ground fault with reversed polarity at dc disco

Status
Not open for further replies.
Thanks for the article on finding ground faults. I read it online after SMA service line could not help me with a peculiar problem. I am hoping that you can.

An insulation resistance error appeared on a customer’s TL series. Zero watts. SMA told me this is indicates a ground fault. System was built by others about ten years ago and the first fault noted about three weeks ago by the customer.

Module string of 12 with pv wire in open air read 444V from pos to neg, polarity normal. This is the expected Voc.
Pos to Gnd read below ten volts and descending. Neg to Gnd also below 10V and descending toward zero.

I reconnected the string to the inverter dc disconnect switch. In free air at the disco I read ‑446 from pos to neg BUT THE POLARITY WAS REVERSED, meaning my meter positive probe on the positive RED conductor and my meter negative probe on the negative WHITE conductor, with the minus sign before the 446 in my display. I double checked to be sure my leads were not crossed and they were not. Pos to Grd read -423 (polarity reversed also, meaning my pos probe on pos and my negative probe on Gnd and a minus sign in display). Neg to Gnd read less than 1 volt (I didn’t note the polarity from N to G).

With the string still connected at the array and the conductors still in free air at the disco, I returned to the array and read 450V Pos to Neg POLARITY NORMAL. Neg to Gnd read 24 V with the positive probe lead on the Gnd and the negative probe on the negative WHITE conductor. Air had gotten cooler, probably accounted for slightly higher Voc.

I cannot grasp how the polarity can reverse, but suspect there is faulty insulation inside the buried conduit between the array and the Dc disco. WHERE WOULD YOU SUSPECT THE PROBLEM TO BE FOUND? To clarify, the array is about 250 feet from the inverter- mounted DC disco. Replacing conductors without a clear idea of which one(s) are at fault would be a risky and costly task. I do not have a megger. I suspect that either the positive conductor or the negative conductor or maybe even both of them have abraded insulation somewhere in the conduit. I strongly believe that no one has reversed the red and white conductors. I am sure that rain water has entered the buried PVC conduit. It may be helpful to note that the conduit serves two arrays and two inverters but only one of the inverters has the fault and the strange polarity. In the conduit are 5 conductors, two white negatives, two red positives, and one green ground common to both arrays and both inverters. One inverter is operating normally.

I am about to run a temporary test conductor about 250 ft. over the surface from array to dc disco to substitute for one of the buried conductors. If that is inconclusive, I will substitute for the other. If still inconclusive, substitute for both with two temporary test conductors over the surface, hoping that it will not be necessary to pull more than one or two conductors out for replacement. I hesitate to do this without agreement from a trusted source because I am so confused by the voltage readings.

Your thoughts, please?

Andy Cazzato
[FONT=&quot]SANTA CRUZ SOLAR[/FONT]
831-685-2786 office
831 251-2786 mobile
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
My thoughts are that it's probably a fault on the white conductor in the conduit, but your post contains some confusing information that may be red-herrings but make it hard to be sure what's going on.

...
An insulation resistance error appeared on a customer’s TL series. Zero watts. SMA told me this is indicates a ground fault. System was built by others about ten years ago and the first fault noted about three weeks ago by the customer.

The TL series is less than 10 years old (only about 6), but if you called SMA and discussed the error code with them, then I'm assuming you understand which inverter you have. Older inverters would not have indicated an insulation (did you mean 'isolation'?) error, only a ground fault. The only reason I'm talking about this is the polarity issue you raise later on...

Module string of 12 with pv wire in open air read 444V from pos to neg, polarity normal. This is the expected Voc.
Pos to Gnd read below ten volts and descending. Neg to Gnd also below 10V and descending toward zero.

You are saying you unplugged the panels at the array to take this reading, right? If so his part of your post indicates that any ground fault is not on the array. Those descending voltages would be phantom.

...I reconnected the string to the inverter dc disconnect switch. In free air at the disco I read ‑446 from pos to neg BUT THE POLARITY WAS REVERSED, meaning my meter positive probe on the positive RED conductor and my meter negative probe on the negative WHITE conductor, with the minus sign before the 446 in my display. I double checked to be sure my leads were not crossed and they were not. Pos to Grd read -423 (polarity reversed also, meaning my pos probe on pos and my negative probe on Gnd and a minus sign in display). Neg to Gnd read less than 1 volt (I didn’t note the polarity from N to G).

If this is really a TL inverter then you have a ground fault on the white conductor. As for the polarity...

If this is really a TL then there should not be a white conductor (code violation to use a grounded conductor color for an ungrounded conductor). If that's the case, then the installers didn't follow a typical color convention, and you shouldn't make assumptions that the red is positive and the white is negative. Maybe it's as simple as that they used white for positive and red for negative. (Using red for negative is anti-conventional, but allowed by code, at least back then). You stated "I strongly believe that no one has reversed the red and white conductors." But you gave no reason for this belief. You need a good reason to believe that.

If this is somehow NOT really a TL inverter, is there any possibly this is a positively grounded Sunpower system? In that case you don't have a ground fault with those readings when panels are connected to inverter and ground fault fuse is not open.

Also, your term 'free in air at the disco' is a little confusing. Why undo the conductor terminations at the disco, rather than just open the disco and measure the terminals? I'm just wondering if I understand what you're describing here.

... WHERE WOULD YOU SUSPECT THE PROBLEM TO BE FOUND? ...

If you get desired voltages at the array when unplugged, but undesired voltages at the disco when the disco is open and the array is plugged in, then the problem is between the array and the disco. I think you have correctly figured out that much.

If the red conductor shows full voltage to ground and the white conductor shows none, then the fault is on the white conductor. White and ground are connected, so you have (roughly) no voltage between them, and (roughly) the same voltage between either and red. If the white-ground 'connection' is through some dirty water then it makes sense if the red-ground voltage is a little lower. If the fault were between panels on the array then you'd have voltage between ground and both circuit conductors, not just one.

As far as the reversed polarity, I do not believe it's physically possible for a fault to cause this. Someone wired something in a manner you have not fully checked out and understood. Either you have made a false assumption about the wire colors, or you crisscrossed the conductors with the other system that shares the conduit, or some other such error.
 
to jaggedben

to jaggedben

Thanks for your thorough analysis. Sorry for the vagueness . I will go back to the job with your notes in hand. Inverters and modules were replaced by others in the past, and I am quite sure that they were not positive ground SPR products. This also explains the original use of white negative conductors connecting to the TL series. I will tape over them with appropriate colors after I get this problem fixed.
 
Status
Not open for further replies.
Top