Megging Question

[Little Bill]

I had a service call today for flickering/dim lights and some receptacles not working. It was in a barn about 200' from the house main.

First I thought he had a loose or open neutral, but the voltage measurements didn't show that. I only showed around 190V to the sub panel at the barn but there wasn't any high voltage on one of the legs.
I had around 72V on one leg-neutral and around 119V -neutral on the other.

So I decided to meg the conductors. This is where I'm confused.

Readings from: Line A to neutral - 11G
Line A to ground - 11G
Same with Line B to neutral/ground

From line to line the reading was 0.0 M

I did determine from the voltage readings that the A leg conductor was bad. Probably underground somewhere since this was an underground feed. I just can't figure out why the reading was the same on both legs to neutral/ground.
I would have thought the bad conductor would have had a different reading.

I'm guessing the low reading on line to line is due to insulation damage along with damage to the wire its self. But why wouldn't it show on line to neutral/ground testing?

 

[GoldDigger]

I had a service call today for flickering/dim lights and some receptacles not working. It was in a barn about 200' from the house main.

First I thought he had a loose or open neutral, but the voltage measurements didn't show that. I only showed around 190V to the sub panel at the barn but there wasn't any high voltage on one of the legs.
I had around 72V on one leg-neutral and around 119V -neutral on the other.

So I decided to meg the conductors. This is where I'm confused.

Readings from: Line A to neutral - 11G
Line A to ground - 11G
Same with Line B to neutral/ground

From line to line the reading was 0.0 M

I did determine from the voltage readings that the A leg conductor was bad. Probably underground somewhere since this was an underground feed. I just can't figure out why the reading was the same on both legs to neutral/ground.
I would have thought the bad conductor would have had a different reading.

I'm guessing the low reading on line to line is due to insulation damage along with damage to the wire its self. But why wouldn't it show on line to neutral/ground testing?

1. Did you confirm that the 72 volt was not just a capacitive or inductive phantom voltage from the other phase?
2. Was there load on the feeder when you measured the 72 volts? If so, then you may have just a high resistance at a near break in one wire. That will not show any resistance to ground and will be too small a series resistance to show up on a megger. It might show on a common ohmmeter.
3. With zero ohms between A and B, either there is a load, a transformer, or some other component bridging between the two or there is a fault connecting the two wires with a low enough impedance that it is lost next to zero on your megger. Or maybe the shunt resistance at a fault in the line goes to zero under the high voltage of the megger but opens at lower voltages.
4. Given that A and B are connected, it makes perfect sense for their resistances to ground to be equal.
5. The one thing you can be sure of is that there is no ground fault!

Supplemental: When you have a fault that shows up at normal line voltages like 120 volts, a megger may not be the ideal tool to track it down.

 

[Little Bill]

1. Did you confirm that the 72 volt was not just a capacitive or inductive phantom voltage from the other phase?
2. Was there load on the feeder when you measured the 72 volts? If so, then you may have just a high resistance at a near break in one wire. That will not show any resistance to ground and will be too small a series resistance to show up on a megger. It might show on a common ohmmeter.
3. With zero ohms between A and B, either there is a load, a transformer, or some other component bridging between the two or there is a fault connecting the two wires with a low enough impedance that it is lost next to zero on your megger. Or maybe the shunt resistance at a fault in the line goes to zero under the high voltage of the megger but opens at lower voltages.
4. Given that A and B are connected, it makes perfect sense for their resistances to ground to be equal.
5. The one thing you can be sure of is that there is no ground fault!

Supplemental: When you have a fault that shows up at normal line voltages like 120 volts, a megger may not be the ideal tool to track it down.

Yes there was some load when I checked the voltage, just a few lights with CFL bulbs.

All the megging was done with all the conductors, on both ends, disconnected.

I also checked with an ohm meter and didn't show any continuity or low resistance.

Since there was only one 240V load in the barn I disconnected that and put all the 120V loads on the working leg. Not much of a load to worry about, only lights and a small water tank heater to keep the water troughs from freezing.

Edit to add: I'm not (at this point) going to try and locate the fault. It's all underground and part of it runs under a paved driveway. I'm just going to run a new feed to the barn.

 

[K8MHZ]

Yes there was some load when I checked the voltage, just a few lights with CFL bulbs.

All the megging was done with all the conductors, on both ends, disconnected.

I also checked with an ohm meter and didn't show any continuity or low resistance.

Since there was only one 240V load in the barn I disconnected that and put all the 120V loads on the working leg. Not much of a load to worry about, only lights and a small water tank heater to keep the water troughs from freezing.

Edit to add: I'm not (at this point) going to try and locate the fault. It's all underground and part of it runs under a paved driveway. I'm just going to run a new feed to the barn.

You mention neutral/ground as if they were one. Is the sub panel in the barn fed with three conductors or four? Remember, a sub panel has to have separate neutral and ground conductors and they can't be bonded in the sub panel. You will need a floating neutral bar and a separate bonded grounding bar. Feeding a 120/240 sub with only three conductors is a violation. You will also need grounding electrodes at the barn.

 

[Little Bill]

You mention neutral/ground as if they were one. Is the sub panel in the barn fed with three conductors or four? Remember, a sub panel has to have separate neutral and ground conductors and they can't be bonded in the sub panel. You will need a floating neutral bar and a separate bonded grounding bar. Feeding a 120/240 sub with only three conductors is a violation. You will also need grounding electrodes at the barn.

If you read my OP, you would see that I listed ground and neutral readings separately.
I only wrote neutral/ground later in the post for brevity.

 

[Smart $]

You mention neutral/ground as if they were one. Is the sub panel in the barn fed with three conductors or four? Remember, a sub panel has to have separate neutral and ground conductors and they can't be bonded in the sub panel. You will need a floating neutral bar and a separate bonded grounding bar. Feeding a 120/240 sub with only three conductors is a violation. You will also need grounding electrodes at the barn.

First, he didn't elaborate as to the panel being service or feeder (i.e. sub).

Second, if sub, it depends on when it was installed. Under older Code editions (2002 and earlier...?), a feeder to another structure without EGC was permitted. The grounded conductor was bonded to both the structure's GES and equipment ground at the disconnecting means.

 

[K8MHZ]

If you read my OP, you would see that I listed ground and neutral readings separately.
I only wrote neutral/ground later in the post for brevity.

Gotcha.

Just curious, did you measure the voltage at the house side with the conductors connected to the breaker? Making sure you are touching only the conductor, not the terminal?

 

[K8MHZ]

One question I would be asking myself was how there could be a dead short between A and B somewhere underground and not show up as a low resistance value from either to the earth.

Or was that not checked?

 

[Smart $]

One question I would be asking myself was how there could be a dead short between A and B somewhere underground and not show up as a low resistance value from either to the earth.

...

Why would you expect a dead short between A and B to have low resistance to ground???

 

[GoldDigger]

Why would you expect a dead short between A and B to have low resistance to ground???

If the fault was produced by a shovel nicking or cutting into the cable, the two would be likely to go together.
But if the fault was entirely internal between two conductors (and there was no bare EGC or neutral) and the sheath was not compromised, you would not get any conduction path to ground.

 

[K8MHZ]

Why would you expect a dead short between A and B to have low resistance to ground???

From an opening in the insulation of the A and B conductors. I doubt they could short out and maintain a perfect seal from the earth.

 

[Smart $]

From an opening in the insulation of the A and B conductors. I doubt they could short out and maintain a perfect seal from the earth.

Perfect... perhaps not. All I'm trying to point out is not to expect low resistance to ground.

 

[Little Bill]

First, he didn't elaborate as to the panel being service or feeder (i.e. sub).

Second, if sub, it depends on when it was installed. Under older Code editions (2002 and earlier...?), a feeder to another structure without EGC was permitted. The grounded conductor was bonded to both the structure's GES and equipment ground at the disconnecting means.

It was in a barn about 200' from the house main.

First I thought he had a loose or open neutral, but the voltage measurements didn't show that. I only showed around 190V to the sub panel at the barn but there wasn't any high voltage on one of the legs.
I had around 72V on one leg-neutral and around 119V -neutral on the other.

Gotcha.

Just curious, did you measure the voltage at the house side with the conductors connected to the breaker? Making sure you are touching only the conductor, not the terminal?

All the megging was done with all the conductors, on both ends, disconnected.

These answers in red above were in post #1 & #3 respectively.

I didn't say before, but I took the readings from both the barn sub panel and the house main and got the same readings.

 

[K8MHZ]

I didn't say before, but I took the readings from both the barn sub panel and the house main and got the same readings.

You got the same voltage readings?

How is changing conductors to a sub panel going to affect the voltage at the main?

Or am I not understanding you correctly?

 

[Little Bill]

You got the same voltage readings?

How is changing conductors to a sub panel going to affect the voltage at the main?

Or am I not understanding you correctly?

No, I'm saying I took the megger readings at both the main and the sub.
So, here's the way I did it:

Disconnect all 4 conductors from both ends

Make sure all conductors were separated, none touching each other or anything at all

Take readings at sub - line to neutral
line to ground
line to line

Take readings at the main panel - line to neutral
line to ground
line to line

Readings identical at each panel

 

[GoldDigger]

Thanks for the additional details.
It is a little hard to be sure with the inability of the Megger to show small (~1k) resistances, but based on all of the data I would say you have one of two things:

1. Some kind of load connected at an intermediate point that you are unaware of, or
2. A fault which combines a high resistance spot in the wire with a moderate resistance fault between conductors on the barn side of the high resistance. The parallel current through the fault would have to be high enough to cause approximately the same voltage drop across the series resistance whether you have a load on that phase at the barn or not.

If these are individual conductors in conduit or PVC, the only way I could explain high resistance to ground would be if the heat fused the insulation of the two wires together around the fault. Stranger things have happened.

If it is USE UF, or other cable, then the high resistance to ground is even easier to understand.

Tapatalk...

 

[K8MHZ]

No, I'm saying I took the megger readings at both the main and the sub.
So, here's the way I did it:

Disconnect all 4 conductors from both ends

Make sure all conductors were separated, none touching each other or anything at all

Take readings at sub - line to neutral
line to ground
line to line

Take readings at the main panel - line to neutral
line to ground
line to line

Readings identical at each panel

Gotcha.

I would have also checked voltages on both sides. I was just wondering if you did that or not.

 

[Little Bill]

Thanks for the additional details.
It is a little hard to be sure with the inability of the Megger to show small (~1k) resistances, but based on all of the data I would say you have one of two things:

1. Some kind of load connected at an intermediate point that you are unaware of, or
2. A fault which combines a high resistance spot in the wire with a moderate resistance fault between conductors on the barn side of the high resistance. The parallel current through the fault would have to be high enough to cause approximately the same voltage drop across the series resistance whether you have a load on that phase at the barn or not.

If these are individual conductors in conduit or PVC, the only way I could explain high resistance to ground would be if the heat fused the insulation of the two wires together around the fault. Stranger things have happened.

If it is USE UF, or other cable, then the high resistance to ground is even easier to understand.

Tapatalk...

If you're talking about reading through a load with the Megger, I don't understand why you would say that when I clearly said all the conductors were disconnected from each panel.

If you're talking about voltage readings they were taken with and without loads. There was about a 1 volt drop with just the lights that were on the good leg on. With the small water tank heater and the lights on there was a 2 volt drop. I had the 240V water heater turned off at the breaker and later just disconnected. Voltage at the bad leg at sub stayed around 72V with loads connected, but nothing actually worked. No phantom voltage as it didn't drop out when tested with loads connected.

The conductors were 2-2-4-6 XLPE Al direct bury. I don't know if they were in conduit past the elbows and conduit emerging from ground. But I'm fairly sure the conductors were direct buried.

Gotcha.

I would have also checked voltages on both sides. I was just wondering if you did that or not.

Yes voltage was checked on both ends. At the main panel: 241V line-line, 120V line to neutral, 121V other line to neutral.
I've already listed the voltages at the sub panel.

 

[kwired]

I may have missed something, but I think you determined you have an open ungrounded conductor somewhere.

Easy way to determine this would have been to connect a reasonably heavy 120 volt load to each line and check voltage readings. The good line will have only a little voltage drop, the bad line with that kind of a load often will drop to nearly nothing and the load will not operate. A 1200-1500 watt heater works well as a test load.

If you have 240 volt loads in an outbuilding like that and an open ungrounded conductor in the feeder, the 240 volt loads will confuse you with voltage readings as you will get feed thru the 240 load to make the open line look like it is mostly working, turn the 240 volt load(s) off and then everything on the bad leg quits working altogether.

 

[S'mise]

Just my 2 ohms

Just my 2 ohms

I would put the megger down and check (for low resistance) of each conductor with a good ohm meter. Disconnect all wires and on one side tie (short) a line wire to ground and you can compare resistances on the other end. Chapter 9 tables will give you a rough idea of what resistance you should get. Sure sounds like a bad underground splice to me.