Stumped by ground/neutral voltage

[PetrosA]

I had a service call today at a farm for a problem in a remote structure (approx. 400' run from main disconnect through house and garage) fed by a 12-2 UF which feeds a skeet thrower and plug-in fan via a few receptacles. The customer reported getting shocked from the metal boxes and EMT run between them. Another electrician had checked it out and removed some questionable wiring in the structure, but hadn't been able to eliminate the voltage on the ground.

When I arrived, the homeowner was not around so I started out at the "bunker" where the skeet thrower is located. I drove about 24" of #2 CU for a ground reference and two meters were showing me about 10V between the UF ground and the reference ground, and about the same between the N and ref. G, with both showing 15V peaks. I then found the other end of the UF in a garage and started checking splices to see if I could find a bad neutral or N to G short. I found one GFI recept on the same circuit with a short between N and G and fixed that. I still had the 10V at the structure, so I checked at the main disconnect. There I found a loose N lug (SquareD 200A disco, lugs are screwed into a plate - crappy design!). Fixed that, still had voltage from N and G in the skeet bunker.

At this point I decided to check if the problem existed at the main disconnect as well (up to this point I was under the impression from what the customer had reported that the problem only affected the skeet bunker). I moved my reference G to the main disco and got some strange readings:

From N to ref. G - 1.4V-9.1V
From either leg to ref. G - approx. 5 volts higher reading than from legs to N; ex. 122V to N, 127V to ref. G

The higher voltage from G and N to ref. G were intermittent with low readings at the disco about 1.4V and in the skeet bunker about 2.9V, which I suspect is partly induced voltage due to the long run.

The customer arrived home about 3 hours into my being there. An interview with him revealed that in the past, they had experienced a similar problem in the basement with the wash sink and an electrician had checked it out, recrimped the neutral connections at the service head, checked the connections in the meter socket and declared a clean bill-of-health in the house, at which point that electrician called in the POCO, who the homeowner thinks installed a ground rod at the nearest pole where the transformer is mounted, checked N connections from transformer to the meter socket again and gave their blessing to the whole thing.

To me this smells like stray voltage on the neutral coming from the POCO's line which in my limited understanding would explain why I get higher voltage to the ref. G. Does that sound like a valid theory, or am I missing something important? I checked all the panels and ATS for a G to N bond, but it all looked fine. I don't have datalogging equipment so I couldn't check if there is a correlation between load and voltage from N to ref. G due to the intermittent aspect of the problem but the house was mostly empty, no central air and peak loads seemed to be limited to a refrigerator or two.

I'd be grateful for any ideas of what theories are likely and what to check to prove or disprove them.

Thanks!

 

[iwire]

120 volts 400 feet on 12/2 UF?

How much load is on the run?

 

[Dennis Alwon]

The first think I would do is turn everything off and see if the voltage is present. If not then turn on one circuit at a time until you find the culprit.

 

[dbuckley]

First step - power the entire installation off (meter out is best - no arguments then) and check N/G volts. Still there? How about at the remote structure still with the power off. There too?

 

[Dennis Alwon]

120 volts 400 feet on 12/2 UF?

How much load is on the run?

Heck with 5 amp loads you would have an 8 volt drop. 10 amp load is duh... 16 volts.

 

[kwired]

Very well could be stray current from a POCO primary neutral problem.

As others stated turn off main disconnect and check if voltage is still present. If it is, it is likely coming from outside of the premises wiring system. Open the neutral connection at the service and your voltage will likely go away but there will be voltage between the service neutral and the rest of your system.

Could be current from a neighbor that has a bad neutral also. Check out any conductive path between them and the place in question including metallic gas or water piping, shielded communications cables.

Make sure you do not have ungrounded equipment with good earth contact with a ground fault within the equipment somewhere on site - especially a water well.

Even if there is stray voltages what if any grounding electrodes are present at each building on the premises? Properly installed grounding elecrode systems will help minimize the shock potential of stray voltages because you have closer to same potential to earth as you have to equipment grounding conductors. Because of the way the utility system is grounded there is normally some current within the earth but when there is a high resistance in a usual path it is forced to flow in higher quantities in other places.

 

[PetrosA]

"How much load?" - I'm not sure. The machine that launches the clay pigeons is large, 6'x4'x4' and has a hydraulic pump and a second motor. I didn't want to turn it on and start launching a mess IIRC, I was getting about 118V to ref. G at the bunker with just a cooling fan turned on.

I will check for the voltage with the meter pulled next week when I go back, and I'll check for any issues with the well pump. I hadn't yet considered that it was a whole house issue when I had the power off today, and by the time I did, the homeowner was back and had company, so I didn't want to shut it all down again. I highly doubt it has anything to do with neighbors - judging by what I see in Google maps, it's about 3500 feet to the nearest one.

From your answers so far, it sounds pretty likely that this is a stray voltage issue. If so, would proving voltage with the meter pulled and potential between the disconnected POCO neutral and premises wiring be enough to force the POCO to install a neutral blocker (or whatever else they might cnsider appropriate) or would I need something else as well?

Edit to add: Regarding additional grounding electrodes - what would be the best way to add an electrode to separate structures which don't have panels in them? IOW, I'm dealing with a single circuit feeding the garage and through to the bunker, so the only point of attachment would be in the junction boxes in those structures. What size conductor from the rod to the box?

 

[gar]

100714-2054 EST

First, I would find where the main panel is grounded. Assume it is both a ground rod or other electrode and a conductive water line. This point then becomes my reference. You want to use a high input impedance meter, such as a Fluke 27 (10 megohms input). The system grounding reference goes to one meter terminal via a very long cord. The other meter lead goes to a probe to check voltage at various points. Probably have an insulated rod connected to the probe and the meter mounted on the handle away from where you hold the insulated shaft. An 8" screwdriver is an adequate probe. Do expect you will find some moderate voltages so protect yourself.

Make some test measurements on a few things in the house like sinks and EGC points in a few receptacles. I expect less than 1 volt, in fact much lower.

Next measure the EGC and the neutral at the skeet thrower, then the earth near the skeet thrower.

Turn the breaker off to the skeet thrower, and repeat the measurements.

Repeat with the main breaker off.

The next step is to move the probe to the ground rod at the pole where the transformer is located, and actually contact the rod with the probe. Still with the main breaker off the voltage should be very close to 0.

The pole ground rod is a very good test lead to the transformer center tap unless there is current flowing in this grounding wire or it is open.

Report your values.

.

 

[SG-1]

From either leg to ref. G - approx. 5 volts higher reading than from legs to N; ex. 122V to N, 127V to ref. G

The higher voltage from G and N to ref. G were intermittent with low readings at the disco about 1.4V and in the skeet bunker about 2.9V, which I suspect is partly induced voltage due to the long run.

The higher readings between the H-G tells you the available voltage to the load,as if you were measuring the voltage in the panelboard. It is normal for this voltage to be higer than the H-N voltage with load on the circuit. No currrent flows between the H-G in a properly wired system, therefore no volt drop.

The H-N voltage is lower than the H-G voltage when a load is running because current is flowing & causing I2R losses in the wire. If the load fluctuates so will the voltage drop. The difference between the two readings will show up on a reading between the N-G.

Good Hunting !

Here is a link to a Fluke article that explains this in more detail.

http://support.fluke.com/educators/Download/Asset/2100112_6003_ENG_B_W.PDF

 

[PetrosA]

The higher readings between the H-G tells you the available voltage to the load,as if you were measuring the voltage in the panelboard. It is normal for this voltage to be higer than the H-N voltage with load on the circuit. No currrent flows between the H-G in a properly wired system, therefore no volt drop.

The H-N voltage is lower than the H-G voltage when a load is running because current is flowing & causing I2R losses in the wire. If the load fluctuates so will the voltage drop. The difference between the two readings will show up on a reading between the N-G.

Good Hunting !

Here is a link to a Fluke article that explains this in more detail.

http://support.fluke.com/educators/Download/Asset/2100112_6003_ENG_B_W.PDF

Thanks for the article! I don't know if it's important, but let me clarify what I meant in that section you quoted. The measurements were taken at the main disconnect (outside, nippled to the meter) where the N is bonded to the GEC and approx. 4'-5' from the ground rods (there are two). My reference ground was 8'-10' from the system rods. Because of the bond at the disconnect, there is obviously no way to measure N-G voltage except to a reference ground unless I disconnect the rods from the neutral.

At the bunker receptacle, my N-G voltage was about 0.58V with no load and rose to about 0.9V with the fan running. These measurements seemed normal.

 

[PetrosA]

100714-2054 EST

First, I would find where the main panel is grounded. Assume it is both a ground rod or other electrode and a conductive water line. This point then becomes my reference. You want to use a high input impedance meter, such as a Fluke 27 (10 megohms input). The system grounding reference goes to one meter terminal via a very long cord. The other meter lead goes to a probe to check voltage at various points. Probably have an insulated rod connected to the probe and the meter mounted on the handle away from where you hold the insulated shaft. An 8" screwdriver is an adequate probe. Do expect you will find some moderate voltages so protect yourself.

Make some test measurements on a few things in the house like sinks and EGC points in a few receptacles. I expect less than 1 volt, in fact much lower.

Next measure the EGC and the neutral at the skeet thrower, then the earth near the skeet thrower.

Turn the breaker off to the skeet thrower, and repeat the measurements.

Repeat with the main breaker off.

The next step is to move the probe to the ground rod at the pole where the transformer is located, and actually contact the rod with the probe. Still with the main breaker off the voltage should be very close to 0.

The pole ground rod is a very good test lead to the transformer center tap unless there is current flowing in this grounding wire or it is open.

Report your values.

.

I'll have to check the specs on my meters to see if any of them have that kind of input impedance (I'm pretty sure the Greenlee doesn't, but the Agilent might - it states 10MOhm impedance for AC/DC volt ranges, but doesn't specify for the resistance range). Considering the distance to the bunker, what size wire should I use for the reference? I'll need about 400-600 feet from the system rod to reach the bunker over and around obstacles (this might be cost prohibitive anyway).

 

[mcclary's electrical]

Was the fan running when you checked? If so, you're probably seeing that due to votage drop accross the neutral.

 

[gar]

100715-0726 EST

PetrosA:

You could use #40 wire, but #22 would be convenient. With a 10 megohm input resistance meter the current is 1/10,000 A at 1000 V. Or 0.1 MA.

The input resistance (impedance) of a meter is a description of the loading a meter places on what is being measured in a voltage mode. Most DVMs, vacuum tube voltmeters, and oscilloscopes have a fixed input impedance for all voltage ranges. A meter such as a Simpson 260/270 has an input resistance that changes with the range. In DC it is 20,000 ohms/volt of full scale reading. The 1000 V range is 20 megohms, and the 2.5 V range is 50,000 ohms. In an AC range older meters were 1000 ohms/volt, and newer ones 5000 ohms/volt.

If the Agilent is fine with the spec you mentioned.

#18 extension cords or even #16 are useful as a long meter probe.

Consider another approach. With no known load at the "bunker" and at the "bunker" measure the voltage from neutral to EGC with main breaker off, then
1. Main breaker off, and bunker breaker off. Should read close to 0.
2. Main breaker on, and bunker breaker off. Should read close to 0.
3. Main breaker on, and bunker breaker on. Still should be close to 0.
4. At the bunker plug in a 1500 heater. Hot the heater should be about 10 ohms, and current about 12 A at 120 V. #12 wire is about 1.6 ohms/1000 ft at 20 deg C. 12 A at 400 ft would be a drop of about 0.4*1.6*12 = 7.7 V. Thus, your neutral to EGC reading would be about 8 V.

If you get these results, then the EGC is probably a good test lead from the bunker to the main panel. At least if you do not put a load on the circuit that has leakage to the EGC.

The test from the main panel to the ground rod at the transformer center tap will possibly indicate if there is high resistance in the neutral circuit between the transformer and the main panel. With no load on the main panel, main breaker open, this voltage should be near 0. With a 12 A heater test load and nothing else the difference should be 12 times the resistance of the neutral wire. You need length, size, and material of the neutral to estimate its resistance.

With the main breaker off and relative to the EGC-neutral point in the main make some measurements of ground potential about 50 ft away, maybe 100 ft, and where ever else this directs you.

If there are stray power company currents in the earth these measurements may help indicate such.

If it appears that there are power company currents, then use my two probes 12 ft apart method to try to determine current direction. Place two probes, 8 to 12 inch screwdrivers, spaced about 12 ft apart in the earth and measure the voltage between them. Use a consistent distance. In my yard readings range from maybe 20 millivolts to 200 millivolts. Also in my yard the direction and magnitude changes with time. Do this N-S (north south), NE-SW, E-W and see what results occur. If 12 ft was 100 millivolts, then this might extrapolate to 3.3 V at 400 ft assuming a distant source.

As you make some measurements the results may lead you in some specific direction.

.

 

[gar]

100716-1004 EST

If the tests in my previous post indicate very little drop on the main to bunker EGC, then either end of the EGC should be useful as a reference back to the main. This shortens your need for a long test lead to about 200 ft. For the voltages to earth you have been reading at the bunker the error from small differences along the EGC should not be significant.

Whether there is a load at the bunker or not the voltage along the EGC should not change unless there is a current flow from something to the EGC. Under normal conditions the EGC is a non-current carrying conductor.


The neutral from the transformer center tap to the main should have no voltage drop along its length with no load, and not much under load. 000 copper has a resistance of 0.062 ohms/1000 ft 20 deg C. 100 ft at 100 A would be 0.62 V.

From your first post:

From N to ref. G - 1.4V-9.1V
From either leg to ref. G - approx. 5 volts higher reading than from legs to N; ex. 122V to N, 127V to ref. G

If both legs to ref.G increase by 5 V, then this implies the ground voltage is shifted by 90 deg. With a 90 deg shift the ground voltage to neutral would need to be about 35 V for a 5 V change of 120 V. If the ground voltage was in phase with one leg, then it would 180 deg out of phase with the other. Thus, one voltage would rise and the other fall by the 5 V.

Knowing the integrity of your EGCs and the neutral from main to transformer helps make sense of other measurements. Determining the integrity of these is one purpose of some of the measurements.

Whenever you perform a voltage measurement between two points you have created a 1 turn loop. If there is a strong AC magnetic field cutting this loop, then an AC error voltage will be inductively coupled to the loop and produce an error voltage added to the desired signal you are trying to measure. Whether this is a problem or not you can determine by creating a similar loop of approximately the same cross sectional area and see what is the induced voltage.

.

 

[SG-1]

"
Edit to add: Regarding additional grounding electrodes - what would be the best way to add an electrode to separate structures which don't have panels in them? IOW, I'm dealing with a single circuit feeding the garage and through to the bunker, so the only point of attachment would be in the junction boxes in those structures. What size conductor from the rod to the box?

250.66(A) under the 2008 code states that a 6 AWG copper conductor is sufficient.

One end of the 6 guage is connected to the ground rod. The other end is connected to the EGC in the junction box ?

 

[hurk27]

250.66(A) under the 2008 code states that a 6 AWG copper conductor is sufficient.

One end of the 6 guage is connected to the ground rod. The other end is connected to the EGC in the junction box ?

Ground rods in a case like this will do nothing to stop a difference of potential between the grounding and Earth, if the last point of neutral to ground bond has voltage drop on the neutral ahead of it.

The higher voltage from G and N to ref. G were intermittent with low readings at the disco about 1.4V and in the skeet bunker about 2.9V, which I suspect is partly induced voltage due to the long run.

This statement tells me you have a intermittent neutral to ground connection some where at the load end, and could be in the wiring of the Skeet machine itself, this is because of the voltage drop on the neutral of the 12/2.

I carry a 1600 watt ceramic heater to load circuits down for testing, unplug the skeet shooter and hook up a substantial load and see if you have any voltage on the grounding, vibrate the box's and conduit to see if it comes and goes, if it does, look for a grounded neutral problem.

You can do this un powered by disconnecting the 12/2 UF feeding this bunker, and placing an audible continuity type tester between the grounding and neutral (on the bunker side of the circuit), then wiggle receptacles and boxes and see which one is causing it to sound, also try wiggling the wiring on the skeet machine, more then likely the skeet machine is vibrating a neutral wire causing it to make a connection to the EGC, this EGC could be a wire, Box, or conduit or even the frame of the skeet machine.

another test to prove this is at the load end, is to place a substantial load hot and neutral on the 12/2 UF feeding the building, with nothing else connected to the grounding, see if there is a voltage from the EGC of the 12/2 to Earth, if there is, the problem is ahead of this point, if not, see above.

 

[gar]

100718-0853 EST


There is something inconsistent with the following:

From N to ref. G - 1.4V-9.1V
From either leg to ref. G - approx. 5 volts higher reading than from legs to N; ex. 122V to N, 127V to ref. G

The higher voltage from G and N to ref. G were intermittent with low readings at the disco about 1.4V and in the skeet bunker about 2.9V, which I suspect is partly induced voltage due to the long run.

In the first paragraph, as I previously described, if you have sine waves of the same frequency and synchronized, then you can not have both legs rising by 5 V in 120 V without the error voltage being about 90 deg shifted from either of the 120 V signals. To make the hypotenuse of a right triangle equal 125 when one leg is 120 requires the other leg to be 35. When you measure from neutral to earth you get at most 9.1 V. This is a long way from 35.

Even if the two voltages were totally uncorrelated the results are the same.

I suggest two small, like 15 W, tungsten incandescent bulbs be connected in series across the two hot legs with the center point between the bulbs connected to your reference ground at the main panel. If the voltages are equal both bulbs will be equally bright. If one is less bright than the other, then the bulbs will have different intensities. Now you can intermittently short your reference ground to neutral and observe any changes in intensity. I can see an intensity change with about a 2V change at 120. The bulbs need to be adjacent to each other for the comparison.

In your second paragraph the word induced is used. I doubt that you have any substantial induced voltage. Rather you have an I^2*R (or Z) voltage drop thru the earth or some other conductor.

Experimentally I created a 1 turn loop on my living room floor. About 2 ft x 10 ft. Using a Fluke 27 the reading from any residual magnetic field was 0.0 millivolts. Then I placed a single wire adjacent to one long side of the loop, and outside the loop. Nothing was precise. With AC 12 A in the single wire the induced voltage read was about 2.3 millivolts.

To get to about 1 V the loop length needs to increase by a factor of 430 or increase to 4300 ft from 10 ft. So in my opinion you do not have an induced voltage problem. Instead of length you could increase the magnetic field intensity, but that means a lot of current. If your coil is 400 ft then the current would be in the 100 A range. If there is a balancing current (second conductor with the return current), then even higher current levels.

.

 

[kwired]

Ground rods in a case like this will do nothing to stop a difference of potential between the grounding and Earth, if the last point of neutral to ground bond has voltage drop on the neutral ahead of it.

Why are we required to bond conductive objects together in many cases - especially around swimming pools, hot tubs, and the like as well as provide equipotential grids?

I'm not saying it is a solution to the unknown current problem, but it does reduce the shock hazard risk by bringing voltage potential between earth in the vicinity of the grounding electrode and equipment grounding conductors and equipment attached to them to a closer level.

 

[hurk27]

Why are we required to bond conductive objects together in many cases - especially around swimming pools, hot tubs, and the like as well as provide exponential grids?

I'm not saying it is a solution to the unknown current problem, but it does reduce the shock hazard risk by bringing voltage potential between earth in the vicinity of the grounding electrode and equipment grounding conductors and equipment attached to them to a closer level.

A grounding grid, will encompass the entire area, bringing you and the voltage to the same plane, a ground rod will not, you can put a 120 volt lead to a ground rod and 3' away have 90 volts to earth, quite a shock hazard within reach, so there is a great difference between equal potential bonding and ground rods, and even though you can have a EPBG, you can still have a potential at the edge of such a grid, or equipment at the edge of it.

Check out this thread:
Earth Shells Experiment = time to eat crow!

Keep in mind the test voltage was 50 volts, and his soil resistivity was about 8 ohms which produced a much smaller SOI, so at 2' he had the 90 volts.

 

[PetrosA]

I just came back from a few days break and will go over all the information you've given me to come up with a list of things to do on my return visit to the customer (planned for Thurs.). Thank you!