SolarEdge DC Isolation Fault diagnosed as comm board failure - is that possible?

[GoldenDragon]

I finally got an explanation today as to why we've waited over two weeks for a comm board:

I've been away at a training conference but I wanted to get back to you regarding your system. I see your system had a gap without production going out on the 5th completely, returning for a few days with very low production and going out again after the 8th. It is producing since yesterday, but it would be great to get someone out to troubleshoot the cause. The error 25 iso fault is typically common if there is a connection somewhere that came loose, but it can also be due to moisture at the points of connection. If the resistance during start up initiation of the inverter fluctuates, this could cause this error reading on the inverter. Using the diagnostic tool on the inverter we can determine where the issue is if there is one. However, according to SolarEdge there are instances when the sensor in the inverter comm board reports issues falsly. It was SolarEdge's recommendation to go to your place with a communication board in case there is no issue with the connections on your system. Our goal is to make sure we can resolve the issue so it would be ideal to have the communication board on hand, but if you want I can send a crew there tomorrow to verify the connections and test the system.

I can't confirm that the communication board can fix the issue, but I can't confirm that resetting all the connections of the system will either if the issue is in fact with the board.

Nonetheless, it is our goal to resolve your issue and if tomorrow works, we would be glad to send a crew out there to troubleshoot the issue.

My response:

Thank you for the explanation. Replacing the comm board finally makes some sense. I still feel like an isolation fault is a serious enough thing to warrant sending someone immediately rather than waiting over two weeks on a part. At the very least, the customer needs to be warned what the fault means and that there could be deadly voltage on exposed parts.

I perhaps should have added that customers should also be informed exactly what's going to be done and why, but then again I expect most people wouldn't understand so I can see why that's not standard policy.

I'll report back when they've found the source of the water leak, but I let them know there's no rush at this point.

[GoldenDragon]

The tech and I wrote back and forth a few times and he seemed mostly on the ball, but sometimes I felt like he forgot parts of what we'd already talked about even an hour or two earlier. He didn't seem to feel the isolation fault was a dangerous situation but his explanations weren't convincing. I had to do my own math to be convinced. Basically, even if the DC to ground had zero resistance (it had 28kOhm resistance in my case), and your body was drenched in sweat so your resistance was 1000ohms (from NIOSH data on human body resistance), the resistance in a 100 foot, 8AWG DC circuit is only 0.063ohms. So, if you touched the ground of the system operating at peak power (theoretically 19A) while holding a grounded metal handrail (which isn't too hard with our install), the current that would travel through you is 19A * (0.063/1000) = 0.001A, well in the range of safe. I also could find no stories online of an isolation fault causing someone injury, so apparently it's just not really a problem.

I'm not happy that they waited 2 weeks on a part that was unlikely to solve the problem, but I understand the desire to have as much stuff on hand as possible so they don't have to make multiple trips to my distant location. I think they should have sent a local electrician up to diagnose it asap before contacting SolarEdge, though I also get the feeling SolarEdge had some reason to think there was a part failure which is why they supplied a replacement part under warranty. That, or SolarEdge tech support is useless.

So, after discussing the isolation fault with my contact, he scheduled a team to come up the next day. Yet when the person in charge of the team reviewed the info, something seemed off, so they called me after they were to supposed to have shown up. I don't really care because I work from home and the system was working, but if I'd taken a day off to be there for them I would have been pissed. Very bad policy to break appointments, especially without prior notice.

Anyway, I'm glad they called because somehow the person I'd been talking to still thought the inverter was detecting an isolation fault that wasn't really there and he didn't realize it was caused by water infiltration. I don't see how the heck he still thought that after my explanations and pictures, but... Looking back at the conversation, it seems clear that he understood at some point so my only theory is they deal with so many different things every day he actually forgot and went back to the previous explanation in his mind (or in notes linked to my account) or something? I don't know. There really isn't a good excuse but at least the guy actually organizing my team noticed something was off and spoke to me directly.

Two guys showed up this morning, 1.5 hours later than expected (maybe because the organizer didn't realize how far away I was). They hadn't brought conduit sealing putty like the organizer specifically said they would... Argh. Attention to detail seems quite lacking through this whole process, though I've found that attention to detail is a rare skill in all walks of life. It's part of why I prefer to do everything myself, but there just isn't enough time. Anyway, he did a thorough job of diagnosing and explained his reasoning, so that was good. It looks like most of the water was coming in through a junction in the conduit that was loose (not screwed down tightly enough). Most of the rust started immediately below that junction. He replaced a couple of pipe sections that had rust and found no rust anywhere else. Wires exposed to water were also replaced.

My original theory was that air was passing through the conduit in our hot attic and then hitting colder pipes on the shaded side of the house where moisture condensed out of the air. It doesn't really look like that was the cause but I had them seal both ends of the conduit with caulk just in case (which NEC reg 300.7 says you're supposed to do when sections of pipe are exposed to significantly different temperatures). He also suggested drilling a weep hole in the DC disconnect which I think is a terrible idea because it means I may never realize there's water getting in and conduits may rust out over the years. It might make some sense if I was unable to check the disconnect for water from time to time.

So, it's definitely sloppy that the original installers didn't tighten all the conduit nuts, but as long as a company is willing to come out and replace parts damaged by their own mistakes (which they did, without argument or prompting), they're only hurting their own bottom line by being sloppy. I have no problem forgiving mistakes as long as a company fixes them. Less forgivable is the money I lost from lack of solar generation for 22 days during the most productive month of the year. They say they'll pay me back if the system doesn't produce the minimum amount they guaranteed, but all solar installers guarantee only about 90% of what they actually think the system will produce. I estimate I lost about 8.7% of the year's production, so it may still produce more than their min guarantee in which case I get nothing. Even if I do get something, it won't be full compensation for what the system should have produced.

[sensij]

Glad they are taking care of you! Check your math on the potential fault current... use 360 V across whatever resistance you expect to calculate current. Very dangerous, that it why it must be detected and shut down.

[J.P.M.]

Without looking at the situation (being there), I'd have trouble believing the condensation theory as explaining the water unless there was a lot of moisture being added to the air after it entered the attic and/or the conduit was regularly at a temp. below the dew point, both of which, while possible, seem unlikely.