What would cause these voltage readings?

[gar]

160813-1024 EDT

iwire:

For some information on current levels that nerves can detect see https://en.wikipedia.org/wiki/Electric_shock .

Now to phantom voltage. It seems this is generally defined as a voltage that is read by a high impedance meter that disappears when a low impedance is placed across the meter.
Basically the phantom voltage source has a high internal impedance.

Some internet discussion on phantom voltage:
http://support.fluke.com/find-sales/Download/Asset/2105317_A_w.pdf
http://www.allaboutcircuits.com/textbook/direct-current/chpt-3/ohms-law-again/ this has good information.

Capacitance in wiring or cables is what normally determines the leakage current that leads to phantom voltage readings.

Look at Romex construction. There are three wires, two are current carrying, and the third is the EGC. The physical topology has the EGC centered between the two current carrying wires. Thus, the capacitance between each current wire and EGC is the same and is proportional to cable length The phantom voltage on an isolated EGC is not dependent upon current flow in the current carrying wires, but on the AC voltage difference between the current wires. Since the two capacitance values are equal the EGC should be at a potential halfway between the two current wires.

The equivalent circuit is a voltage source equal to 1/2 the voltage between the two current wires and an internal impedance from the sum of the two capacitors. In reality there are other factors that modify this equivalent circuit some. The original post had somewhat unequal voltages. Thus, there are external factors that modify the circuit. For example the meter internal impedance.

.

 

[iwire]

160813-1024 EDT

iwire:

For some information on current levels that nerves can detect see https://en.wikipedia.org/wiki/Electric_shock .

Now to phantom voltage. It seems this is generally defined as a voltage that is read by a high impedance meter that disappears when a low impedance is placed across the meter.
Basically the phantom voltage source has a high internal impedance.

Some internet discussion on phantom voltage:
http://support.fluke.com/find-sales/Download/Asset/2105317_A_w.pdf
http://www.allaboutcircuits.com/textbook/direct-current/chpt-3/ohms-law-again/ this has good information.

Capacitance in wiring or cables is what normally determines the leakage current that leads to phantom voltage readings.

Look at Romex construction. There are three wires, two are current carrying, and the third is the EGC. The physical topology has the EGC centered between the two current carrying wires. Thus, the capacitance between each current wire and EGC is the same and is proportional to cable length The phantom voltage on an isolated EGC is not dependent upon current flow in the current carrying wires, but on the AC voltage difference between the current wires. Since the two capacitance values are equal the EGC should be at a potential halfway between the two current wires.

The equivalent circuit is a voltage source equal to 1/2 the voltage between the two current wires and an internal impedance from the sum of the two capacitors. In reality there are other factors that modify this equivalent circuit some. The original post had somewhat unequal voltages. Thus, there are external factors that modify the circuit. For example the meter internal impedance.

.

All very nice.

I still maintain you will not get a 'tingle' as the OP described it from phantom voltage in a residence.

 

[peter d]

All very nice.

I still maintain you will not get a 'tingle' as the OP described it from phantom voltage in a residence.

I agree. Common situation is where you have a MWBC and you shut one circuit off, the conductor that is turned off will almost always read ghost voltage from the proximity of the live conductor in the cable or conduit. This conductor will not cause a shock or even a tingle.

 

[gar]

160813-1745 EDT

iwire:

Today my hand to hand resistance is about 2 to 3 megohms using a Fluke 27, and about 1.5 to 2 megohms using a Simpson 260.

With a DC supply and 100,000 ohms current limiting resistance I start to sense current at about 100 microamperes. At this level it is not yet a tingle.

You can experiment on yourself with 120 V or less as the source. At 120 V and a 1 megohm current limiting resistance maximum short circuit current is 120 microamperes. With 680 k max current is 177 microamps, with 330 k max current is 364 micro amps, and with 100 k max current is 1200 microamps or 1.2 mA. Because of the likely high resistance of your skin the stimulus currents will be less than the maximums under short circuit conditions.

.

 

[kwired]

Bob I agree with you. But here is what I think is happening. I may not even be crazy because I have seen it happen.

Hot to neutral = 113
Hot to ground= 68
Ground to neutral=42

You would assume that the 42V that he is reading is from the neutral to ground but I'm thinking the voltage is on the ground wire.

Had a problem with the same crazy reading a few years ago. Turned out to be a big old flying splice in the attic above the bathroom. Tape had burned to carbon creating a resistor and I'm reading the voltage drop across the resistor from hot to ground (open ground).

I think he's going to find a bunch of burned wire, wire nuts or tape.

Needs to test with a load applied to the circuit and see what changes there are on the measurements, even just measuring with a low impedance meter may be sufficient enough to get more idea of what is going on here.

I don't run into the carbon trace problems with inside wiring too much, but do get funky readings often on failed underground wiring, adding (or attempting to add) some load, or sometimes removing load, will usually make the problem more obvious after you see how the measurements respond to that action.

 

[GoldDigger]

If the bond is left out at the service panel where the EGC is created, the bond at the POCO trans is not relevant. The potential between the homes EGC and neutral could be anything.

If there is no significant load current (no load connected), then the house EGC voltage and the POCO neutral voltage will not differ by 50V since they are both connected to earth, even though the ground electrode resistance may be high.

Think some more about what you said, and draw a wire diagram including ground electrodes. Now if you leave out both the neutral to EGC bond AND the ground connection of the house EGC, what you said is possible. But I would just call that an open EGC, not an unbonded EGC.

 

[iwire]

If there is no significant load current (no load connected), then the house EGC voltage and the POCO neutral voltage will not differ by 50V since they are both connected to earth, even though the ground electrode resistance may be high.

If there is no load you are correct and why I said the voltage could be any value.

I don't know why you want to assume no load, the OP is in the house working, very likely there are other loads on. The branch circuit the OP is working on may have no load while the homes service neutral likely is loaded to some point.

Think some more about what you said, and draw a wire diagram including ground electrodes. Now if you leave out both the neutral to EGC bond AND the ground connection of the house EGC, what you said is possible. But I would just call that an open EGC, not an unbonded EGC.

I don't need to draw it.

 

[GoldDigger]

Phantom voltage will not produce a shock.

And the chances of two phantom voltages happening to add up to more or less line voltage is slim.

Think on that some more.
.

1. Depends on what you define as a shock. If you allow "a small tingle" to be called a shock, then capacitive coupling in a long (several hundred feet would certainly do, probably less) run of NM or other closely coupled wires could produce a shock.

2. I thought about that some more. Here is what I thought:

Assume an open ground wire with little coupling to any actually grounded surface (a run of NM through a wood framed wall, for example).
The voltage on the EGC will be somewhere between the voltage on the hot and the voltage on the neutral since there is a capacitive voltage divider from hot to EGC to neutral.
By the very nature of a voltage divider, the sum of the hot to EGC voltage and the EGC to neutral voltage must equal (roughly) the hot to neutral voltage in the absence of a meter load. Because there is some parallel resistance even with a high input impedance meter both readings will be slightly less than the "free" voltage reading and so we expect, as was seen by the OP, that the sum of the two voltages will be slightly less than the hot to neutral voltage. The larger the capacitances involved the lower the discrepancy.

Hot to neutral = 113
Hot to ground= 68
Ground to neutral=42

Both the measured voltages and the perceived shock are consistent with a rather large capacitive coupling. There are other explanations though. But IMHO an open neutral is not one of them.

 

[iwire]

1. Depends on what you define as a shock. If you allow "a small tingle" to be called a shock, then capacitive coupling in a long (several hundred feet would certainly do, probably less) run of NM or other closely coupled wires could produce a shock.

The OP is working in a house which suggests fairly short runs and less than max 240 volts. There is not enough capacitance coupling in those situations to provide any real current.

The runs are not 'several hundred feet' in a homes branch circuit.

On the other hand the parking lot lighting runs I work on are and even with a phantom reading over 200 volts there is not enough current to give me a shock.

2. I thought about that some more. Here is what I thought:

Assume an open ground wire with little coupling to any actually grounded surface (a run of NM through a wood framed wall, for example).
The voltage on the EGC will be somewhere between the voltage on the hot and the voltage on the neutral since there is a capacitive voltage divider from hot to EGC to neutral.
By the very nature of a voltage divider, the sum of the hot to EGC voltage and the EGC to neutral voltage must equal (roughly) the hot to neutral voltage in the absence of a meter load. Because there is some parallel resistance even with a high input impedance meter both readings will be slightly less than the "free" voltage reading and so we expect, as was seen by the OP, that the sum of the two voltages will be slightly less than the hot to neutral voltage. The larger the capacitances involved the lower the discrepancy.

Both the measured voltages and the perceived shock are consistent with a rather large capacitive coupling. There are other explanations though. But IMHO an open neutral is not one of them.

That is something about which even two intelligent people might disagree, so I am not at all surprised that you disagree with me.

 

[kwired]

The runs are not 'several hundred feet' in a homes branch circuit.

On the other hand the parking lot lighting runs I work on are and even with a phantom reading over 200 volts there is not enough current to give me a shock.

If there is enough power ability behind the voltage to give you a shock it is real voltage and not phantom voltage. If it can sustain that voltage with even as high as 10k ohms of load across it it is likely real voltage and not phantom voltage.

 

[big john]

To me "phantom" is just synonymous with "coupled voltage."

Trying to make a distinction where it can't be enough to cause a shock seems unnecessarily complicated: We're only using the term to describe that there is no physical conductive path to the wire being tested with a DMM.

More on topic with the OP, ignoring for a moment the many detailed replies here, it sounds like a simple open EGC.

 

[mivey]

All very nice.

I still maintain you will not get a 'tingle' as the OP described it from phantom voltage in a residence.

I have experienced it first-hand and recently as well (within the past few years). I am within a 40 minute drive of that very fixture. Test-lighted and metered it but it did not shock me the second time (being the consumate professonal I could not resist bare-handing it for a check).

Don't follow the mistaken belief that phantom voltage is only capacitive coupling as it can also be a dielectric breakdown like a high-impedance fault. It may cause a tingle but not be enough to carry normal currents associated with a normal impedance load.

 

[iwire]

I have experienced it first-hand and recently as well (within the past few years). I am within a 40 minute drive of that very fixture. Test-lighted and metered it but it did not shock me the second time (being the consumate professonal I could not resist bare-handing it for a check).

Don't follow the mistaken belief that phantom voltage is only capacitive coupling as it can also be a dielectric breakdown like a high-impedance fault. It may cause a tingle but not be enough to carry normal currents associated with a normal impedance load.

I made clear my view was an open conductor with no real connection as described in the NEMA bulletin I linked to.

 

[kwired]

I guess it depends on how you want to define "phantom voltage".

Merriam Webster definition of phantom says:1
a : something apparent to sense but with no substantial existence : apparition
b : something elusive or visionary
c : an object of continual dread or abhorrence <the phantom of disease and want>
2
: something existing in appearance only


So I would think the voltage is real, but it is the current limiting ability of what may be going on that causes it to appear to not be real when you try to make it deliver any significant current to a load. Most common cases is capacitive coupling where the capacitor that is present just isn't enough of a capacitor to allow any significant current to flow.

A high impedance fault will have sufficient power source behind it, you just happen at be at two points in the circuit where if you touch them both you happen to be a higher resistance then whatever else is between those two points, leaving low enough voltage across you that no significant current flows through you.

 

[mivey]

I made clear my view was an open conductor with no real connection as described in the NEMA bulletin I linked to.

If restricting the phantom voltage to capacitive coupling on an otherwise unconnected and uncoupled conductor in a residence then it is not likely that is a scenario that will create a tingle.

Since the OP got a tingle, it is worth exploring other types of phantom voltages that can't sustain appreciable currents. The term "phantom voltage" was around long before NEMA published the bulletin you referenced so they hold no patent on the term's use.