Trouble shooting objectionable voltage on grounding wire

[greenjeans]

Hi all,
I'm working in a home that has 53-110 volts on the grounding wire. This was discovered by the cable guy's, apparently they have changed out the modems and cable box's three times(Internet/TV dropping out) I started t/s the problem and found, a multi-wire feed the 2nd floor.
Problem was, as I opened boxes and light fixtures and checked for junction boxes in attic, to find where the neutral and grounding wire were joined or sliced together, the voltages started climbing, 110V. I opened every switch and ceiling fixtures(no recessed yet) hoping to find this splice. I know it the multi wire circuit, cause it drops to 0-1v when turned off.
Does anyone have an idea that would help t/s this better or If this has been posted before i would like to get that post please, thanks you for all your ideas and comments.

 

[hurk27]

I sounds like somwhere there is neutral current on this grounding conductor, and this grounding conductor is not bonded back to the main panel neutral bar, look for receptacles with a boot leg jumper between the neutral and ground on the receptacles, also look for any load that might be using the grounding conductor as a neutral.

A grounding conductor can not have voltage on it if it is connected to the neutral at the service panel.

All I can say is try turning off loads on those circuits and see which one or ones are causing this voltage then back trace from there.

 

[gar]

110326-2056 EDT

greenjeans:

I do not follow your description. So some assumptions to start.

Assumptions:
1. All EGC (Equipment Grounding Conductors) are securely connected to the EGC bus bar in the main panel.
2. EGC bus bar in the main panel is securely connected to the neutral bus bar.
3. the neutral bus is properly terminated to the required grounding electrode(s).

Connect a 3 prong receptacle to its own circuit in the main panel. Get a long 3 conductor extension cord. Plug into this receptacle. This provides you with reference points within the panel to which to make relative measurements. If I do not intend to put a load on the extension cord, then #18 is fine. Otherwise use an appropriate wire size cord. Also get a small 1500 W heater, this is about 10 ohms and therefore about 12 A at 120 V. Use a good DVM that can resolve 0.1V at 120 V.

Go to a location where you have excess voltage on the EGC. For the first tests use the extension cord EGC as your meter reference point. You need to know that #12 copper wire has a resistance of about 1.59 ohms per 1000 ft. At 10 A thru 10 ft of #12 copper the voltage drop will be about 0.159 V. For #14 and 10 ft it is about 0.253 V.

If you had a 50 ft run for an EGC, then at 20 A your drop might be 1 V ball park range. Thus, I believe you have open circuit in your problem EGC or a high resistance joint some where in the EGC

You first need to find the open or semi-open point. Pick a bad location and connect your meter on a range for 120 V so one meter lead goes to the extension cord EGC, the other meter lead goes to the bad EGC point. This should read many volts based on your original post. Next check all the EGCs on all the other outlets on this circuit. Are any of them near zero or less than maybe 2 V? If any are OK, then the high resistance is between the OK and not OK outlets. Find and fix this problem point. Then check for any others on the same circuit.

Note: you really want to see 0 voltage drop along an EGC conductor. But in some cases there may be some noise voltage that appears, but of high frequency and low energy. My TED monitoring system is such a culprit because of its carrier current signalling.

After the EGC is made good, then make sure that a single breaker de-energizes this circuit. That means all breakers are on except for the one that drops this circuit.

Turn the breaker for this circuit back on, and go to the outlet on the far end of this circuit. Use the extension cord EGC for the meter reference. Use the 1500 W heater to be a switched load from the outlet being tested. Separately measure the change in voltage of both the hot and neutral relative to the extension EGC. If this circuit is 50 ft of #14 you should see about 0.25 V of change in either the hot or neutral for the wire alone. But there are other items normally in this path that may and probably do result in a larger drop, fuses and/or breakers.

At my main panel the voltage between neutral and EGC with no load on the outlet is less than 0.001 V. Adding a 12 A load this increases to about 0.37 V. This is more than I expect and I do not want to open the panel tonight to see if screws need tightening. You can do your own test in panels to see what to expect. On the hot side you will see more change resulting from fuses and/or breakers in the panel and from service supply side impedance.

You will have to develop judgement on how much voltage change should be normal and what might imply loose connections in the circuit.

I might guess between 3 and 10 V for a 12 A load should be considered a potential problem indicator. But you need to guess at wire length and know what is occurring within the panel and what the main change looks like at entry to the main panel.

See if this might help. Basically looking for points in a circuit with abnormal vs normal voltage drop can help you isolate a location.

.

 

[gar]

110326-2352 EDT

I made a calculation error in one spot, so see if I have missed any others.

Turn the breaker for this circuit back on, and go to the outlet on the far end of this circuit. Use the extension cord EGC for the meter reference. Use the 1500 W heater to be a switched load from the outlet being tested. Separately measure the change in voltage of both the hot and neutral relative to the extension EGC. If this circuit is 50 ft of #14 you should see about 0.25 V of change in either the hot or neutral for the wire alone. But there are other items normally in this path that may and probably do result in a larger drop, fuses and/or breakers.

The 50 ft of #14 at 12 A should be about 5 times the 0.25 or about 1.25 V, actually multiply by another 1.2 for 12 vs 10 A to get 1.5 V expected change.

.

 

[greenjeans]

Gar,

Yes all the EGC are connected, as they should be, and H20 and GRD rods are wired and installed correctly.
At first i thought it may be an open neutral(utility side to pole), but its only happens on the multi-circuit, and yes at first i didn't realize this, cause i didn't open the panel cover..........DAAAA. Always better to start in the front of the circuit, instead of the middle.

So using an extension cord18/2 (100') test from the cord-EGC and the outlet on far side of house. Question: do i take the outlet apart and separate the EGC's and test, OR test to outlet ground as its hooked up to duplex. I'm a bit confused, I've open all outlets on the circuit, and thought i would have opened the 14-2 that would be bad(EGC+NEUTRAL touching/tied together) and problem would stop, i guess its still present, cause i haven't found it yet. Will be opening all recessed lighting and ex fans (switch legs) next week. Hope i not to confusing, i do appreciate all the advice, Joe

 

[gar]

110327-1419 EDT

greenjeans:

To make it clear on how to use the extension cord. Its purpose is to provide you long test leads into the main panel. At the moment the most important one of these leads is the one that attaches to the EGC in the main panel. You might use an existing outlet as the source of the connection. But with the problems you describe I would like you to wire this reference point outlet directly into the panel close to the panel. Then you can be fairly sure you are connected to panel points without any other current flowing thru the leads.

My point about 20 A and the expected drop in a #14 copper wire of 50 ft was to point out that you could not have a large voltage on the EGC wire without an open or high resistance connection some place in the EGC circuit where the problem is located.

The problem you are describing can not be caused by an open neutral at the power company transformer. The reason is that all EGCs and neutrals are bonded together in the main panel.

The use of the neutral-EGC-ground reference in the main panel as one measurement point is that this is the point in the system where all three potentials must be the same, or not too many millivolts apart.

When you measure the voltage on an EGC relative to the main panel neutral-EGC-ground point and that voltage is excessive, then that measured EGC point is beyond the location of the high resistance or break in the EGC circuit being tested.

.

 

[gar]

110327-1516 EDT

greenjeans:

To do the voltage tests you do not need to open any boxes. Just put your test lead into the EGC hole and make contact. Or you probably should make a plug with a wire attached to its EGC and at the other end have a banana plug to go into your meter for measuring the EGC voltage.

After you find the points between which the open exists, then you start opening things.

.

 

[greenjeans]

Gar:

Thanks for the clarity, going back in a week, that should help allot. Gar, do think I'm looking for an open? I do not, I think its stray voltage dew to a neural and EGC touching each other, but I've been wrong before. Any way thanks again and I'll let you know the out come.

 

[gar]

110327-1912 EDT

greenjeans:

It can not be stray voltage, if by this you mean capacitively voltage to a floating wire, unless the EGC is not connected to the EGC bus in the main panel. If it is not connected, then there is an open or high resistance location someplace.

You need to get well paid for your time to straighten someone's old mess.

.

 

[gar]

110327-2225 EDT

greenjean:

My previous post has a missing word. This happens to me a lot, or words get doubled, or messed up some other way.

"capacitively voltage" should have been "capacitively coupled voltage".

What this means is the voltage measured on the floating wire resulted from capacitive coupling from something else with a voltage on it. If you have a piece of Romex, say 50 ft long, ground the neutral wire, connect the hot wire to 120 volts, and leave the EGC wire float, then using ground as a reference and measure with a high impedance meter, 120 megohm input impedance, the voltage anywhere along any of the three wires you will read essentially 0 on the neutral, 120 on the hot, and something in the range of 60 on the EGC. Next connect the EGC to ground and it will read essentially the same as the neutral.

Next run a 12 A load thru and neutral. The voltage drop along either the hot or neutral will be about 1.5 V for #14 copper wire.

If there was no insulation on the EGC and neutral wires and you measured the voltage difference between these two as you move from the grounding point to the load end you would see a linearly increasing voltage difference. At half the distance it would be 0.75 V. Note there is no voltage change along the EGC because no current flows thru it, but there is a linearly increasing voltage along the neutral because of the current thru it.

.