DC feeder test

[Weezy]

Hello all,

I am testing a pv system and discovered an unexpected reading. The test is a ground/feeder continuity test. It verifies continuity of the positive feeder, the ground system, and correct polarity of the feeders. Two 500mcm feeders are pulled from a disconnect, one positive and one negative, and terminated at a combiner box at the array. The outputs are pulled to an inverter right next to the disconnect and terminated. The positive bus and the ground bus at the inverter are jumped to created continuity. This is a positively grounded system.

Two thirds of the combiner boxes passed the test, testing at less than 1 ohm between the positive bus and the ground bus at the combiner boxes. The negative feeders all tested OL as they should, proving correct polarity. I have about ten boxes that failed, from four different inverters, all tested the same way. All ten of the combiner boxes that failed are in the center the structure. The feeders rang out but read from 9 ohms all the way to 14 ohms from the positive bus to the ground bus. I have checked for voltage at the combiner boxes and got nothing. I tried the test at the disconnect and they passed. Voltage at the disconnect was occasionally -0.1 v. I've tried three different meters and testing at different times of day and they always fail. Why would most pass and some fail? I established continuity but high resistance. I feel the ground system is solid. I'm curious why if I change leads during the test and use the black probe on my fluke on the positive bus instead of the red, the reading goes from 10 ohms to 0. Why is this? Does this mean there is some voltage? I am not finding any. The Ac side of the inverter is already terminated at both ends but none of the equipment has been turned on. I've already completed a thorough string test and everything passed.

Any thoughts or suggestions are appreciated. Thank you.

 

[jaggedben]

Did you ring out the equipment grounding conductor? You might have a bad one that is common to all the combiner boxes in that one area. You said you 'feel the ground system is solid'. That doesn't sound like you've tested the EGCs to the same standard as the other conductors. If your 'feeling' is based on ground rods being everywhere or something like that, that is no good. You need a solid EGC connection from all combiner boxes back to the inverters.

 

[Weezy]

Did you ring out the equipment grounding conductor? You might have a bad one that is common to all the combiner boxes in that one area. You said you 'feel the ground system is solid'. That doesn't sound like you've tested the EGCs to the same standard as the other conductors. If your 'feeling' is based on ground rods being everywhere or something like that, that is no good. You need a solid EGC connection from all combiner boxes back to the inverters.

Thank you for the comment. Here's a some more detail about the grounding system:

every combiner box has its own egc. It is bonded on bushings at the Meyers hub, then at the cable tray and the can then finally at the ground bus at the disconnect, pulled all the way, not spliced. The rest of the grounding is weebs between the panels and the purlins and braids between all the purlins, and the tray and trough.

Do do you have any comment about the two readings I get with my meters? I get an average of 11 ohms with my red probe on the positive bus and the black probe on the ground bus at the combiner box. When I switch the leads and the black is on the positive bus I read 0 ohms. Why is this?

 

[GoldDigger]

Thank you for the comment. Here's a some more detail about the grounding system:

every combiner box has its own egc. It is bonded on bushings at the Meyers hub, then at the cable tray and the can then finally at the ground bus at the disconnect, pulled all the way, not spliced. The rest of the grounding is weebs between the panels and the purlins and braids between all the purlins, and the tray and trough.

Do do you have any comment about the two readings I get with my meters? I get an average of 11 ohms with my red probe on the positive bus and the black probe on the ground bus at the combiner box. When I switch the leads and the black is on the positive bus I read 0 ohms. Why is this?

Possibly the result of the bypass diodes in the panels if the negative DC lead is grounded? And remember that the ohmmeter will not work if there is any DC between the two places you connect it.
The EGC plays no role in this and except for any GF sensing mechanism inside the inverter there should not be any continuity between the ungrounded conductor and the EGC other than through the panels.
The grounded conductor from the array, if there is a solid ground at the inverter, should measure close to zero ohms to the EGC unless there is a break in the wiring.
I think you do not know for sure how (or if) the array DC conductors are effectively grounded.

If you disconnect the panels and the inverter you should measure an open circuit between both DC leads and the EGC unless you have explicitly grounded one of the leads somewhere.

 

[Weezy]

Did you ring out the equipment grounding conductor? You might have a bad one that is common to all the combiner boxes in that one area. You said you 'feel the ground system is solid'. That doesn't sound like you've tested the EGCs to the same standard as the other conductors. If your 'feeling' is based on ground rods being everywhere or something like that, that is no good. You need a solid EGC connection from all combiner boxes back to the inverters.

Possibly the result of the bypass diodes in the panels if the negative DC lead is grounded? And remember that the ohmmeter will not work if there is any DC between the two places you connect it.
The EGC plays no role in this and except for any GF sensing mechanism inside the inverter there should not be any continuity between the ungrounded conductor and the EGC other than through the panels.
The grounded conductor from the array, if there is a solid ground at the inverter, should measure close to zero ohms to the EGC unless there is a break in the wiring.
I think you do not know for sure how (or if) the array DC conductors are effectively grounded.

If you disconnect the panels and the inverter you should measure an open circuit between both DC leads and the EGC unless you have explicitly grounded one of the leads somewhere.

Thank you for your comments. That is correct, there is no continuity between the ungrounded feeder and egc, so that tests good. The negative feeder tests open to ground. And yes, I expected to read close to zero ohms between the grounded conductor and the egc. That is where my problem is. There are no breaks in these conductors. I do know that the positive feeders are grounded. Twenty one of these boxes tested good and there's ten that have this issue. Three feeders from one inverter, two from another, three feeders from the third inverter, and two more feeders belonging to the fourth inverter. I will try disconnecting the lugs at the inverter and the combiner box to see if it reads open. As it is now, I continue to have no voltage readings at the box, disconnect, or inverters. I assumed the same thing, that the ohmmeter was not working correctly because there is voltage present, but I just can't seem to find where it is. Thank you for your help.

 

[jaggedben]

The EGC plays no role in this and except for any GF sensing mechanism inside the inverter there should not be any continuity between the ungrounded conductor and the EGC other than through the panels.

He's testing for continuity between the grounded conductor and the EGC at the combiner. Since the grounding occurs at the inverter, the EGC definitely plays a role in this. I think you misread some part of his post.

The grounded conductor from the array, if there is a solid ground at the inverter, should measure close to zero ohms to the EGC unless there is a break in the wiring. I think you do not know for sure how (or if) the array DC conductors are effectively grounded.

If you disconnect the panels and the inverter you should measure an open circuit between both DC leads and the EGC unless you have explicitly grounded one of the leads somewhere.

Good point. The positive conductors are probably grounded by a fuse in the inverters, but that's an assumption on my part. In order to diagnose what's going on one ought to know exactly the method. It could be a bad fuse or any of a number of things going on at the inverters.

Thank you for your comments. That is correct, there is no continuity between the ungrounded feeder and egc, so that tests good. The negative feeder tests open to ground. And yes, I expected to read close to zero ohms between the grounded conductor and the egc. That is where my problem is. There are no breaks in these conductors. I do know that the positive feeders are grounded. Twenty one of these boxes tested good and there's ten that have this issue. Three feeders from one inverter, two from another, three feeders from the third inverter, and two more feeders belonging to the fourth inverter. I will try disconnecting the lugs at the inverter and the combiner box to see if it reads open. As it is now, I continue to have no voltage readings at the box, disconnect, or inverters. I assumed the same thing, that the ohmmeter was not working correctly because there is voltage present, but I just can't seem to find where it is. Thank you for your help.

I'd want to know if these particular 'feeders' share a particular conduit run or tray, or not. (BTW, the NEC term would be 'PV output circuit'.) You did say that they are all in one part of the structure of the array.

 

[Weezy]

Thank you for your thoughts. I am prohibited to post any photos of this project, but if I could I would show you our terminations at the inverter. My company is subcontracted to do the DC side only. I have never been inside the inverter other than the small portion on the bottom where we crimped and landed the lugs. All the positive outputs from the disconnect are landed on a bus in the bottom of the inverter. All the egc's are landed on a bus there too. The negative outputs are landed on separate stand off lugs, because they are fused. To the right of all that is the 480v AC lugs, pulled and terminated by a different contractor. I'm describing this in detail so you understand that the positive conductors come straight through from the combiners and land collectively on the bus. Wouldn't I have this problem with all the positives if there was something wrong internally with the inverter? I have four inverters and a small portion of these circuits with each system have this issue. I am going to locate a key for the inverter and see what I can find out about how the conductors are grounded beyond the terminations I described. When I do this continuity test, I put a jumper between the positive and ground busses in the inverter to establish continuity. And yes, each circuit shares a path down a cable tray after in reaches the inverter room. Thank you for your time and your help.

 

[jaggedben]

As far as I can think, leading suspects would be either:

1) a bad EGC that is common to all the circuits with the issue. You still haven't given enough detail to rule this out. It's rare but not unheard of for a wire to have a manufacturing defect that leaves a bad or open connection.
2) a small ground fault somewhere on a negative conductor that is putting DC on the EGC(s) you are testing. What you've said seems to rule this out if you did your tests for voltage thoroughly, but maybe worth redoing?

...Wouldn't I have this problem with all the positives if there was something wrong internally with the inverter? ....

Yes, one would think it would occur per inverter. But also, since you also say you install a jumper from positive to ground when you test, this seems less likely to be a cause of the problem.