SolarEdge string is down.

[ericy]

We have a SolarEdge system that has two strings, but one of the strings seems to be down. While the system was installed in 2014 - we bought the house last fall, and I am still coming up to speed on some of this. The system has 39 panels, and is rated for 9.75kW, with a SE10000 inverter.

I can tell from the monitoring portal that the system operated correctly for all of 2015 - the output dropped by about 60% in May/Jun of 2016. This past summer, peak output was about 4kW.

I can look on the monitoring portal - I can see there is one optimizer that has not reported status since 6/25/16. Over 2 years, in other words. This more or less coincides with the power drop. There a few other optimizers that oftentimes are late in reporting - as I write this, anywhere from a day to a month late in reporting.

The remainder of the optimizers on the string are reporting on a regular basis, but they seem to remain in "Safe Mode" with an optimizer voltage of 1V.

The original installer is still around, but they are not all that responsive (I suspect they were slammed with new install business). I had booked the appointment in mid-Nov, and on the day of, there was a noreaster - there was no way anyone sane was going on the roof in that kind of weather. At the time they were inspecting any interior wiring and refreshing their memory as to how they did the install back in 2014 (mainly wondering how they routed the wiring through our rather unusual house). He wondered about whether the system was showing any fault codes - that's something that apparently I can't currently see in the management portal. He said he could grant me installer access, but that never happened. They said they were going to come out when the weather was better so they could go up on the roof, but I see no indication that this has happened either.

When they were here, they were kicking around various ideas. Mainly checking the connectors where they did internal wiring, but perhaps a bad ground on one of the optimizer? They had wondered whether one of the connectors for the string had popped off (if that had happened, wouldn't that imply that all of the other optimizers on that string would be unable to report??). I am wondering if a failed optimizer could take down an entire string. And I am curious as to why it is that a few of the other optimizers are having intermittent communications.

I would rather not go on the roof myself - my first choice was to have the original installer fix the thing, but I am wondering if I have a plan-B. Without installer access to the monitoring site, I am not sure what other options I really have - a different vendor probably wouldn't have the ability to configure a new optimizer, for example.

After that one day where they did come out, I got horribly busy with stuff, so I back-burnered the whole thing, but I thought I would ask here to see if anyone had any thoughts as to what the underlying problem might actually be. But my busy-ness is coming to an end, and I want to get something going here. In the next day or so, I guess I could call yet again and see if I can get the original installer to do anything at all.

Thoughts?

[ButchDeal]

Have they tried re-pairing the optimizers? this should have them reconnect to the inverter and if there is a break some of them will drop out.
It is pretty rare for a single optimizer to take out a string but there were some bugs with communication so a call to SolarEdge to push latest firmware to the inverter could help as well.
if they gave you full owner access you should be able to see the errors on the the inverter monitoring.

[ericy]

Have they tried re-pairing the optimizers? this should have them reconnect to the inverter and if there is a break some of them will drop out.
It is pretty rare for a single optimizer to take out a string but there were some bugs with communication so a call to SolarEdge to push latest firmware to the inverter could help as well.
if they gave you full owner access you should be able to see the errors on the the inverter monitoring.

I don't think they have done much of anything as of yet. That was part of the reason for posting here - so I could get another set of eyes on the thing, and when I get the installer back on the phone I can give him more concrete things to do for me.

Any idea of the version of the latest firmware? Currently I have this:

Name	Inverter 1

Manufacture

SolarEdge

Model

SE10000

DSP2 Version

2.52.154

Communication Board (CPU) Version

2.640.0

DSP1 Version

1.210.1060

I RCD [mA]

0.01

Inverter Status

Standby

Last Isolation Value [KOhm]

0

P AC [W]

0

Power Limit [%]

100

How do inverter firmware updates work? Can SE push the things or do I need to get the installer to do this? Or can I do it myself (I would need to find the actual firmware file to download first, I guess).

[ButchDeal]

communication board version is up to 3.x now

SolarEdge pushes the update to firmware (only way).
installer or solaredge can set to repair remotely. you can do it locally as well

[ericy]

communication board version is up to 3.x now

SolarEdge pushes the update to firmware (only way).
installer or solaredge can set to repair remotely. you can do it locally as well

Sounds like it is worth a shot, then.

If the only (current) connection from the inverter to the outside world is via ZigBee, can they still push? Do I need to physically be there when they do this, or can I ask them to do it while I am at the office?

[ButchDeal]

Sounds like it is worth a shot, then.

If the only (current) connection from the inverter to the outside world is via ZigBee, can they still push? Do I need to physically be there when they do this, or can I ask them to do it while I am at the office?

yes the zigbee allows access directly. You can ask them to take a look and push latest firmware as well as re-pair. re-pair takes less than 10 minutes.
pushing new firmware can take 24 hours but usually much faster on zigbee and ethernet connections, slower on cellular.

you do not need to be there for any of it.

[ericy]

yes the zigbee allows access directly. You can ask them to take a look and push latest firmware as well as re-pair. re-pair takes less than 10 minutes.
pushing new firmware can take 24 hours but usually much faster on zigbee and ethernet connections, slower on cellular.

you do not need to be there for any of it.

I talked to them on the phone today - they said they could see that there was a ground fault on the DC side. I believe that the "Last Isolation Value" is the key to this - when there is sunlight and the thing is on, it shows values like 2-3000. I am not sure what is a normal value - I used the SolarEdge API to pull values at 30 day intervals - when the thing was first installed numbers like 6000 were common, but not long after that the values started dropping:

{"date":"2014-11-01 14:01:44","totalActivePower":659.5,"groundFaultRes istance":6000.0}
{"date":"2014-12-01 13:01:29","totalActivePower":5427.0,"groundFaultRe sistance":6000.0}
{"date":"2015-03-31 14:03:58","totalActivePower":8567.0,"groundFaultRe sistance":6000.0}
{"date":"2015-04-30 14:00:33","totalActivePower":4942.0,"groundFaultRe sistance":6000.0}
{"date":"2015-05-30 14:00:06","totalActivePower":7793.5,"groundFaultRe sistance":2406.13}
{"date":"2015-06-29 14:02:39","totalActivePower":8120.0,"groundFaultRe sistance":4514.09}
{"date":"2015-07-29 14:04:27","totalActivePower":1535.5,"groundFaultRe sistance":2604.96}
{"date":"2015-08-28 14:02:41","totalActivePower":7578.0,"groundFaultRe sistance":3634.48}
{"date":"2015-09-27 14:03:28","totalActivePower":2388.0,"groundFaultRe sistance":2943.44}
{"date":"2015-10-27 14:02:07","totalActivePower":1708.5,"groundFaultRe sistance":6000.0}
{"date":"2015-11-26 13:04:42","totalActivePower":5858.0,"groundFaultRe sistance":3946.86}
{"date":"2015-12-26 13:03:43","totalActivePower":500.0,"groundFaultRes istance":6000.0}
{"date":"2016-03-25 14:02:38","totalActivePower":2887.5,"groundFaultRe sistance":1109.76}
{"date":"2016-04-24 14:00:49","totalActivePower":8414.5,"groundFaultRe sistance":6000.0}
{"date":"2016-05-24 14:01:29","totalActivePower":3610.0,"groundFaultRe sistance":2564.5}
{"date":"2016-06-23 14:01:40","totalActivePower":3352.0,"groundFaultRe sistance":5803.32}
{"date":"2016-07-23 14:03:44","totalActivePower":3701.0,"groundFaultRe sistance":2140.04}
{"date":"2016-08-22 14:03:59","totalActivePower":3572.5,"groundFaultRe sistance":1754.18}
{"date":"2016-09-21 14:01:32","totalActivePower":1208.5,"groundFaultRe sistance":2022.64}
{"date":"2016-10-21 14:03:56","totalActivePower":3502.5,"groundFaultRe sistance":1855.6}
{"date":"2016-11-20 13:01:12","totalActivePower":1185.5,"groundFaultRe sistance":6000.0}
{"date":"2017-04-19 14:00:44","totalActivePower":2156.5,"groundFaultRe sistance":5949.48}
{"date":"2017-05-19 14:01:05","totalActivePower":3644.0,"groundFaultRe sistance":3310.35}
{"date":"2017-06-18 14:04:51","totalActivePower":4143.5,"groundFaultRe sistance":1715.06}
{"date":"2017-07-18 14:01:01","totalActivePower":2270.0,"groundFaultRe sistance":1371.42}
{"date":"2017-08-17 14:03:32","totalActivePower":3374.0,"groundFaultRe sistance":1169.11}
{"date":"2017-09-16 14:00:03","totalActivePower":3338.0,"groundFaultRe sistance":1221.52}
{"date":"2017-10-16 14:02:29","totalActivePower":325.0,"groundFaultRes istance":1534.81}
{"date":"2017-11-15 13:01:30","totalActivePower":788.0,"groundFaultRes istance":2174.47}
{"date":"2017-12-15 13:01:07","totalActivePower":269.5,"groundFaultRes istance":5498.17}
{"date":"2018-04-14 14:01:06","totalActivePower":3711.0,"groundFaultRe sistance":2160.13}
{"date":"2018-05-14 14:01:51","totalActivePower":3534.0,"groundFaultRe sistance":828.989}
{"date":"2018-06-13 14:04:38","totalActivePower":3480.0,"groundFaultRe sistance":673.64}
{"date":"2018-07-13 14:01:39","totalActivePower":3690.0,"groundFaultRe sistance":666.378}
{"date":"2018-08-12 14:02:25","totalActivePower":3354.0,"groundFaultRe sistance":794.756}
{"date":"2018-10-11 14:03:15","totalActivePower":3390.0,"groundFaultRe sistance":619.383}
{"date":"2018-11-10 13:01:09","totalActivePower":2824.5,"groundFaultRe sistance":1840.74}
{"date":"2018-12-10 13:03:07","totalActivePower":3127.0,"groundFaultRe sistance":2040.74}

I don't know what those numbers really represent, however, and I don't know what is considered "normal".

They did push a software update of some sort, but the CPU version is now

Communication Board (CPU) Version

2.1054.0

If there is really a ground fault, then it didn't seem like there is much point in re-pairing the optimizers to the inverter.

[ButchDeal]

Wellyour installer should have seen that error or at the minimum gotten the information from solaredge.

[Ben25]

Sounds like you need go have a stern chat with the installer. If there is a fault(sounds like there is), the isolation value may change with temp/humidity. Definitely a problem and if they had looked at the monitoring, they would have seen it. Also, it is possible to update the firmware downloaded to an SD card, but having solaredge push it is *usually* easier.

[foo1bar]

When I had strings fail it was due to the fuse holders in the inverter failing (visible scorch marks).
You wouldn't have fuse holders so it would have to be something else.
It could be something with one optimizer going bad and taking out the string - an optimizer could have a problem on the panel side or on the inverter side. If it's on the panel side (ex. panel gets unplugged) then you'll see zero production from that optimizer but the rest of the string is fine. If it's on the inverter side, it could essentially stop the string from functioning.
If you're not used to doing electrical work, it's probably best to just keep bugging the installer.

[ericy]

If you're not used to doing electrical work, it's probably best to just keep bugging the installer.

I have done all kinds of elecrical in the past - my problem is that I am just far too busy with other things at the moment, and I would rather spend my time on other things. And it is a single-story house where the pitch of the roof isn't all that bad. I just would need to find time that I don't have in abundance.

Without doing any further analysis, I would be tempted to start looking at the one optimizer that hasn't reported status in over 2 years - use an MC4 bypass cable and take that one out of the string. That assumes of course that the representation in the layout page is a close approximation of reality, which might or might not be the case, but I guess it would be a starting point. But it would probably be better to take a more systematic approach and not make any assumptions.

SolarEdge has a troubleshooting guide - there is something on the inverter that helps pinpoint issues on a string, but it only helps if the installer correctly did the stringing in the layout, and I already know there are errors there (some panels are on the wrong strings - the installer admitted that he probably made some mistakes entering the data). If I get a nice day sometime, I could probably follow those instructions as it would tell me something.

[Ben25]

I would check all of the junction boxes first if there are any. I've seen a few optimizers short out, they shut down the whole inverter until that string is disconnected. It's a PITA to find them unless solaredge is able to run telemetry and pinpoint which one it is. The the string will test ok for voltage but shorts when circuit is closed

[foo1bar]

SolarEdge has a troubleshooting guide - there is something on the inverter that helps pinpoint issues on a string, but it only helps if the installer correctly did the stringing in the layout, and I already know there are errors there

Been there.
Worst is when I only had myself to blame.
But for that I at least could just go up on my roof, cover up a panel with cardboard for ~30 min and figure out the right layout that way. Fortunately I didn't have that many panels (and with different orientation groups it was split into smaller sets to start with)

[ericy]

I guess the advantage of doing it yourself is if you keep good notes, that you know what the configuration looks like. I have to reverse-engineer it, bit by bit.

I have a few more datapoints - I went out to look more closely at the inverter, and the LCD screen is *really* hard to read. It looks to me like there is condensation in there which is fogging the window over the LCD display. So that's a new concern, which may tie into all of the rest of this.

Tried calling the solar company - left a message with the head of the outfit, but no callback as of yet. I am at the point where I am going to be hounding them regularly until I get some action from them. Now that I think about it, I might just drop by their office in person tomorrow and see if that gets me anywhere.

I turned off the inverter and went away for 10 minutes or so. When I came back, it only saw 19V. I have 39 panels, so I would have expected 39V. Not sure what this implies - broken string? I went through the re-pairing procedure when I turned it back on, and it brought me back to the same place I was before. I honestly didn't expect this to have any positive effect. Despite all of this, all but 3 optimizers are reporting status to the monitoring site.

The weather was cold, and I didn't feel like standing out there for all too long - plus I am trying to get work for my day job done. But I would like to open up the inverter and look for any sign of moisture in there.

The house design is a bit weird - there is no attic that runs from one side to the other, so in order for the installer to wire the thing, they needed to run a conduit through the crawlspace from one side of the house to the other. There are conduit LL boxes on each side of the house where it comes out from the crawlspace, and then the conduit goes up the walls to the roof. I would like to pop the covers off of those boxes, and make sure there is no sign of moisture in there either.

The one time I did get the installer out here (when we had a noreaster blowing), they were talking about taking some of these same initial steps (finding, opening and inspecting the junction boxes).

It may well be that there are multiple issues with this system.

I read through the troubleshooting guide from SolarEdge - it sounds like the most foolproof way of finding ground faults would involve having a special type of ohmmeter and some branch cables so you can take some of the optimizers out of the string. Sounds straightforward - in theory, but I don't have any of those things lying around.

[Ben25]

The safety voltage will only be an indicator of the longest string in series, not total optimizers. There are probably 2 strings connected in parallel.