I was on a job troubleshooting some a pair of three way switching systems when I noticed that when I checked the voltage on the "unenergized" switch leg I was reading 34 volts. The guy running the job said it was induced voltage and not to worry about it. I didn't think about it again until I was on a job and turning on the the two breakers that feed a 12/3 mc that fed two circuits of recepticles. I turned one on and just to do some checking I checked the wire that was on the red wire of the mc and although the breaker was off, it had 34 volts on it.
I touched the wire and put my finger to ground and could feel nothing. I remembered someone telling me about a kind of phantom instrument reading or something and wondered if anyone had any explanations on why this would be.
Could it be a neutral or proximity problem, or do you agree with the opinion I got that it was "induced"?

High impedance, digital meters see a "ghost" voltage because they put virtually no load on the unenergized conductor.

The trick is to use a solenoid type voltage tester.

In the electronics world very high impedance is a wonderful thing. You meter does not affect what the circuit is doing. In "power" circuits, this high impedance is not so great. You can see induced voltages on "floating" wires that have no effect. I am very surprised digital meters meant for the "electrical" trades don't have a built in load resistor (maybe switchable) that makes these "phantom voltage" readings go away. A megohm would probably do it and still have negligible affect on your readings.

Quality digital meters do have a switchable load resistor, at least in Europe (Benning Duspol for example). Problems like this are very common with DIYers using cheap DMMs or neon screwdrivers on 230V lines and then complain about the weird readings.

On the other hand, sometimes a neon screwdriver that lights up in both holes of a receptacle does indicate severe trouble... once found a broken neutral that way.

A good ole fashioned analog meter comes in handy for these occassions. We've made it standard procedure to doublecheck with a lower-impedance meter whenever a fluke picks up suspected ghost voltages.

When do you suspect phantom voltages?

Obviously when the breaker is off and you measure a voltage?

Can you suspect Phantom voltage in other situations? How high can the Phantom voltage go? 120?

If the circuit is off, then what is inducing the voltage? Can it be a shared neutral?

Sorry about the multiple questions, I am curious about this phenomena..

You will see induced voltage any time you have a wire in a cable or raceway that is not connected to a load, power or ground at either end. The 3 way traveler is one example. When the switches are set so the load is on, the other traveler is not connected to anything but it is magnetically coupled to the current carrying traveler.

Even if other energized conductors are not actually carrying current to a load, you will still get a reading from the capacitive coupling. Two conductors running parallel separated by just their insulation form a very effective capacitor. The longer the cable run, the higher the capacitance and the higher the voltage reading.

You'll see the same capacitive effect if you use an older-style analog megger for insulation testing. When you press the button, the needle will "kick" over to the right, then gradually drop back to the left. The longer the cable, the greater the "kick" and the longer it will take to drop back to its steady value. The D.C. you are applying to test is charging the capacitance of the cable.

A number of years ago, I came up with the expression, "All the world is a voltage divider" I think it started with explaining the voltages to ground on battery strings measured with Fluke 77s Vs Simpson 260s. To gain greater understanding, first grab your meter manual and check the specs for the Mohms and pF that reflect it's input impedance. The cut sheets don't usually show these specs but they should be in the manual's expanded specs. Next, draw a circle with a sine wave in it that represents an AC voltage source. Place an "N" and the ground symbol below the source and write 120 VAC or Vrms beside it. To the right, draw a vertical resistor in parallel with a vertical capacitor. Add a ground symbol below them, the values from the manual, and the "M" in a circle meter symbol pointing across them. These things represent what you know. Next, draw a line up and then to the right of the source to a rectangle, then from the right side of the rectangle, over and down to the RC combo. In the rectangle, write inductive coupling, capacitive coupling, and resistive losses??? These are the values that you don't know, yet measure the effects of on your meter.

Now, I guess you decided to perform the death defying 34 volt to ground touch. Yet, you sought out the tingle without the enlightenment. Had you had a meter probe tip in each hand while touching the wire and ground, you would have gained another useful measurement. Had you then switched to ohms and touched the tips with the same pressure, you would know about the amount of parallel resistance that you added to the RC mix.

In a simple, purely resistive world, we could see that we dropped about 1/4 of the voltage with our meter loading. We would assume that if our meter was 10Mohms, that the other 3/4 of the 120 volts was getting dropped over about 30Mohms of lumped effects. But in the real world of AC circuits and reactive components, the voltages across all of the series components will likely add up to more than the total due to phase shifts involved.

I never really liked the term "phantom voltage" because thinking it isn't really there could turn you into a ghost! So could someone killed by induced voltage, be accused of over reacting???
Joe

We ran into this a lot when IBM gave everyone a Fluke 8060. Suddenly they had changed all those "zero" readings on floating wires to what I called a random number generator.
My "support" calls spiked.
We also quickly found out the classic difference between "precision" and "accuracy" when guys started comparing their road worn flukes on the same source. After a while it became apparent calibration was a must if you really wanted to believe all the numbers that fluke gave you. Most had an inaccurate reading, "precise" out to 3 decimal places.
A lot of guys ordered the Triplett analog meter that was much smaller than the Fluke and as accurate for most real world measurements

I think that the proliferation of digital meters with their umpteen-digit presentation has resulted in lot of people losing track of the difference between precision and accuracy. I've seen some real cheap DMMs (made you-know-where) in which the accuracy on A.C. volts is something like 4% plus 2 counts -- Worse than a good quality analog meter with a conventional d'Arsonval movement -- at least at the upper end of the scale.

I still have a Simpson 260 which I like to use, along with an old Avo 8, not to mention several lesser meters. I also have an old BPL (British Physical Labs) SuperRanger, which with a basic sensitivity of 1000 ohms-per-volt is much less susceptible to "phantom" readings on power work.

You guys are awesome. It's great to have this kind of resource of wisdom and experience. Thanks.

I learned about this the hard way a few years ago. I was also troubleshooting a 3-way situation and was using a voltage sensor to figure out what was live and what wasn't. The sensor, however, had a sensitivity adjusting dial on it which must have been set a little too high. It was indicating live wires which had no way of being live. I know this because I had disconnected the wire at both ends of the pipe run and verified that it was the same wire at both ends.
Mystified at this point, I got out my DMM and took a "real" measurement which was around...30-some volts. It seemed to me that the voltage pretty much had to be a result of induction from the other conductors in the pipe.
This phenomenon can really leave a guy scratching his head for a while until he figures it out.

I agree with wiking, it really is great to have access to the kind of wisdom available here.


Shawn.

I find this effect especially prevalant in old, cloth wrapped wire (circa '50s?). Used to drive me nuts when I was starting out. An inductive "sniffer" even lights up on an unconnected conductor at times.

The reason I gave up my Vol-Con Elite was because of false-positives on de-energized wires. Same situation as a lot of the ones here: A 277 volt three-way I was trouble-shooting; wires run in pipe. Kept getting strong 120V readings on some of the wires. All the wires are the same color so I'm cussing the original installer thinking I have two different voltages in the same box with the same wire colors. Can't even remember how I finally figured out I was reading induced voltage on de-energized travelers, I just know it took me a long while to do it. I agree. Went to T/S a light a helper was working on. We're getting approximately 120V on a neutral wire. Being the studious guy, he suggests "phantom voltage", and I was inclined to agree. The neutral was in a conduit about 150' long with about fifteen other conductors. Normally ghost voltage is nothing, but in that case it would've been more than enough to knock someone on their butt. Even with two wiggys on the wire, there wasn't enough of a load to drop the reading below 100V!

-John