[RE-wrenches] Unisolar Intermittent Ground Fault

[Robert Nuese]

Dear Wrenches

I've been experiencing a troubleshooting problem that you may be able to
help me with.  Or, can anyone refer me to someone, maybe an electrical
engineer?, who I could hire to help.

About 7 years ago I installed Unisolar modules on all the roofs of a complex
of 12 mixed use buildings and one commercial building. These fed the
grid through a variety of Fronius IG and SMA SunnyBoy inverters.

Ever since, we have had occasional intermittent ground faults on a few of
the roofs, and more frequent ones on one roof. At least some of these I
suspect were caused by the roofer. Standard procedure on the Unisolar
modules is to put two screws through them that hold on the roof ridge
z-strip metal closure piece. These also keep the modules from sliding off
the roof in extreme hot weather. I found a number of screws placed
slightly outside the allowed screw zone. After I'd moved the screws
(and removed the raised metal edges the screws had made in the roof,
and put insulation between roof and module in those locations), this has
seemed to fix some problems. Similarly, putting tefzel patches, per Uni-
solar instructions, on some big dings and scratches may have fixed some
others.

However, I still occasionally, rather rarely, get ground fault indications
on some inverters, and rather regularly get them on one specific IG 3000.
Some times the indication is just in the software, and the problem disappears
through resetting by disconnecting the inverter and reconnecting to reset the
software. However, on that worst Fronius, it has usually blown the 1 amp
ground fault fuse, indicating that there very likely has been a ground fault.

I suspect that these ground faults are brought on by thermal expansion and
contraction, that may only occasionally, and briefly, bring hot and ground
elements of the system into close enough proximity. Dampness seems to
play a part, the ground faults almost always occur during the rainy season,
and usually during a rainy period.

But whenever I get out to the site to test the system, the ground fault is
always gone. I test by disconnecting both wires of each string, measuring
the total voltage across, and then the voltage from each to ground. If there
is a clear voltage to ground, then there is a ground fault, and its location in
the string can be easily determined. If the voltage steadily goes toward
zero, then a ground fault is not clearly present, and the location of where
one had been is totally obscure.

I've done very thorough visual inspection of the roofs, and also of the
disconnects, junction boxes, and gutters that the wires pass through.
As mentioned, the modules have had some problems, and I've fixed them
all, as well as can be done. The wires all seem fine. Well connected, neat,
not close to raw metal edges, insulation ok, etc.

I've asked Unisolar (before they went bankrupt, now they don't answer the
phones), and they've just recommended the testing I've done, and not had
any other ideas. Same with Fronius - they think it's outside of the inverter,
and the blown fuses support that position.

I could go on with more details about what I've done and thought about,
for instance, is the Fronius too sensitive? But I'll catch my breath and see
if anyone else has some input before I go into more detail.

SO what am I missing? What test should I do? How can I fix this problem?

OR does anyone know of somebody with more expertise who I could hire
to help? I'm a fairly experienced solar installer, but I think this might be
something that requires an electrical (or electronic?) engineer with a deep
understanding of electrical theory, and with some advanced test equipment.
Any recommendations? Probably needs to be someone within 100 miles of
Sebastopol CA.

Thanks,
Robert Nuese






-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.re-wrenches.org/pipermail/re-wrenches-re-wrenches.org/attachments/20121126/ebbc6a9b/attachment-0002.html>