Very unusual phantom voltages

[PetrosA]

I'm stumped. I had a service call today to repair a few two-tube wraparounds in the second floor (finished space) of a two bay garage. When I got there, it turns out that the whole project must have been done by a handyman or homeowner at some point in the past - white conductors in the NMs were used as hots; each of the three lights was on a separate switch; there were three handy boxes for the switches with individual conductors passed through adjacent KOs, and the worst was what I saw when I took the cover off the ballast compartment: the black and white wires from the ballast were passed through a 1/4" hole in the top of the fixture and disappeared into the ceiling. The one fixture didn't work at all, so I got my pencil tester out, which rang on the white wire. I checked the switch, and sure enough it was bad. I repaired the switch and the light worked again. Out of curiosity I put my pencil tester against the frame of the wraparound and it lit up. At that point I decided I'd better make sure that I was seeing juice on the white so I got my meter out and an extension cord to get a good ground... or so I thought.

And here's where it gets weird. Building is supplied with 120/240 single phase. Sure enough, the extension cord showed that the neutral slot was hot just like the lighting circuit, and the ground and hot slots were not. From neutral slot to wraparound frame my Hioki showed 265V, but it's a high impedance meter so I went and got my Agilent U1272A. It showed 359V on high impedance setting, and zero volts on the Low Z setting between frame and any of the slots on the cord. Based on that reading, I don't think there is actually dangerous voltage on the light but I am really confused at the high impedance readings I got on both meters.

I've seen phantom voltage many times, but it was never higher than phase to ground or phase to phase voltage. Why would I be seeing a reading that was higher than phase to phase on a single phase service? At first I thought that maybe the lighting and receptacle circuits were on opposite legs, but the zero reading in Low Z mode should have shown me something in that case.

I phoned the property manager and let him know there are electrical issues in the building and that I would send him a report to pass on to the owner. In the meantime I'm hoping someone here can shed some light on those readings for me.

 

[GoldDigger]

All that I can think of would be that the ballast(s) contain, in effect, a step up winding feeding the tombstones and the capacitive coupling from that to the metal of the light is what is producing the high voltage you saw.

 

[mgookin]

Did you follow the circuit back to the service?

 

[SG-1]

Your floating conductor may be charging toward a peak to peak value.

 

[ptonsparky]

Were the fixtures actually grounded via a known good EGC?

 

[PetrosA]

All that I can think of would be that the ballast(s) contain, in effect, a step up winding feeding the tombstones and the capacitive coupling from that to the metal of the light is what is producing the high voltage you saw.

The fixture I measured still had a magnetic ballast in it. You may be on to something.

Did you follow the circuit back to the service?

I did not. I was only called in to get the lights working again, and when I asked to see the panel, it was in the lower part of the garage and not accessible for all the junk in the way. I will try and convince the owner to let me back in there to get things straightened out.

Your floating conductor may be charging toward a peak to peak value.

That was one of my thoughts but I wasn't sure how (two) TRMS meters would be reading a peak to peak value. It also seemed strange that each meter was reading so differently, but I scratched that off to the fact that this is some kind of induced voltage reading.

Were the fixtures actually grounded via a known good EGC?

I'm not sure that there's even a good ground anywhere in the building. What I saw at the panel is a conduit coming up from slab to a meter socket that contained a not-so-POCO-looking meter, then SEU out of that to a 6 circuit SquareD panel. I wasn't able to determine (due to lack of access) whether this was a separate service or whether it's fed from one of the multi-family houses on the property or whether there's a ground rod or anything. It's a bit of a mess.

 

[SG-1]

I did an experiment using a 3kva 480 to 120 transformer & was able to obtain phantom voltage readings greater than the supply voltage. That was many years ago & I do not recall the specific set up.

 

[PetrosA]

I did an experiment using a 3kva 480 to 120 transformer & was able to obtain phantom voltage readings greater than the supply voltage. That was many years ago & I do not recall the specific set up.

Between your thoughts and Golddigger's, I think I have enough theory to not write anything alarmist about voltage in my mini-report to the owner. Thanks!

The thing I hate about jobs like this is the fear factor. On a property where a lot of electrical has obviously been done by unqualified individuals, you never know what's going to be energized - could be luminaires, could be pipes, could be plaster walls, could be device plates, could be a small chunk of knob and tube wire sticking out of a joist. Fear Central.

 

[kwired]

Non bonded metal objects in close proximity to ungrounded conductors will capacitively couple enough voltage to light up a non contact tester, it is one fairly simple way to look for disconnected or misconnected EGC's. EI's around here test about every switch and luminaire they can easily reach with a non contact tester while it is energized. I once had one that lit up even though it should have been well grounded, he let it go after looking harder at it and determined it was grounded, but it is an easy way to catch non bonded objects without having to disassemble anything.