Enphase S280 troubleshooting

[idnominal]

If you own the system, see if the installer will give you the log in information.

The companies I considered all said (both SolarEdge and enphase) that there was a lot of detail technical information that most customers do not want or understand. I explained that I might never change anything, or even look at it very often, if ever. However, I made access to those parts of software a condition of my purchase.

Reminds me of a whole home burglar alarm system on Long Island back in the 1990's. I got the set up login, and spent hours changing a few bits here and there to do some odd functions. Then, I never looked at it again. I suppose these systems are different, because as is the case here, when something goes wrong, there can be long lead times to get the experts out to look at it.

I remember somewhere in one of the online comparisons I looked at, I think there was a failure mode or two, where one bad enphase micro inverter can take out [disable, not damage] a string of micro inverters. Is that correct?

[andrewket]

I'm sure there are failure modes where one inverter can take out all a common circuit. I hope it's rare, as the distributed fault isolation is one of the value props of enphase.

[idnominal]

I did. No dice. My installer is coming out Friday. I've asked him to please share the current error that enphase is reporting to help point me in the right direction.

How did the repair go?

[andrewket]

Still not working. My installer was here are verified that the wiring was ok in the attic. They waited 2 hours on hold to speak with a tier 1 rep who couldn't help. Now we're waiting for a tier 2 rep to call back on Tuesday after the holiday.

[idnominal]

Bummer, tell us how it goes later. If you have an envoy-S, the name and password from the manual you were asking about last week are envoy and the last six of your ser. no. Even after you log in, I do not think there is much more to see there. There might be some interesting granular information from looking around the web, but it would take some programming to see it.

[andrewket]

Thanks, that u/p combo worked. But as it turns out, it just gives you access to the network diagnostic tools (ping, routes, etc.). The only other option that looks like it could be useful is "installer login", and that redirects to enphase. If I use my u/p for enphase's site, I get back an unauthorized / no permissions error.

I did experiment on my own a bit today. I removed my eGuage monitoring again just to double verify that it wasn't the cause. I also powered off all 3 circuits and power cycled the envoy, and then energized the three circuits on at a time. After 5-10 minutes, I'm right back to where I began: 14 inverters on the same circuit won't engage. The other inverters across the other two circuits came back. The envoy says it's communicating with all 38. I think it has to be a failed inverter, and unfortunately it failed in such a way as to take out the rest of the common circuit. Hopefully enphase can remotely disable the failed inverter so that the other 13 can work while we wait for a replacement. :\

[idnominal]

The S 280 manual talks about detecting DC resistance type faults, but that would probably only take out the one failure on the branch.

I suppose if there were some sort of AC failure in one module of either of the 240V legs to neutral or ground, that could possibly disable the whole branch. However, if you had a short on the AC side or very low resistance failure (or a cable problem, such as insulation failure, or pinched insulation ground either L1 or L2 of the 240V buss), that branch breaker would probably open, or there would be havoc across all three of the branches. Unless, the fault is detected so quickly that all branch inverters are down (or they do not even start) before the breaker can open. But, it couldn't be a hard short of L1 or L2 to ground/neutral anyway, because the panel would still be connected and the branch breaker would open ...

I wonder if there are any low level control failures, where some fault bit gets set and takes down all S 280s on a common branch. Do S-280s and/or the envoy "know" what micro inverters are on what branch?

hmm, but an open in either L1 or L2 or both (e.g. a broken section of wire) from the electrical panel to the first part of the branch (enphase recommends a "center" fed branch) would for sure take out an entire branch and probably not trip the breaker (because with either or both L1 or L2 unpowered, the micro-inverters probably do not start). A faulty breaker could probably cause this fault too, if one or both sides of the breaker of the failed branch are not actually closing electrically (even though the handle moves part or all of the way from off to on).

okay, so one easy test (if qualified, and not with a $2 plastic meter), with the branch breaker closed, do you see 120V to neutral on both terminals on the branch side of the breaker of the failed branch and 240V across the two terminals? (NOTE, if the breaker of the failed branch refuses to close electrically, there might actually be an AC SHORT and the breaker might be protecting as it should!)

Then, I guess that same test has to somehow be made where the same cable meets the failed branch at the roof (is there a junction box maybe?). If there are one or more junction boxes at or near the roof (or, anywhere between the panel breaker of the failed branch and the connection to that branch at the roof), an open L1 and/or L2 connection there too could take down a common branch.

It is hard to imagine a critter such as a squirrel chewing through #12 copper wire, but I guess that could also open L1 or L2 anywhere from the panel to the junction box/connection of the failed string.

I wonder if the S 280s can still communicate with the envoy if both L1 and L2 are open (e.g. does the envoy still "see" the S 280s of that branch with the branch breaker open). If not, then probably at least one of the L1 or L2 legs is still connected to the failed branch, and only one leg is open (if there is an AC open as part of this fault).

I'm thinking odds are (still just technical guessing) it's a cable fault / open (possibly a broken conductor, bad breaker, or a bad connection in a junction box at or anywhere between the panel and the roof) based on not seeing any S 280 enphase branch failures on the web (a S 280 fault that takes out the entire branch).

(all of this is no more than technical speculation, I have no training on the enphase systems, it might be better to leave all this to the installer, getting even a light shock on the roof and falling off would be bad!)

[bcroe]

hmm, but an open in either L1 or L2 or both (e.g. a broken section of wire) from the electrical panel to the first part of the branch (enphase recommends a "center" fed branch) would for sure take out an entire branch and probably not trip the breaker (because with either or both L1 or L2 unpowered, the micro-inverters probably do not start). A faulty breaker could probably cause this fault too, if one or both sides of the breaker of the failed branch are not actually closing electrically (even though the handle moves part or all of the way from off to on).

It still sounds to me like bad voltage to the 14 micros, more likely a simple open circuit. Which would be found by measuring
the inverter output terminals, as I said on 27 Aug. Bruce Roe

[idnominal]

It still sounds to me like bad voltage to the 14 micros, more likely a simple open circuit. Which would be found by measuring
the inverter output terminals, as I said on 27 Aug. Bruce Roe

Right, I agree with the open circuit, but it has to be (I think) before the first S 280 in the branch, otherwise one or more S 280s of that branch would still be alive. If the open thought is correct (leading to a low voltage as you said), the break in the wire or bad connection is between the panel and the connection to the branch on the roof. Do you think that's right?

Also, is there a way to probe the output terminals when the S 280 is connected to the branch? A S 280 will only power up when there is branch power. Oh, but you could disconnect and look at the cable side, not literally at the S 280 output terminals (unless they can be probed while connected).

I think the problem (an open if we are right) might show most easily at a junction box on the roof, probably where the enphase branch cable(s) begin.

[bcroe]

Get as close to the S280 output as possible. Not familiar here with micro AC busing, but perhaps unplug
one and either probe the bus connector or plug in a dummy cord and check voltage it delivers. Bruce Roe

[andrewket]

Not possible. The micros are all up on the roof. The attic is the closest we can get without climbing up there and unmounting a panel, which is likely what's going to happen anyway.

[bcroe]

Not possible. The micros are all up on the roof. The attic is the closest we can get without climbing up there
and unmounting a panel, which is likely what's going to happen anyway.

I suppose, the micros plug into a bus that is also on the roof? How badly do you want to fix it? This is already on
my list, of advantages to a ground mount. Bruce Roe

[idnominal]

The S 280 shuts down without bus voltage in limits. I see no advantage to initially trying to probe the S 280 output, which we know is presently disabled (from the envoy data). I think the next key step (which likely requires the installer to be done safely) is to verify the L1/L2 legs between the panel and the junction box which then connects to one or more enphase branch circuits (of 14 S280 max, which could be one bus or two split in the middle). There is nothing to measure at the output terminals of a disabled S 280 (it is "off") that cannot be first checked at the earlier junction box.

This is looking more like an installation problem than an ephase issue. It could be anything in the feed from the breaker itself (e.g. not actually closing one pole), the cable, and any number of one or more junction boxes between the panel and the junction box. If there is only one enphase branch, the open could be anywhere between the panel and the first S 280. If proper L1/L2 240V gets to even the first S 280, at least one of them should be on.

If we are right about the open circuit, it could be any number of faults, even something as stupid as a cable that got a large impact dent while still on the reel, and now is an intermittent gone open (perhaps made worse by pulling the cable during installation).

Some repair techs are probably too dependent on the computer error messages and diagnostics and have limited understanding of the S 280 operation. It is unclear why the installer needs Tier 2 support if he did not even check the connections yet and measure the L1/L2 voltages all the way out to the branch connections on the roof?

Of course, it could still be some weird fault, such as too long a run, or the cable out to the branch is too small a wire gauge (unlikely), or some weird data communications between the envoy and the branch and it just needs some kind of program reset/restart of the S 280s of that branch. It will be interesting to learn the answer!

Stay safe, wait for the installer.

[andrewket]

Keep in mind all 14 of the micros are communicating with the envoy via data over power line. So they are seeing grid voltage and are able to communicate over the circuit.

[andrewket]

Also, we did measure 120v on both legs in the junction box in the attic.

[andrewket]

My installer found "a short" in one of the junction boxes on the roof. Based on the description (I haven't seen a picture) it sounds like one of the insulators failed. No signs of water or physical damage. The only thing that comes to mind is possibly thermal failure. I am wondering if they ran the wrong gauge of wire. For the 3 days leading up to the failure, the array hit max output of the inverters. 14 inverters * 280w = 3920w (at 100% efficiency) or 16.3A at 240v. The breaker is 20A. Per code, max current allowed for continuous duty is 80%, or 16A. So it was right at the limit. Combined with the 90F+ temps, I wonder if the insulator melted? The breaker did NOT trip.