My Solaredge system dead. How to figure if it's an optimizer or inverter that's bad?

[ButchDeal]

Too bad you didn't just go with micro inverters...None of this would have happened, and if you lost one, your whole system wouldn't be down.

This is less then helpful. Had the installer installed micro inverters as poorly as they installed the SolarEdge system, it likely would be in the exact same situation.
Further as Sensij pointed out, micros really were not an option.

[sensij]

Too bad you didn't just go with micro inverters...None of this would have happened, and if you lost one, your whole system wouldn't be down.

You can read more about the system planning in this thread:

https://www.solarpaneltalk.com/forum...hadow-elements

If there were any micro-inverters available that were compatible with those panels, maybe it would have been an option, but there weren't (and still aren't, as far as I know).

[Six4KilowWatt]

You should be under warranty. So with a few months you ought to have a warrantied replacement Look on the bright side, There isn't a helluva lot you can do. Once the new board works or doesn't work. A new inverter should be issued.

Too bad you didn't just go with micro inverters...None of this would have happened, and if you lost one, your whole system wouldn't be down.

[sensij]

Hard to fathom it being anything but a problem internal to the inverter.

Yeah, I'm not saying it isn't the inverter, just trying to offer suggestions on how to rule out everything else.

I totally agree that trying to replace parts in the inverter is ridiculous, they should swap in a new one and repair the other at their convenience.

[kny]

I'm out the door not to return until Monday, so further assessment must wait until then.

Keep in mind, however, two separate wire runs - one 10 AWG, one 8AWG - from inverter to panel make no difference. Next door neighbor has solar and is operating fine, so highly unlikely a grid voltage issue.

Hard to fathom it being anything but a problem internal to the inverter.

[sensij]

The 10-15 W ac power on the screen *before* the clicks seems like a bad sign. I think I've only ever seen *0* until after the clicking is done.

[kny]

The click click is the last thing that happens before power is released (relays actuating, I guess). AC too high immediately would mean the power getting released has no where to go (running into high impedance), if all systems are healthy.

Do you own a volt meter?

Is the Curb fast enough that if you moved it to the inverter, it could measure voltage at those terminals instead of the service panel?

The Curb ain't moving. I do own a multimeter.

[sensij]

The click click is the last thing that happens before power is released (relays actuating, I guess). AC too high immediately would mean the power getting released has no where to go (running into high impedance), if all systems are healthy.

Do you own a volt meter?

Is the Curb fast enough that if you moved it to the inverter, it could measure voltage at those terminals instead of the service panel?

[kny]

I have just replaced the wiring from the inverter to ac disconnect to service panel and upgraded from 10AWG to 8AWG. Breaker also upgraded from 30 amp to 40 amp.

No change in behavior. Inverter goes through 5 minute waking up period, starts finding optimizers, mini screen reads 243 AC Volts, 410 DC Volts, hovers around 10-15 AC Power, goes click-click-click-click and then poof - AC VOLTAGE TOO HIGH

[sensij]

One more observation... The Curb (installed at the service panel) was showing clipping at 5900 W. Clipping is confirmed by the high DC voltage at that time. If it is accurate, AC transmission loss could be found by comparing that number to what the inverter says it was putting out. For example, if the inverter was clipping at 6000 W, that is 1.7% transmission loss.

Until someone tries to determine if the AC spikes are real or just imagined by the inverter, I don't see the OP's situation getting better any time soon.

[sensij]

The voltage on a simple string will be high when no current is being drawn. But the output of the individual optimzers should be only 1V DC when they are not in communication with the inverter.
And when the inverter is in control, I would not necessarily expect it to let the string voltage rise that high just because it is not willing to produce an AC output.

The DC voltage is allowed to rise to well above operating levels during the startup cycle / grid qualification stage. I've seen my inverter hang at that high DC voltage with no output when there was an abnormal shutdown that eventually required re-pairing the optimizers before successfully restarting. Although the optimizers return to 1 V as expected when the inverter is disconnected from the grid or they are asked to shut down by flipping the inverter's toggle switch, there seems to be at least one point in the state diagram in which they are allowed to run high.

After a quick look through my data, I can see another state in which DC voltage is allowed to rise is when the inverter goes into clipping, which may also explain some of what the OP observed in the first post in this thread.

[inetdog]

It would be miraculously coincidental that after 18 months the utility grid voltage increased and increased permanently at the exact moment the PV system was handling its all-time load high of 6700 watts, and it was at that moment that the AC VOLTAGE TOO HIGH error appeared never to go away.

I've checked the wiring connections at the breaker, the ac disconnect and within the inverter. All looks good.

FWIW, the voltage at the grid tie inverter terminals will be greater than the grid voltage at your POCO transformer secondary by an amount which is proportional to the current being delivered and the resistance of the wiring and connections, including the service drop and the transformer windings. So first seeing the error when at maximum production is reasonable. What is not reasonable is that it also come up every time the inverter starts, including while the sunlight on the panel array is still low in the morning.

Possibly something happened to greatly increase the circuit resistance.
One simple test to conduct would be to turn off the inverter and then temporarily connect a load, like a heat gun, hot plate, or portable electric heater at the inverter AC connection point and looking at the voltage drop. It would take an electrician to wire that up if you do not have a receptacle installed right now.

[inetdog]

The high DC voltage is typical when power is not being produced, even on a healthy inverter. I'm not sure I'd agree that it suggests a common root cause. Thanks for sharing your experience, though...

The voltage on a simple string will be high when no current is being drawn. But the output of the individual optimzers should be only 1V DC when they are not in communication with the inverter.
And when the inverter is in control, I would not necessarily expect it to let the string voltage rise that high just because it is not willing to produce an AC output.

[solardreamer]

Experience thus far with SolarEdge is terrible. I'm now down for nearly a month. Took 2 weeks for my installer to get P400s to swap out the incorrect P300s. But, of course, that solved nothing. SolarEdge is supposed to send out a new "communications board" under RMA, but it's been 2 weeks and still nothing shipped. And I have no doubt this will fail to solve the problem and I will wait more weeks for an inverter replacement to ship out under RMA.

I'm pretty close to simply buying a new inverter and installing myself.

Does SolarEdge think your problem is likely due to a bad communications board?

[sensij]

I've checked the wiring connections at the breaker, the ac disconnect and within the inverter. All looks good.

Hopefully there isn't a bad spliced connection hidden in the conduit somewhere. Again, based on what you've shared of this installers work, it is hard to rule anything out.

If it is shutting down almost immediately now, has anyone tried taking high speed voltage readings at the inverter to at least see if the voltage error is real? A regular multimeter may not be fast enough, and SolarEdge might not be able to do that remotely. If the voltage does spike at the inverter, does it also spike at the service panel? If the voltage does not actually spike at the inverter, you can have more confidence that inverter repair/replacement will help.