Sanity Checking + Advice on SolarEdge

[ButchDeal]

His initial rationale was that he divided my home into three "zones" and was pairing one inverter with each zone. Presented with my (and Bruce's) arguments for the benefits of eliminating an inverter, he re-ran the numbers and agreed that using just two would be fine. .

Well a common mistake for people more familiar with string inverters. With optimized systems you want to have some modules from each azimuth or "zone" so that it can keep for longer during the day.

[Ne0eN]

That is my one concern with my 56 roof mount panels. I had 2 DOAs out of 80 inveters. As luck would have it, one was in the top row and the other in the bottom row, but I'm sure the day will come when I need to deal with an inverter in one of the 2 middle rows. It's a non issue for my ground mount array.

How exactly do these optimizers behave when they are DOA? Do they still put out 1V when connected to a panel? I'm about to connect 35 panels up and would want to know immediately when I come across a dead inverter before finishing the install. I could just have a guy monitoring the string voltage as the panels are connected...

[Magius]

How exactly do these optimizers behave when they are DOA?

This was one of my biggest questions, and the more research I've done it seems there are anecdotes claiming just about every failure mode you can imagine. I haven't saved links, but I have read one guy's thread where he claims an optimizer died and took down the entire string, and another guy's thread where he claims an optimizer died and the power simply routed around that panel, delivering 100% of the expected output of the other panels. I'm sure there are probably scenarios in which either of these are possible, though the best I can gather from SE they seem to claim that the former shouldn't be able to happen..?

What you propose sounds like a good sanity check, make sure it puts out the 1V safety voltage when connected to a panel, but I'm no expert. I don't know if that guarantees that it would function correctly when connected to the inverter and "turned on" for full voltage? Might be interesting to know, but hey, that's my installer's problem :P

[cebury]

More anecdotal. One optimizer of mine was performing intermittently, perhaps a rare fault, I don't know. But it would work normally then stop working at random times in the day. When it stopped, it reported zero voltage. When installer came to test, it was malfunctioning luckily, and showed zero volts when disconnect from the array. However, even when it went bad, the string appeared to,operate just fine. All other optimizers worked fine and total kwh output was only reduced by the one panel output, at least as shown via the SE site. Now ymmv based on number of panels per string and how total voltage is affected, eto. I'm not the one to answer that.

[emartin00]

In most cases when an optimizer fails, the rest of the string shouldn't be affected. I would imagine if there was some internal wiring problem, it could cause the entire string to open, but it's not very likely.

[Magius]

I thought I'd come back and update how this turned out with my installer. He did agree with my request, based on the advice above, to swap two of the 6kW inverters for an 11.4kW. He also credited me back $560 or so (I can't remember exactly now) for the change. So the plan was going to be an 11.4 kW and a 6kW, supporting 57x LG320 modules (18.24 kW DC)

We signed the papers and my project then went to engineering, but when it came back they'd changed the setup to two 7.6kW inverters. I questioned the change, since it reduced my AC output capacity significantly. Given my multiple azimuths I don't think it's any real problem in terms of oversizing, but I still wanted to understand why it changed from what we agreed to. It turns out that they use a device called a ConnectDER to attach to the utility meter, and this device is limited to an 80A input breaker and 15kW of AC input. The two 7.6kW inverters with up to 32A output each max out both the 80A breaker and the ConnectDER device. I was told that to go any higher than that, they have to use a different way of connecting to the utility, which requires ripping out my brand new electrical panel, and would cost an extra $1000 or so.

Since I don't expect to clip the 7.6kW inverters, as long as they string up the modules intelligently, it wasn't worth it to me to redo the panel to support an 11.4kW and larger breaker. As such, I agreed to their inverter change, and requested another credit back, since they're obviously saving some BOM cost vs. what we agreed to. I estimated the cost savings to be around $325, and they offered me $280, so I figure that's close enough.

My project is now going to the city for permits, and hopefully they'll start the installation in 7-10 days.

On the plus side, if my state ever cancels net metering, at least the 7.6kW inverters leave me set up nicely for a Powerwall... lol

[aleenoor]

I thought I'd come back and update how this turned out with my installer.

.....

My project is now going to the city for permits, and hopefully they'll start the installation in 7-10 days.

On the plus side, if my state ever cancels net metering, at least the 7.6kW inverters leave me set up nicely for a Powerwall... lol

I Magius, is your install complete ? How did it turn out ? Do you clip at all ? Does SE system tell you if you clip ?

I am also considering very similar equipment albeit about 11.7 kW DC to avoid million dollar insurance here in FL.
I am considering either SE 2 x HD-Wave 5k (when available) + 300W Q cell or 300W Canadian panels. OR for low cost option with KACO TL inverters.

I wanted to ask you how your experience has been with SE and LG panels?

[aleenoor]

Magius, I also wanted to ask how much of a % drop in PV output/panel are you seeing between you S facing vs W facing array (on 6:12) slope.
I am limited on S facing (6:12) space by 38 (60c) panels or 30 (72c) Panels. If I use a string inverter , I have to push at least 25% of my panels to West facing. I wanted to gauge how much of a output loss will I be looking at.

[sensij]

Magius, I also wanted to ask how much of a % drop in PV output/panel are you seeing between you S facing vs W facing array (on 6:12) slope.
I am limited on S facing (6:12) space by 38 (60c) panels or 30 (72c) Panels. If I use a string inverter , I have to push at least 25% of my panels to West facing. I wanted to gauge how much of a output loss will I be looking at.

You can get some idea of the production difference of the different orientations available to you by modeling each roof face in PVWatts.

[Magius]

Aleenoor,
Thank you for your interest in my installation, and I'd love to help answer your questions!
To kick things off, here are links to my public SolarEdge monitoring portal, as well as my PVOutput page. You might be able to get a few answers of your own browsing around those pages.
https://monitoringpublic.solaredge.c...Marc_Lake_Nona

http://pvoutput.org/list.jsp?id=54006&sid=50534

Now to address some questions directly:
Yes, my installation is complete, it turned out fantastic, and I'm loving it! SE + LG has turned out to be a great combination, very well matched and efficient from what I can tell. I'm very happy that I did not go with my original plan of LG + Enphase, as I can see now that I would have been clipping much more frequently than I'd initially suspected.

That said, it's hard to tell if I've actually "clipped" (maxed out) my panels or optimizers, as the portal doesn't explicitly say "optimizer #3 is clipping". What I can say is that I've seen individual south-facing panels putting out instantaneous readings of 330-335W on a few perfect days. So that's a 320W nameplate panel (-0/+5% tolerance means it could be up to a 336W STC panel) going through the 320W optimizer, but producing ~335W for short periods of time. Was the optimizer clipped at 335W? or was there just not enough solar energy on the panel to get an extra watt or two out of it? it's really hard to know that answer. I can also say for sure that I have not yet clipped my 7.6kW inverters, though I have seen them both produce more than 7.6kW output. The highest combined instantaneous output I've seen to date is ~15.75kW, but they're rater to about 8.4kW each, so I have a little headroom left there. The system is well balanced such that both inverters are usually under very similar loads, rather than one being busy in the morning and the other in the afternoon, or one producing 1.5kW more than the other at any given time. They're within 0.5 kW of each other at all times of day from what I've seen.

Regarding the million dollar insurance, unfortunately if you're on FPL I've heard there's no way around that. I'm on OUC so I don't have to get the million dollar policy. OUC is hugely friendly to solar installations, to the point that up until a few months ago, they were giving all solar customers a $0.05/KwH production credit on top of net metering, even if you consumed the energy on premises. Unfortunately that program ended shortly before my system went live, but OUC has been great to work with in every way and thankfully they don't use the insurance hammer to try to force systems to arbitrarily smaller sizes...

Lastly, regarding my South vs. West production, I should clarify that my "West" is actually slightly south-west, and my "East" is actually slightly north-east, however I do have some data that I produced for another friend that I'll copy-paste below. He has a solar system composed of due South and due East arrays, and he asked me to compare to make sure he and I were seeing similar efficiency drops on our non-south arrays. His system is SolarWorld panels with Enphase micros, and his house is ~55 miles away from me, however the directionality comparisons still came out exactly as we'd expected. Here's that data:

[Magius]

[Deleted accidental duplicate post ... also continued original post which got cut off in the middle due to certain symbols?]

On my array, month to date, Ive gotten ~23.25kWh from each South panel, ~18.25kWh (78.5%) from each East (theyre slightly north-east) and ~19.5kWh (84%) from each West (theyre slightly south-west)
On your array, month to date, youve gotten ~19.5kWh from each South panel, and ~15.75kWh (81%) from each East. That seems right where it should be. Your south panels are making 25% more than your east, or your east are making 20% less than your south, whichever way you prefer to say it. And your due east output is right between my slightly-south-west and my slightly-north-east, so it all makes sense.

You can also use the PVWatts site to compare production of identical arrays pointed South vs. West, but from the data I pasted above, you should expect your west-facing panels to be like my friend's east-facing panels, and produce ~81% of the power produced by your south-facing panels. This all assumes zero shading, as neither mine nor my friends installation has any shading.
I hope this is useful info for you, please let me know if I can be of any further assistance.
Thanks,
Marc

[sensij]

That said, it's hard to tell if I've actually "clipped" (maxed out) my panels or optimizers, as the portal doesn't explicitly say "optimizer #3 is clipping". What I can say is that I've seen individual south-facing panels putting out instantaneous readings of 330-335W on a few perfect days. So that's a 320W nameplate panel (-0/+5% tolerance means it could be up to a 336W STC panel) going through the 320W optimizer, but producing ~335W for short periods of time. Was the optimizer clipped at 335W? or was there just not enough solar energy on the panel to get an extra watt or two out of it? it's really hard to know that answer. I can also say for sure that I have not yet clipped my 7.6kW inverters, though I have seen them both produce more than 7.6kW output. The highest combined instantaneous output I've seen to date is ~15.75kW, but they're rater to about 8.4kW each, so I have a little headroom left there.

A few things...

Optimizers won't clip under any real world conditions. Their power rating is not a limit like it is with the inverters.

The solar panel's STC rating is not a limit, it is just the panel's output under a specific set of test conditions. It is capable of producing more than that if the actual conditions are more favorable than those test conditions. A quick scan through your data suggests that the times in which you've approached or exceeded STC are on cool, partly cloudy days. The temperature helps... STC is at 25 deg, and panels will produce more when cooler. Some clouds can help, since sunlight will reflect on the edges of the clouds, so the array can get a share of sunlight that would have otherwise hit the ground somewhere else if the clouds weren't there.

On your PVOutput setup, hopefully you've set the "Energy" to disabled (I just compared PVOutput to the portal data, and see that it is). Because of how SolarEdge reports energy, the value tends to be inflated by about 4%. Reports based on power are more accurate, usually within 1% of a revenue grade meter.

The 7600 W inverter will only produce 7600 W continuously. Note the output current rating of 32 A... 32 A * 240 V = 7680. In very dynamic conditions, the Solaredge 15 minute reporting can get wonky. It is likely that anytime you see a power in excess of 15.2 kW, it will have points much lower on at least one of the readings immediately before or after. You really need a fast responding grade revenue meter to get accurate instantaneous data in those conditions, although the inverter's integrated power over hourly timeframes tends to be OK.

[aleenoor]

You can get some idea of the production difference of the different orientations available to you by modeling each roof face in PVWatts.

Thanks Sensji, I did run various scenarios on PV Watts and saw 15% drop on West Facing (6:12 pitch) and 12.5% drop on East facing (6:12) pitch. I was interested in Marc's real world readings as some of his arrays are almost exact tilt/azimuth as mine.

Thanks for sharing the insight regarding the SE Optimizers caps and setting up PVOutput with SE.

[aleenoor]

Marc,
I really appreciate you detailed reply. I think I had already added your setup to my favs on PVOutput today. My test setup is HF1 (renamed to 'Windermere test' today).

I have Duke energy here in Dr. Phillips/Windermere and I don't expect them to show me any love for going solar.

I have been doing a lot of research into my setup and am almost ready to pull the trigger. I might just be over analyzing it at this time. However, reviewing your outputs will put a lot of my questions to rest.

In my case, a couple of good college friends back in OH are in the Solar Installation business and I could source good quality Panels/Inverters and Racking for very cheap (about a dollar a watt for everything). They will handle any equipment warranty stuff as well. That coupled with the fact that I am an Electrical Engineer, I am trying to see if a good installer will be willing to properly permit and install the setup for me for reasonable labor cost.

I have a relatively simple setup with only 2 arrays all south (for 60 cell) or 3 array S, S , W for 72c modules.

1 big decision is SE or String. SE will cost me about $4000~ where as KACO Transformerless String Inverters 2x 5.0 KW will cost me $1400 for 2. I will have a total of 4 MPPTs. Web connected , monitoring, self consumption and all.

http://kaco-newenergy.com/products/solar-pv-inverters/residential-pv-systems/blueplanet-3-5-kva/

Thanks again for your offer to help. I might hit you up with some targeted question.

[sensij]

Marc,

1 big decision is SE or String. SE will cost me about $4000~ where as KACO Transformerless String Inverters 2x 5.0 KW will cost me $1400 for 2. I will have a total of 4 MPPTs. Web connected , monitoring, self consumption and all.

http://kaco-newenergy.com/products/s...lanet-3-5-kva/

If you are still on 2011 NEC, string inverters are hard to resist. On the system I will be putting on my house, the cost of the rapid shutdown system ate most of the difference in price, so I'm going to spend a few hundred more for Solaredge to keep the design and installation simple. Unless you've got serious shade or other design problems, I'd go with the Kaco Inverters. I have a couple years experience with Solaredge already, and while there is enough to like that I'm ok using it again, the system is not without its flaws.

BTW, is $700 for a transformerless 5 kW inverter a publicly available price? The cost of re-submitting my permit application might be prohibitive, but I'd have gone that way with my design if I could buy a current model inverter for <$1k. The only place I see it listed has the 5 kW for about $2k.