Benefits of Solaredge vs. string inverter for non-shaded

[inetdog]

I see +5W/-0W as the power tolerance a lot more often than i see +5%/-0%. What panels are you looking at? Even with that tolerance, the chance that two entire arrays would be built on opposite ends of the tolerance isn't the most likely explanation for the observed power output difference, in my opinion.

Statistics tells us that for randomly selected panels an individual panel variation of 5% will lead to a much smaller variability in an array.
if you have 9 panels in an array the likely variation would be roughly 5%/3, or 1.66%

[solarix]

I kind of quit using SolarEdge because of all the problems, but as far as I know the new versions are pretty reliable.
Now that SMA's new string inverters are quite good at multiple orientations, I rarely use SolarEdge. Better to do whatever it takes to get full sun than to mitigate shade with a band-aid....

[sensij]

http://www.lg.com/us/commercial/sola...lg-LG310N1C-G4 is +3% / 0-%

http://www.sunpreme.com/wp-content/u...le-Rev-1.3.pdf is +5% / -3%.

What's a +5W / -0W example?

Look at panels made by Solarworld, Canadian Solar, and Q-Cells, to name a few of the checked before I posted.

[DaveDE2]

On a system that large, and assuming say 310W panels, you're talking about 80 or so panels on each roof face. You're probably going to have to remove one or two panels to get to a bad optimizer when it fails, something to consider maintenance-wise. May or may not be a big deal depending on how steep the roof is. With 160 panels and optimizers it will probably happen on the order of every two years or so (assuming optimizer 300 year MTBF), again, may or may not be a big deal. Alternatively, with only diodes on the roof, a failure will be less likely to ever happen over 20-30 years or so. Just saying.

[DanKegel]

Out of the 30 or so SolarEdge systems I installed back in the 2010 time frame, I've had 5 inverter failures (two of which literally blew up) and at least 12 Powerbox failures.

Good data. Do you have any data on more recent SolarEdge equipment reliability?

[rsilvers]

http://us.sunpower.com/sites/sunpower/files/media-library/data-sheets/ds-e20-series-327-residential-solar-panels.pdf

[rsilvers]

http://www.lg.com/us/commercial/sola...lg-LG310N1C-G4 is +3% / 0-%

http://www.sunpreme.com/wp-content/u...le-Rev-1.3.pdf is +5% / -3%.

What's a +5W / -0W example?

Here is a +-5%:

http://us.sunpower.com/sites/sunpower/files/media-library/data-sheets/ds-e20-series-435-solar-panel-datasheet.pdf

Here is a +5% - 0%:

http://us.sunpower.com/sites/sunpower/files/media-library/data-sheets/ds-e20-series-327-residential-solar-panels.pdf

[J.P.M.]

My two identical arrays were originally one 24 panel system running on a 5kW Xantrex inverter. Trina 230w panels purchased all together on the same pallet. Because I'm an installer, I had a customer wanting to upgrade from a 3.3kW system to a 5kW system and through a whole series of events I ended up trading him my 5kW for his 3.3kW Xantrex and splitting my array - putting a SolarEdge on the other half as I was curious to see just how good it was. I did try swapping the arrays and saw no difference. Also (as a result of all the failures we experienced with the early SolarEdge version) I had a SolarEdge rep come visit us and showed him my head to head comparison and he could not do much more than smile. I think the conversion losses from multiple individual optimizers are higher than in one central string inverter and these losses blow any gaines made from optimization. On my two systems, the SolarEdge does better at first when the sun comes up, then through the day - the Xantrex gets ahead, and in late afternoon the Solaredge tries to catch up but never quite gets there. Anyway, the difference is slight and I can tell you for sure that optimizer systems are not worth it unless you have serious shading issues. On my system, the SolarEdge inverter has been reliable (so far) but have had to replace two of the Powerboxes (now they call them optimizers). Out of the 30 or so SolarEdge systems I installed back in the 2010 time frame, I've had 5 inverter failures (two of which literally blew up) and at least 12 Powerbox failures. All this was done under warranty, but of course - they did not compensate us for the time involved, not to mention the considerable customer loss of confidence. I really like the SolarEdge design concept - but they just did not quite get the manufacturing quality needed.

This is good data, and pretty valuable as a real world comparison of actual output. There may be things encountered in the future in either the methodology, the equipment or the data that may cause some reexamination of the results, but that does not retract from the value of what's been done so far. Whether or not everyone agrees with the conclusions doesn't matter too much. What's important is the data and its quality. Interpretation of what it all means followed by improvements are what counts. There's nothing like real world data that's well gathered and documented.

FWIW, IMO, nicely done.

[solarix]

My two identical arrays were originally one 24 panel system running on a 5kW Xantrex inverter. Trina 230w panels purchased all together on the same pallet. Because I'm an installer, I had a customer wanting to upgrade from a 3.3kW system to a 5kW system and through a whole series of events I ended up trading him my 5kW for his 3.3kW Xantrex and splitting my array - putting a SolarEdge on the other half as I was curious to see just how good it was. I did try swapping the arrays and saw no difference. Also (as a result of all the failures we experienced with the early SolarEdge version) I had a SolarEdge rep come visit us and showed him my head to head comparison and he could not do much more than smile. I think the conversion losses from multiple individual optimizers are higher than in one central string inverter and these losses blow any gaines made from optimization. On my two systems, the SolarEdge does better at first when the sun comes up, then through the day - the Xantrex gets ahead, and in late afternoon the Solaredge tries to catch up but never quite gets there. Anyway, the difference is slight and I can tell you for sure that optimizer systems are not worth it unless you have serious shading issues. On my system, the SolarEdge inverter has been reliable (so far) but have had to replace two of the Powerboxes (now they call them optimizers). Out of the 30 or so SolarEdge systems I installed back in the 2010 time frame, I've had 5 inverter failures (two of which literally blew up) and at least 12 Powerbox failures. All this was done under warranty, but of course - they did not compensate us for the time involved, not to mention the considerable customer loss of confidence. I really like the SolarEdge design concept - but they just did not quite get the manufacturing quality needed.

[J.P.M.]

Thanks for your thoughts.

You're welcome.

[foo1bar]

Vandalism - fenced or not.
Critters chewing on things.
Put some $$ in the budget for roundup and payroll $$ to apply it. It'll help keep things looking pretty.

2nd story roof - so not too worried about vandalism or roundup (if the weeds get that high, I'll be looking around for a guy named Jack)
I'm not expecting much for critters either - I think it's too far from edge of the roof for squirrels to jump to a tree - and wouldn't be a spot I think they'd find attractive nor find reason to be traversing.

Keep spares. Keep them locked up.

Probably won't happen...
$500 for spares just in case two panels fail or live with 1% less output if two panels fail...
I think the choice will be 1% less output.

Monitor input. Even if not required, still a good idea. Check utility requirements (if any) and if allowed, get Davis Pro II+ or a Li-Cor pyranometer, both w/ data logging, and maintain them. The Davis will set you back ~~ $1K +/- some, and comes with enough instrumentation for performance estimates. The Licor costs a bit more to get all the bells/whistles for data logging and only measures irradiance, but it's a bit more professional.

Interesting - but I wouldn't know how we'd use it. Nor how it'd be any better than just comparing different string's outputs or comparing to other installations nearby.

On cleaning, I'd think long/hard about a cleaning schedule and more so about who does it. Without some controls on when and who cleans the array, I'd plan on some well intentioned but solar ignorant person(s) hitting an array with a hose when the panel glass temp. is ~ 60 C. or so. Just sayin'.

Good thing to think about.

I haven't had longer than a 7 week spell w/ no rain to confirm/refute. A decent rain seems to restore ~~ 1/2 to 2/3 or so of the performance lost to fouling. I usually suggest to people as a practical matter that they add ~ 3% to the array size to account for performance degradation and hit the array with a hose if it hasn't rained in 3 - 4 weeks.

5-6 months without rain is fairly common here.
I don't think we'll be cleaning more often than ~3 months... We'll have to see who will go up on the roof.

Thanks for your thoughts.

[J.P.M.]

I'm not anticipating much in maintenance and service.
Any issues that are related to warranty (ex. inverter fails) I expect the installer to handle.
I would guess with 10+ inverters that there will be one infant mortality in the first 2 years.

The only other thing I expect is that occasionally the panels may benefit from being washed. (It often won't rain from May until Oct.)
I expect that to be done by facility staff. (Probably will use a hose and something like the "Mr. Clean AutoDry Car Wash System" for the final rinse)
And at that time I'd expect they'd look for any wires that had slipped out of the cable management.

What other things do you think we should anticipate?

Since you ask:

Vandalism - fenced or not.
Critters chewing on things.
Put some $$ in the budget for roundup and payroll $$ to apply it. It'll help keep things looking pretty.
Keep spares. Keep them locked up.
Keep people informed of performance. They'll bitch less when they see it's reducing the bills.

Monitor input. Even if not required, still a good idea. Check utility requirements (if any) and if allowed, get Davis Pro II+ or a Li-Cor pyranometer, both w/ data logging, and maintain them. The Davis will set you back ~~ $1K +/- some, and comes with enough instrumentation for performance estimates. The Licor costs a bit more to get all the bells/whistles for data logging and only measures irradiance, but it's a bit more professional.

On cleaning, I'd think long/hard about a cleaning schedule and more so about who does it. Without some controls on when and who cleans the array, I'd plan on some well intentioned but solar ignorant person(s) hitting an array with a hose when the panel glass temp. is ~ 60 C. or so. Just sayin'.

The folks at the Furnace Creek Resort ( Death Valley) clean their array (1.23 mW) every other year as a guano wash only. They monitor performance fairly well (Li-cor + Davis) and tell me that fouling does not seem to increase (and thus performance decrease) after a short initiation period of desert dust buildup.The array near me that feeds the pumps for the Valley Center H2O district (1.1 mW) is in plain sight to me as I write this. It has yet to see any cleaning other than rain in the 7+ yrs. it's been on line that I've seen.

Keeping track of my array's output (and input), suggest to me that, without rain, my array fouls at a rate that causes, very approximately, and very non linearly, ~~ 1% performance decrease per week. I suspect that without rain, that rate might becomes asymptotic and level off @ ~ 7 - 8 % or so after about 7 weeks or so, sort of in agreement with the Death Valley observations, but I haven't had longer than a 7 week spell w/ no rain to confirm/refute. A decent rain seems to restore ~~ 1/2 to 2/3 or so of the performance lost to fouling. I usually suggest to people as a practical matter that they add ~ 3% to the array size to account for performance degradation and hit the array with a hose if it hasn't rained in 3 - 4 weeks.

Like most things of a solar utilization nature, this is not an exact science.

[DanKegel]

http://www.lg.com/us/commercial/sola...lg-LG310N1C-G4 is +3% / 0-%

http://www.sunpreme.com/wp-content/u...le-Rev-1.3.pdf is +5% / -3%.

What's a +5W / -0W example?

[sensij]

His word has nothing to do with it. Panels are +5% -0%. One array could be 5% more output than the other even if they are all brand new and identical.

I see +5W/-0W as the power tolerance a lot more often than i see +5%/-0%. What panels are you looking at? Even with that tolerance, the chance that two entire arrays would be built on opposite ends of the tolerance isn't the most likely explanation for the observed power output difference, in my opinion.

[rsilvers]

I'd take him at his word that the arrays are identical in the most obvious ways.

His word has nothing to do with it. Panels are +5% -0%. One array could be 5% more output than the other even if they are all brand new and identical.