43v in circuit when breaker switched off

[SDmark]

What voltage is your track lighting ? Does it have a dimmer?

There is no track lighting yet, just a junction box in the ceiling with wires sticking out. It will be line voltage. I've installed an old X10 dimmer switch for now; this switch has a little safety toggle (for changing bulbs) which is currently off. Also, that is on circuit #4, not #3 or #13.

Is your breaker box bonded to ground properly ?

The main box certainly is--a thick wire runs from it to a grounding bar. I was here when Cox hammered that in--3 or 4 feet of copper half an inch in diameter. (Cox did it for cable TV; I attached the main panel ground wire to it, maybe because I thought it was better than copper plumbing, or maybe because I was moving plumbing around.) That should cover the subpanel too, right?

 

[Jadnashua]

Living in SD, you might need a MUCH longer ground rod to get a good one, and I think, code requires two.

 

[Sponsor]

 

[Widgit Maker]

Disconnect the neutral going to addition from neutral buss and do a voltage check between wire add neutral buss to see if you have voltage on neutral. Shouldn't have.

At least the cable under the shingles is UF, not Rromes, better insulation.

 

[SDmark]

Living in SD, you might need a MUCH longer ground rod to get a good one, and I think, code requires two.

Okay now you guys are just making stuff up, lol . Actually, it was probably 8 feet and not 4 feet. The reason I think that is that when I moved and upgraded the pool equipment in 2003, I installed a new subpanel for the pool electric. The inspector made me install a separate ground rod for that. I found the 6/28/2003 Home Depot receipt: UPC 782856331506 - 5/8 GRND R - $12.50 (plus tax). If you Google that UPC, you'll see it's for an 8-foot ground rod. The inspector did not say anything about installing more than one per panel. Considering how dense the mud is in my yard, I'm surprised the Cox guy buried the rod with a sledgehammer, as I recall. I used a rented rotary hammer, basically a mini jackhammer.

[From photo]Is this white wire part of 240 circuit?

No, that is breaker #12, one of the ones that is always off.

At least the cable under the shingles is UF, not Rromes, better insulation.

I have a tiny attic over part of the house, above the subpanel. In there, the cable is an old-style non-metallic with a kind of fabric wrap. This, I think, is what goes near the roof, then back down to the junction box in the kitchen. Exiting the kitchen, it goes to Underground Feeder, which is buried under a patio for a short run (15 or 20 feet) to the master bedroom addition.

Disconnect the neutral going to addition from neutral buss

I did see this, thanks. Will try to get it done in the next day or two.

 

[hj]

WE are chasing rainbows trying to diagnose this over the Internet using your testing procedures. Someone has to be there to do his own testing and evaluation.

 

[DonL]

WE are chasing rainbows trying to diagnose this over the Internet using your testing procedures. Someone has to be there to do his own testing and evaluation.

So far, I blame it on X-10 junk. I know what it is, because I own a bunch of it.

 

[SDmark]

Disconnect the neutral going to addition from neutral buss and do a voltage check between wire add neutral buss to see if you have voltage on neutral.

This time also measuring millamps. The multimeter shows down to the hundredth mA. Common neutral from #3 and #13 to buss:

Both off: 19.6v, 0.00 mA
#3 on: 62.6v, 0.01 mA
#13 on: 76.5v, 0.02 mA
Both on: 4.6v, 0.00 mA

[From photo]Is this white wire part of 240 circuit?

Got curious about why a white wire on a breaker. White is on #12, below the 240v breaker on the right. Black is on #5, opposite bus bar (above the 240v breaker, on the left). Why would someone do that? Both are unused and stay off.

So far, I blame it on X-10 junk.

Seems unlikely since I get "ghost" voltage with every breaker except #3 off (see post #16). Unless one of the breakers is bad and is allowing current through even when switched off...

 

[DonL]

Seems unlikely since I get "ghost" voltage with every breaker except #3 off (see post #16). Unless one of the breakers is bad and is allowing current through even when switched off...

They can feed voltage back thru another circuit, Because they use power so they can transmit and receive using the Power lines on both phases.

You should remove it from the equation, Just for grins.

If you are measuring current on AC lines that may or may not be hot, You should get some spare fuses for your meter.

Good Luck.

 

[Widgit Maker]

If you have 19.6 volts between the white wire and the neutral buss with both breakers off and the wires in the junction box disconnected, you have something back feeding voltage to the neutral (white wire). With both breakers off and the neutral (white wire) disconnected in the sub-panel, all wires disconnected in the junction box, the cable between the sub-panel and the junction box is just a piece of wire in the wall. It is not connected to your electrical system at all. Therefore there must be some other connection to the electrical system.

If I understood your description correctly, the cable leaves the sub-panel goes up to the attic, then goes under the roofing shingles and then down the kitchen wall to the junction box. If you have access to the attic I suggest looking for some connecting cable there.
I afraid that this is going to require physically tracing the cable to find this voltage source.

Let us say that someone pulled a hot to a light bulb in the attic, then pulled a neutral form the cable between the sub panel and the junction box. That could put a voltage on the neutral when the light is on. Now wiring something like that would not meet any provision of NEC code. It certainly wouldn't meet the "Don't be stupid" provision of the NEC.

While that seems pretty far out, from the your description the wiring it seems to be pretty hodge podge. Anyone who would run cable underneath the shingles just might do something like that.

 

[SDmark]

I afraid that this is going to require physically tracing the cable to find this voltage source.

Of the estimated 80' run, maybe 15' is in the attic and 4' near the roofing material The remaining 60' feet or so are inside walls or beams. So, not so easy.

you have something back feeding voltage to the neutral (white wire). With both breakers off and the neutral (white wire) disconnected in the sub-panel, all wires disconnected in the junction box, the cable between the sub-panel and the junction box is just a piece of wire in the wall.

I understand this. What I don't understand: Doesn't that power have to come FROM somewhere? I.e. shouldn't I be able to identify the circuit feeding the 43v?

I did another test. The subpanel is fully reconnected (all breakers inserted, neutral wire reconnected, cover on). At the outside main panel, I turned off all breakers (pool subpanel, office, sauna, etc.) except the 60A to the indoor subpanel. At the indoor subpanel, I turned off all breakers but #13. That should eliminate all X10 devices, the doorbell transformer, the yard lighting transformers, the computer UPS. Then in the kitchen, with the wires at the junction box separated, I measured voltage. I get 120v on #13 and 46.7v on #3. Then at the main panel outside, I turned off the indoor subpanel. At the kitchen junction box, I got 4 millivolts on #13 and 0.9v on #3 . So the "ghost voltage" drops significantly when the indoor panel is switched off from the main panel, but turning off every other breaker in the subpanel doesn't affect the ghost voltage. Where in the world could the ghost voltage be coming from?

I guess I could re-do the test but instead of turning breakers off, physically pull them from the subpanel, taking the opportunity to confirm no continuity across the breaker when turned off. That should eliminate power seeping through a bad breaker.

What about grounding...the rooftop TV antenna as well as the Cox cable share a grounding rod with the main electrical panel...could one of those be putting power on the neutral buss? And some of the kitchen ground is off an independent wire that comes from who knows where, maybe a plumbing pipe, although as far as I know, there is no ground in the kitchen junction box we're dealing with.

Jim stated in post #34 that GFCI would trip on 5 mA. If these ghost voltages are at max 0.02 mA (post #47), is all this really worth worrying about?

 

[Jadnashua]

Power=Volts*amps...you have no power there. If it was a good connection to the mains generating that voltage, there would be LOTS of power used. You can chase your tail forever on this. When one wire runs next to another, you can get induced voltage, but unless it is ideal, like say in a transformer where it is supposed to pass power from one side to the other...you have minimal power generated in the alternate set of wires.

 

[Widgit Maker]

Where in the world could the ghost voltage be coming from?

That is what we are trying to figure out.

Right now we are not trying to figure how we are trying to figure where it got there.
When you know where you can find out how.
If I understand your last post correctly, you turn off everything in the main panel except the feed to the sub-panel. In the sub-panel you turned off everything except #13 (black). Yet you have 46 volts between #3 (red) and neutral.
That means the only source of the voltage on #3 (red) is #13 (black) between the sub-panel and the junction box, or one of the cables between the main panel and the sub-panel, or the service lines into the main panel. I think we can eliminate the main lines into the house. Lets see if we can eliminate the cables between main panel and the sub-panel. In the main panel, if you can turn off the leg that feeds the upper half of the sub-panel where #3 (red) is. Check the voltages at the junction box. Then repeat turn off the cable to the lower half of the sub=panel, turn off #13 (black) and turn on #3 (red) Check your voltages. If you have a double breaker in the main panel and can't turn off the legs independently, Just disconnect the wire from the breaker. I wouldn't expect to see any differences in the reading, but we are desperate here.
Another way would be to run a temporary 2-wire cable, doesn't matter what size, between the main panel and the sub-panel.

Now there is no point in doing any of this if the cable between the main and the sub-panel and the cable between the sub-panel and the junction box are physically separated such that there is no possibility of them coming in contact with each other.

After doing this or not doing this you should be able to say with certainty that the that the voltage on the red is coming from the black between the sub-panel and the junction box. You have voltage on a neutral, but not with everything else in the house turned off. Therefore the voltage on the red cannot be coming from the white.

Two things to keep in mind, although I don't know what they mean in this little adventure right now.

You could have a connection between two wires that is so poor or bad that almost no current flows between the two wires. A voltage check would show voltage on the second wire. It might show full voltage or it might show a reduced voltage, all depending on the connection and the resistance of the connection. An example of that might be that you are working with stranded wire and one strand sticks out and comes in contact with another wire. A voltage check would show voltage but there is almost no current flow (amperage).

Second thing to remember is that when you do a voltage check you are measuring the difference in voltage between two conductors, not how much voltage is on either wire. When you do a voltage check between hot and neutral you assume that the neutral is properly grounded. Therefore you reading is an accurate reading of the voltage on the hot. If the neutral is not properly grounded and the connection has a lot of resistance, your reading will not be accurate with respect actual voltage on the hot. You could also have a connection between the black and the red but the connection is so poor or so weak that there is a 46 volt drop across the connection. In other words you have 120 volts on the black and 74 volts on the red. Your meter would show a 46 volt difference. How could you have such a connection? You could have cracked insulation and have moisture making the connection. Remember this connection is so weak that no or almost no amperage is passing.

How is the weather there? Having any rain, any snow on the roof?

You may have to set up that separate ground to get accurate voltage measurements.

God my fingers are tired, I am not very proficient at the keyboard and I am sure that this is more than you ever wanted to know..

 

[DonL]

What about grounding...the rooftop TV antenna as well as the Cox cable share a grounding rod with the main electrical panel...could one of those be putting power on the neutral buss?

There is power on the main cable line, But it should not make it to your house, Unless something is wrong with a line tap.

 

[Jadnashua]

One thing to keep in mind...the voltage flows through BOTH the hot and the neutral. If it didn't, there would be an open circuit, and no current COULD flow. It doesn't just magically disappear to ground...it needs a complete path back to the power station. In the way most homes are wired in the USA, the neutral is often referenced to ground, but it is still a current carrying conductor. At the transformer supplying your home, the input is much higher voltage, and on the secondary (the side going to your home), L1 and L2 come from the ends of the secondary winding, while the neutral is from the center...measuring from one end to the other gives you the 240vac, but from either end to the middle, 120vac. Now, the neutral (centertap) might not be perfectly in the center, but it will be close which is why you might get slightly different voltages between L1-N and L2-N. And, if your load isn't balanced between L1 and L2, the neutral will be carrying more power than ideal.

When you insert your meter between two wires, you are completing a circuit. A modern DMM has a very insignificant load (hundreds of megaohms, typically), so that it does not load the circuit down. This is the reason why you can read voltages that have no current carrying capacity behind them. If you were to use an older, analog meter, it creates a much bigger load on the circuit, and you might just get readings of zero where you're seeing higher ones now. It is not all that hard to induce some voltage potential in adjacent conductors. This becomes very important when dealing with small signals, and is one reason why they use twisted pairs in things like low-level audio connections and Ethernet cables. You can chase your tail a long time on this. This might help a little after some study...

http://physics.about.com/od/electromagnetics/f/KirchhoffRule.htm

 

[SDmark]

Folks, I REALLY appreciate all your effort and input on this. It is frustrating that we haven't been able to fully resolve it yet. Maybe additional testing could narrow it down (in addition to Widget Maker's suggestions, we also haven't tried testing other circuits from the subpanel), but I have to repeat the last question of my previous post: is it worth worrying about? If a GFCI trips before amps become dangerous at 5ma, and my maximum 43v readings are fluctuating between 0.00 mA and 0.02 mA, doesn't that mean I would need at least 250 times as much power as I have for it to be dangerous? Maybe I can discuss with an electrician acquaintance, but if it's safe, I may have to leave the mystery unsolved.

 

[DonL]

Maybe I can discuss with an electrician acquaintanc

Get a 10-100K resistor from him. Have him tell you how to use it properly/safely with your meter and your PPE.

Good Luck.

 

[WorthFlorida]

SDmark,
Your beyond your expertise trying to figure this out. I mentioned about 50 posts ago to get a electrician and the first thing is to check the grounding. So now your have your head buried in that. And that picture of the panel with a WHITE wire connected to a breaker, RED FLAG 1. Even a poorly trained electrician wouldn't do that but you did mention a previous homeowner pulled the permit. RED FLAG 2. There are times that a electrician needs to use the white wire (switched circuit) but they must wrap the ends with black tape but never be on a breaker. It seems like you're a trouble shooter for data or telephony as I was for 40 years and you love to find causes of problems. And I did spent many hours on the phone troubleshooting by giving advice or asking for advice. But the AC power for your home is not the same.

I do not know of any telephone or data wiring every cause a building fire or deaths due to electrocution. So drop everything and get a qualified electrician there and spend the money, first by replacing that scary looking panel. Electrical parts do wear out, leakage can start, bakelite does get brittle over time if subject to high heat for many years. I've even had a breaker appeared to snap open but it was still electrically connected. The contacts internally were welded together.

If you read the 2nd and third nail posts of this forum it mentions safety as first and foremost and no advice is to be given that can cause harm to a person asking for help.

Enough said.

 

[SDmark]

WHITE wire connected to a breaker

At least that breaker was marked "unused" and not turned on. Earlier today I went through and disconnected and capped wires from three unused breakers, including the white wire.

 

[SDmark]

get a electrician

Just curious: is anyone participating in this discussion an electrician?

 

[Jadnashua]

FWIW, I spent years troubleshooting radars and missile launchers, so I have a fair idea of electricity and electronics. Not to say that I may be wrong, nobody's perfect. A worn or defective CB may not fully make the disconnect, as could corrosion on the bus bar. Not enough continuity to provide any appreciable power, but enough to measure when you use a DMM with hundreds of Mohms of resistance...essentially, almost an open circuit, just like a dead battery showing voltage until you apply a load. That's why an old, analog meter might show a different value...it provides a load on the circuit...not huge, but thousands to millions of times larger than a modern DMM.