possible, too

 

[Sea Nile]

If they're under the same lug screw, really good chance.

If there's a good ground but bad neutral, I would expect the h-g reading to be good.

I'll give you that it's possible, but the only place that could happen would be the panel, which would not be subject to the water and corrosion. Also, every light would be affected. Assuming every light being affected I suggested moving the bulb to a different post to rule that out. It could be that the particular bulb affected just has a slightly different voltage threshold than the others.
I agree with hillbilly1 about the increased current providing a small arc at the bad connection to partially fuse the connection. The question now is locating where the bad connection is. Is it at the panel where two conductors are under the same screw? Or is it outside where the one light is being spliced in to the circuit? Don't know untill more tests are done

 

[Joethemechanic]

I've been known to stick one of my meter leads in the dirt to check neutrals and grounds.

 

[Sea Nile]

it is generally prohibited to land two N conductors or a G and N conductor. This is because the N conductors carry current and are thus subject to cyclic heating and cooling.

Thanks, I've always wondered what the reasons were. I assumed it was only because you didn't want to risk inadvertently loosening a neutral for a MWBC when isolating another circuit, which wouldn't be possible if every neutral has it's own designated spot. But cycle heating improves my understanding a bit, so thank you for that.

 

[Little Bill]

I would put the incandescent bulb back in and leave it. Watch it for a few days. That or try another LED.

 

[ptonsparky]

Put a temp receptacle on the circuit, load it up with a heat gun. That may clear up the intermittent connection.

 

[Sea Nile]

Put a temp receptacle on the circuit, load it up with a heat gun. That may clear up the intermittent connection.

Noooooo. That will just make the bad connection harder to find. Its like cracking in egg in a car radiator before selling the car. There is a bad connection, it should be found.

 

[ptonsparky]

Noooooo. That will just make the bad connection harder to find. Its like cracking in egg in a car radiator before selling the car. There is a bad connection, it should be found.

It should open it.