Abnormal voltages to ground

Status
Not open for further replies.

jlcman

New member
We were called to an office today to inspect several outlets. The customer complained that when they connected their computers with surge protectors to the outlets that several of the surge strips "popped" and burned up. When measuring the voltage to neutral at all outlets, the readings were normal 120volt, but the voltages to ground were 112 at some outlets and 136 at others. When we checked the voltages at the panel the voltages were all normal.

The circuits are multiwire and installed in EMT conduit to a metal junction box in the attic with a ground wire. After the first junction box, someone installed ENT conduit between the junction box and another section of EMT conduit which disappears into an inaccessible space. The wiring from this junction box going into the ENT conduit does not contain a ground wire. When the wiring reaches the plugs in the office it is type MC cable in metal boxes with bonding screws.

Does anyone have an idea of what the problem is. My first thought is that someone probably connected the ground and neutral wires in some junction box that we can't find and that wherever this connection is, it has come loose...The only reason I second guess this is that I'd expect to have abnormal voltages in the L-N readings at the plugs somewhere if this were true instead of in the ground wire. My other thought was that with the break in the conduit with the small section of ENT and no apparent ground wire, that somewhere there is a fault that is going onto the metal conduits and MC and that because the conduit does not have a solid path it is traveling back to the ground terminal on the screws of the plugs and giving me these readings...the building is wood framing instead of iron and sheetmetal, so the conduits are essentially isolated from a path to ground with the break in the conduit. This break could be causing the lack of the potential and thus not allowing the breaker to trip even though there is a fault.

Any help or ideas would be appreciated. We are going back in the morning and were going to try and turn off the breakers individually while testing the outlets to see if we can find out which circuit is causing the readings and then attempting to track down the problem.....
 
Sounds a lot like an open or high resistance connection in the neutral feeding those receptacles.
The ground (EGC) path sounds solid, and there is no good reason for the ground to be offset one way or the other from neutral if only the ground is open.
With an open neutral, putting a line to neutral load on one phase will offset the neutral to the point where it will blow line to neutral surge protectors on the other phase.
To my mind that is the first thing an experienced electrician would check for on a service call given those symptoms.
If that has been ruled out, the next likely problem would be two simultaneous faults: an open ground and a short or high leakage from one of the lines to ground, possibly in connected equipment.
Given that line to neutral has been checked (under load???) to be 120 at each receptacle the double fault becomes more likely.
A power strip probably has voltage limiters from phase to ground as well as or instead of phase to neutral.

Sent from my XT1585 using Tapatalk
 
Last edited:
See the last half of my analysis or provide an alternate explanation for consistent 120 line to neutral. :)

Sent from my XT1585 using Tapatalk

Poor use of terminology. High voltage on one side low voltage on the other and stuff blowing up screams bad neutral no matter what the claimed voltage reference is. Bad grounded conductor lets the smoke out, bad grounding doesn't.
 
Poor use of terminology. High voltage on one side low voltage on the other and stuff blowing up screams bad neutral no matter what the claimed voltage reference is. Bad grounded conductor lets the smoke out, bad grounding doesn't.
Yes and no.
Bad neutral lets smoke out of the equipment and some surge protectors but bad EGC combined with another fault lets smoke out of the surge protectors only and can shock people, where a bad neutral would not.

Sent from my XT1585 using Tapatalk
 
Yes and no.
Bad neutral lets smoke out of the equipment and some surge protectors but bad EGC combined with another fault lets smoke out of the surge protectors only and can shock people, where a bad neutral would not.

Sent from my XT1585 using Tapatalk

Not saying you are wrong and more importantly not saying that I'm not wrong but if what you are thinking can happen turns out to be what really did happen it would be the first time I have ever heard of it happening.
 
170103-2133 EST

jlcman:

We were called to an office today to inspect several outlets. The customer complained that when they connected their computers with surge protectors to the outlets that several of the surge strips "popped" and burned up. When measuring the voltage to neutral at all outlets, the readings were normal 120volt, but the voltages to ground were 112 at some outlets and 136 at others. When we checked the voltages at the panel the voltages were all normal.

I believe you are using the word ground to mean the EGC at an outlet.

Since you really don't know what caused the surge protectors to fail, and you can not find something obvious I would do the following.

Assume at the main panel that neutral, EGC, and earthing ground are well bonded together.

1. Unplug everything on the problem circuits.

2. Use a single wire or an extension cord to provide a test lead from the EGC bus bar at the main panel.

3. Get a 1500 W space heater to use as a test load. These often are in the 10 to 12 ohms range when heated, and somewhat lower when not heated.

4. With no loads on the circuits, and using a high impedance meter like a Fluke 27 or 87 measure the voltages from each phase (hot wire) to the EGC at the main panel. Also from neutral to main EGC, and local outlet EGC to main EGC. This provides you with some references and you know what these should be.

5. Now we do some loading with the heater. The load is in the ballpark of 10 A at 120 V.

6. At an outlet where a problem (damage) has occurred connect the heater load between phase A and its neutral. Redo the measurements of (4). Do also from phase B hot to neutral, a different outlet. This may immediately point you in the direction of the problem.

Report back.

.
 
Not saying you are wrong and more importantly not saying that I'm not wrong but if what you are thinking can happen turns out to be what really did happen it would be the first time I have ever heard of it happening.

Engineers when making not making the world a better place they fill there time making everyday simple repairs complicated. :D

GoldDigger, Gar. Have you never heard of K.I.S.S. (The acronym, not the band):)

Seriously guys like Dave and myself would not get anything done if we started troubleshooting for the unusual first. We are talking basic building wiring here, not MV distribution, not computerized systems or anything remotely complicated.

I sometimes I think engineers don't realize some of us have been doing these types of calls daily for decade now, not just imagining them. :cool:
 
Engineers when making not making the world a better place they fill there time making everyday simple repairs complicated. :D

GoldDigger, Gar. Have you never heard of K.I.S.S. (The acronym, not the band):)

Seriously guys like Dave and myself would not get anything done if we started troubleshooting for the unusual first. We are talking basic building wiring here, not MV distribution, not computerized systems or anything remotely complicated.

I sometimes I think engineers don't realize some of us have been doing these types of calls daily for decade now, not just imagining them. :cool:

Note that the first thing I said after describing a neutral problem was that I expected that a competent, experienced service electrician would alreay have checked for that, and went from there to look at other possibilities. I may have been overestimating the people working for the OP, but since he sounced desperate I wanted to go on to the less likely possibilities.

Eagerly awaiting the final result. :)
 
Eagerly awaiting the final result. :)

FWIW I meant both parts sincerely, making the world a better place and the complications. :)


There was an engineer that used to participate his level of knowledge about transformer design, construction, theory and use was outstanding. Very helpful, nice guy.

He would also have no problem writing a 100 pages on how to sharpen a pencil in such detail that your eyes would bleed. :D
 
We were called to an office today to inspect several outlets. The customer complained that when they connected their computers with surge protectors to the outlets that several of the surge strips "popped" and burned up. When measuring the voltage to neutral at all outlets, the readings were normal 120volt, but the voltages to ground were 112 at some outlets and 136 at others. When we checked the voltages at the panel the voltages were all normal.

The circuits are multiwire and installed in EMT conduit to a metal junction box in the attic with a ground wire. After the first junction box, someone installed ENT conduit between the junction box and another section of EMT conduit which disappears into an inaccessible space. The wiring from this junction box going into the ENT conduit does not contain a ground wire. When the wiring reaches the plugs in the office it is type MC cable in metal boxes with bonding screws.

Does anyone have an idea of what the problem is. My first thought is that someone probably connected the ground and neutral wires in some junction box that we can't find and that wherever this connection is, it has come loose...The only reason I second guess this is that I'd expect to have abnormal voltages in the L-N readings at the plugs somewhere if this were true instead of in the ground wire. My other thought was that with the break in the conduit with the small section of ENT and no apparent ground wire, that somewhere there is a fault that is going onto the metal conduits and MC and that because the conduit does not have a solid path it is traveling back to the ground terminal on the screws of the plugs and giving me these readings...the building is wood framing instead of iron and sheetmetal, so the conduits are essentially isolated from a path to ground with the break in the conduit. This break could be causing the lack of the potential and thus not allowing the breaker to trip even though there is a fault.

Any help or ideas would be appreciated. We are going back in the morning and were going to try and turn off the breakers individually while testing the outlets to see if we can find out which circuit is causing the readings and then attempting to track down the problem.....

wag: neutral and grounds are mixed up somewhere, and you have an open ground which was serving as the neutral. Did you check for continuity between neutral and ground? ALL panel voltages (L-L, L-N, L-G x3ea) are good?
 
FWIW I meant both parts sincerely, making the world a better place and the complications. :)


There was an engineer that used to participate his level of knowledge about transformer design, construction, theory and use was outstanding. Very helpful, nice guy.

He would also have no problem writing a 100 pages on how to sharpen a pencil in such detail that your eyes would bleed. :D
Sounds like some of the consultants I've come across.............:D
 
wag: neutral and grounds are mixed up somewhere, and you have an open ground which was serving as the neutral. Did you check for continuity between neutral and ground? ALL panel voltages (L-L, L-N, L-G x3ea) are good?
On the right track here, I think.

Is possible there is a low resistance path back to the source on the grounded conductor at the outlets in question.

Equipment ground sounds like possibly interconnected with a current carrying conductor somewhere - and has open circuit issues back to the source. Somewhere the equipment connected to this that is putting current on that (intermixed neutral) line has to have voltage issues.

Maybe a circuit tracer connected from ungrounded to EGC at the outlets in question will help you find the "L2" side of the circuit where this "open neutral" condition originates back at the panel? Keep in mind it may be a circuit that was not suspected at all but does cross paths in a junction box somewhere.
 
170104-0925 EST

iwire:

I have no problem with the KISS approach. That was tried by the OP in post 1.

Guessing at where the problem is and then by trial and error opening outlets and junction boxes all over the place probably does not make sense. A more direct approach that uses measurements to point you in the correct direct probably can save a lot of time.

Outlet strips probably contain MOVs that TRY to prevent excessive over voltage. You have to understand the characteristics of MOVs and what your loads can tolerate to have a feeling of what happens under different fault conditions. Other devices than MOVs can be used.

To have outlet strip MOVs fail means a large over voltage with a lot of source current occurred.

If you remove all loads on the affected circuits, apply source voltage, make voltage measurements in the affected area as a reference, then apply a single 120 V load, and compare the voltage readings the results may quickly point you in the direction to make other measurements. The load may have to be tried in different locations.

With a Fluke 27 or 87 with the capability to resolve 0.1 V at 120 V, and a 10 A load I can tell which of two outlets is closer to the main panel even when #12 copper is used between the outlets.

The 27 is better because it is a 3.5 digit display and it does not jump around as much from normal line fluctuations.

#12 copper is about 1.5 ohms per 1000 ft. Two outlets 10 ft apart have a #12 current path of 20 ft. Thus, resistance is 0.03 ohms. 10 A at 0.03 ohms is an 0.3 V drop. The outlet with an 0.3 higher voltage with this 10 A load is closer to the main panel. I would suggest an extension cord between the outlets so you can quickly switch the meter from one outlet to the other to minimize the effects of supply voltage changes.

Appropriately used voltage measurements can provide a lot of useful information. Picking an appropriate voltage reference point can be important.

.
 
Last edited:
When measuring the voltage to neutral at all outlets, the readings were normal 120volt, but the voltages to ground were 112 at some outlets and 136 at others. When we checked the voltages at the panel the voltages were all normal.

The circuits are multiwire and installed in EMT conduit to a metal junction box in the attic with a ground wire. After the first junction box, someone installed ENT conduit between the junction box and another section of EMT conduit which disappears into an inaccessible space. The wiring from this junction box going into the ENT conduit does not contain a ground wire.

Does anyone have an idea of what the problem is.

wag: neutral and grounds are mixed up somewhere, and you have an open ground which was serving as the neutral.

When finding things like ENT without a ground conductor I think I would check every junction box between the receptacles and the panel.

The symptoms do sound like at bad neutral but there may be bad connections in more than one box. And that ENT does need a ground conductor.
 
170104-1594 EDT

Continuing with my comments on voltage measurement.

To track the path of a circuit greater sensitivity is obtained by measuring the voltage drop along the neutral relative to the potential of the EGC bus at the main panel.

You can get the main panel EGC reference in two ways.

One is to run a wire or extension cord from the EGC bus at the main panel to wherever you want to make a voltage measurement relative to said main panel point.

The second way is to use the EGC wire at the measurement point. This second way requires that the EGC is intact back to the main panel, and that no current is flowing in the EGC. For this thread this is not a good assumption.

This approach of measuring the neutral voltage relative to the main panel EGC bus is only going to be seeing small voltages normally. Thus, I can possibly use a 100 W incandescent rather than a 1500 W heater as the load.

My following sample voltages are read with a Fluke 27 with no filtering. Ideally I should have a narrow band 60 Hz band pass filter at the meter input.

At my main panel the Fluke reads 0 or 1 mV between the Sq-D EGC bus and the neutral bus. At my work bench area, about 50 ft away, the reading is about 3 mV. Also about 3 mV between an extension cord EGC and the bench neutral.

With a 100 W lamp load at the bench area at the lamp outlet neutral to EGC is about
120 mV. Going to a point on the benches closer to the wall the reading is about 50 mV. There are three benches with cascading of power from one to another, not short cords. The lamp was located at the middle bench. I use #8 for 50 ft from the main panel to the bench area. Without making an actual measurement I estimate the #8 drop at about 25 mV.

.
 
Last edited:
For there to be a voltage drop (or rise) on that presumed EGC, there about has to be current somewhere on that conductor upstream, which likely means it is intermixed with neutral conductors or with what should be connected to a neutral conductor. The challenge may be finding out where.
 
For there to be a voltage drop (or rise) on that presumed EGC, there about has to be current somewhere on that conductor upstream, which likely means it is intermixed with neutral conductors or with what should be connected to a neutral conductor. The challenge may be finding out where.

Yes that can be a challenge? If it's a drop ceiling or open ceiling I just follow the conduit to the first junction box but since this is a wood building it may have hard ceilings and require some circuit traceing.

From experience I normally find burned wire caused by loose connections ( wire nuts ).

I hope the OP comes back and lets us know what he or his guys find.
 
Status
Not open for further replies.
Top