New owner, trying to figure out this clipping deal

[Alan1296]

Hello all. I've had my grid tied system running about one month here in Missouri. We have a ground mount system, 36 panels, 12.96kW. The installer went with a Solar Edge SE10000H inverter. I questioned the choice after the fact, wondering why an 11400H inverter wouldn't of been better. He assured me the 10KW was the best one for my system. I've seen DIY kits with 12.96kW selling online with both inverters, so not sure when one makes since over the other.

Today we had our best output so far, nearly 74kWh produced. Looking at the days production things start heating up at 6:45 am and by 10:15 the system is clipping. The output remains maxed at 10kW till about 2:45, so about 4.5 hours of clipping. By 5:30pm the day is done. Today was a bright sunny but cool day, mid 40's for a high.

When I look at this a couple questions come to mind - could I had went with 32 panels, a 11.52kW system, and still be producing this same power? Or with the 11400H would I have produced roughly 6 more kWh's then I did with the current set up? (1.4kWh for 4.5 hours) It makes you wonder if your getting the best bang for your buck when you have these questions. Can I get some feedback on your thoughts of my systems setup? Thanks!

[J.P.M.]

If you had gone with, or if you now reduce down to 32 panels, you would lose less to clipping but your year long output would be less overall. It's a question of a half empty or half full glass - either overpaneled or under invertered.

Provided you don't have a lot of shading, to get a VERY ROUGH estimate of annual loss to clipping with your current inverter and array, do the following:

1.) Get the max. design power output for the inverter. Unless I'm missing something (at least possible if not likely), the spec sheet for the SE 10000H inverter says the max. AC output is 10 kW which seems to agree with what you report for max. output.
2.) Run PVWatts with a 10 % system loss parameter.
3.) Get the hourly output option.
4.) For hours where the PVWatts modeled hourly output is > the max. design inverter output (which looks like that 10 kW), subtract the PVWatts modeled hourly output from the max. design inverter output ( a negative #), otherwise enter a zero for than hour.
5.) Sum all the hourly entries from # 4 above. The result will be a very rough estimate of annual loss to clipping, probably good to +/- 10% or 20 % or so.

If you get tempted to compare the PVWatts output to what you have generated so far - don't. The results will be misleading and not representative or meaningful.

What's your zip and array orientation ? Any shading ? Without that information, it's hard to provide any meaningful #'s as to how much clipping you may be experiencing over a year. Even with that, it'll be a SWAG.

[discodanman45]

With the difference in pricing of about $200 for the 11400H, it was not a wise decision IMO for a 130% overclock. If you do max out at 10 kW for 3 hours a day, you are not only missing out on electricity but could be shortening the life of the inverter running at high temps. My installer decided to put 11.25 kW on a 7600 inverter and I was stressing out the inverter 4 hours a day. I would monitor your inverter and watch for clipping. If it is something that is everyday, I would contact back your installer and asked if they used PVWatts or the SolarEdge Designer to estimate how much clipping would happen per year. I would also see if you can get the 25 year warranty on the inverter, because if clipping is common that inverter will have a much shorter life.

[J.P.M.]

With the difference in pricing of about $200 for the 11400H, it was not a wise decision IMO for a 130% overclock. If you do max out at 10 kW for 3 hours a day, you are not only missing out on electricity but could be shortening the life of the inverter running at high temps. My installer decided to put 11.25 kW on a 7600 inverter and I was stressing out the inverter 4 hours a day. I would monitor your inverter and watch for clipping. If it is something that is everyday, I would contact back your installer and asked if they used PVWatts or the SolarEdge Designer to estimate how much clipping would happen per year. I would also see if you can get the 25 year warranty on the inverter, because if clipping is common that inverter will have a much shorter life.

Pretty much agree, but I'd comment that peddlers and installers routinely oversize systems. To my experience, the amount of oversize is often something like 10 % or so. If so, that'll reduce any clipping they come up with by a substantial amount. I'd run PVWatts on my own and verify what a vendor is saying.

[littleharbor]

The cold temperatures added to your output. Could it also be that your roof pitch is optimum for this time of year? Summer peak output could be lower, midday, than you are seeing now.

[Alan1296]

With the difference in pricing of about $200 for the 11400H, it was not a wise decision IMO for a 130% overclock. If you do max out at 10 kW for 3 hours a day, you are not only missing out on electricity but could be shortening the life of the inverter running at high temps. My installer decided to put 11.25 kW on a 7600 inverter and I was stressing out the inverter 4 hours a day. I would monitor your inverter and watch for clipping. If it is something that is everyday, I would contact back your installer and asked if they used PVWatts or the SolarEdge Designer to estimate how much clipping would happen per year. I would also see if you can get the 25 year warranty on the inverter, because if clipping is common that inverter will have a much shorter life.

Thank you for the reply. He did say they size all units through SolarEdge and they did extend the inverter warranty to 25 years already.

[Alan1296]

The cold temperatures added to your output. Could it also be that your roof pitch is optimum for this time of year? Summer peak output could be lower, midday, than you are seeing now.

Thanks for the reply. We have a ground mount system, so no roof pitch. But yes I did read that the panels actually perform better on these cool sunny days then they will in the heat of the summer.

[Alan1296]

If you had gone with, or if you now reduce down to 32 panels, you would lose less to clipping but your year long output would be less overall. It's a question of a half empty or half full glass - either overpaneled or under invertered.

Provided you don't have a lot of shading, to get a VERY ROUGH estimate of annual loss to clipping with your current inverter and array, do the following:

1.) Get the max. design power output for the inverter. Unless I'm missing something (at least possible if not likely), the spec sheet for the SE 10000H inverter says the max. AC output is 10 kW which seems to agree with what you report for max. output.
2.) Run PVWatts with a 10 % system loss parameter.
3.) Get the hourly output option.
4.) For hours where the PVWatts modeled hourly output is > the max. design inverter output (which looks like that 10 kW), subtract the PVWatts modeled hourly output from the max. design inverter output ( a negative #), otherwise enter a zero for than hour.
5.) Sum all the hourly entries from # 4 above. The result will be a very rough estimate of annual loss to clipping, probably good to +/- 10% or 20 % or so.

If you get tempted to compare the PVWatts output to what you have generated so far - don't. The results will be misleading and not representative or meaningful.

What's your zip and array orientation ? Any shading ? Without that information, it's hard to provide any meaningful #'s as to how much clipping you may be experiencing over a year. Even with that, it'll be a SWAG.

Thanks for the detailed reply. Zip is 65757. Array is due south. We may have a little shade in the morning but for the most part no shade. They solar company even paid to remove several trees from a neighbors yard to make this system work. (with her permission of course) lol. Removing panels at this point isn't an option, but a larger inverter or a separate additional inverter could be feasible.

I'll need a little help with PVWatts. I've been to the website before so familiar with it but not sure everything is correct. My parameters are -

12.96
Standard
Fixed (open rack)
10% loss (this number was like 14% but I lowered to 10%)
Tilt 30 (unsure)
Azimuth 180

Under advanced parameters -
DC to AC size 1.3
Inverter efficiency 96% (I just left the number alone, seems high maybe)
Ground coverage ratio .4

This gives me 19,457kWh per year. Sounds a little high. Solar company estimated 18,566. I'm looking at the hourly report that seems to go on forever, seeing some DC array output numbers over 10,000w but all AC outputs under 10,000w, which would seem to make sense with the inverter max. I'm just not sure what numbers to subtract from what.

[ButchDeal]

Hello all. I've had my grid tied system running about one month here in Missouri. We have a ground mount system, 36 panels, 12.96kW. The installer went with a Solar Edge SE10000H inverter. I questioned the choice after the fact, wondering why an 11400H inverter wouldn't of been better. He assured me the 10KW was the best one for my system. I've seen DIY kits with 12.96kW selling online with both inverters, so not sure when one makes since over the other.

Today we had our best output so far, nearly 74kWh produced. Looking at the days production things start heating up at 6:45 am and by 10:15 the system is clipping. The output remains maxed at 10kW till about 2:45, so about 4.5 hours of clipping. By 5:30pm the day is done. Today was a bright sunny but cool day, mid 40's for a high.

When I look at this a couple questions come to mind - could I had went with 32 panels, a 11.52kW system, and still be producing this same power? Or with the 11400H would I have produced roughly 6 more kWh's then I did with the current set up? (1.4kWh for 4.5 hours) It makes you wonder if your getting the best bang for your buck when you have these questions. Can I get some feedback on your thoughts of my systems setup? Thanks!

Well if you look at your daily curves, it wouldn't have been 1.4kW more production solid for that 4.5 hours. In fact it may not have even peaked out at 11.4kW

Personally I find a few of the sizes odd and worthless to install.
The black balled inverters are the SE3000, SE6000, and SE10000.
There is just very little reason to use these 3 inverters. The next size up is very little cost difference, has same size breaker and two of them have more options.
The SE3800H and SE7600H both have EV charger options.

Some distributors don't stock these as well so orders take longer as well.

[ButchDeal]

With the difference in pricing of about $200 for the 11400H, it was not a wise decision IMO for a 130% overclock. If you do max out at 10 kW for 3 hours a day, you are not only missing out on electricity but could be shortening the life of the inverter running at high temps. My installer decided to put 11.25 kW on a 7600 inverter and I was stressing out the inverter 4 hours a day.

11.3kW on a 7.6kW is not a bad thing.
It depends on the layout of the system. For example an East/West mix of 11.3kW would not clip on a 7.6kW in most locations.
Also another consideration is that any inverter larger than the 7.6kW would require MSP derating or lineside tap which can add cost. If the array is situation such that the clipping is minimal, it might not warrant the added expense to avoid.

Yet another consideration is that the 7.6kW has the most other options with optional StorEdge battery backup or EV charger.

The OPs situation is a bit different with SE10000h vs SE11400h the features and connection requirements are nearly identical.

[discodanman45]

12.96
Standard - Use premium - CoP of almost all panels these days have numbers close to premium
Fixed (open rack)
10% loss (this number was like 14% but I lowered to 10%)
Tilt 30 (unsure)
Azimuth 180

Under advanced parameters -
DC to AC size 1.3 - Use 1 -1.3 is what your poorly choosen inverter uses
Inverter efficiency 96% (I just left the number alone, seems high maybe) - SolarEdge Inverters are rated at 99%
Ground coverage ratio .4

This gives me 19,457kWh per year. Sounds a little high. Solar company estimated 18,566. I'm looking at the hourly report that seems to go on forever, seeing some DC array output numbers over 10,000w but all AC outputs under 10,000w, which would seem to make sense with the inverter max. I'm just not sure what numbers to subtract from what.

I ran your numbers and got 250 kWh of clipping with 373 hours of overclocking your inverter. I did my own 7600 inverter with 11.25 kW, before I got another inverter added, and it was 500 kWh. This number is even low for me, I am already producing numbers that are far outweighing any suggested clipping for February. I think that 250 kW is a conservative number and you will see a few hours of clipping on most full sun days.

[discodanman45]

11.3kW on a 7.6kW is not a bad thing.
It depends on the layout of the system. For example an East/West mix of 11.3kW would not clip on a 7.6kW in most locations.

In my area(Chowchilla, CA) I ran PVWatts and I would have clipped almost 400 kWh with a 11.25 kW 50/50 E/W arrays with a 7.6 kW system. This may even be a low number from the simulations I used for my system and the clipping I produced last year. Not only do you have to worry about clipping, but straining the inverter 600 hours a year is not a good thing as well. I am probably in one of the best areas for solar production, so I agree most areas won't see the clipping I get. However, if a company has a DC/AC ratio above 1.2, they better justify it to the consumer.

[ButchDeal]

In my area(Chowchilla, CA) I ran PVWatts and I would have clipped almost 400 kWh with a 11.25 kW 50/50 E/W arrays with a 7.6 kW system. This may even be a low number from the simulations I used for my system and the clipping I produced last year. Not only do you have to worry about clipping, but straining the inverter 600 hours a year is not a good thing as well. I am probably in one of the best areas for solar production, so I agree most areas won't see the clipping I get. However, if a company has a DC/AC ratio above 1.2, they better justify it to the consumer.

I should have specified, it depends on the pitch of the east/west arrays. low pitch would clip, high pitch not so much.

I agree very much there should not be a static DC/AC ratio but designed systems. We modeled everything and tried to keep clipping to under 1% unless there were design constraints like limits on AC feed in....

[J.P.M.]

Thanks for the detailed reply. Zip is 65757. Array is due south. We may have a little shade in the morning but for the most part no shade. They solar company even paid to remove several trees from a neighbors yard to make this system work. (with her permission of course) lol. Removing panels at this point isn't an option, but a larger inverter or a separate additional inverter could be feasible.

I'll need a little help with PVWatts. I've been to the website before so familiar with it but not sure everything is correct. My parameters are -

12.96
Standard
Fixed (open rack)
10% loss (this number was like 14% but I lowered to 10%)
Tilt 30 (unsure)
Azimuth 180

Under advanced parameters -
DC to AC size 1.3
Inverter efficiency 96% (I just left the number alone, seems high maybe)
Ground coverage ratio .4

This gives me 19,457kWh per year. Sounds a little high. Solar company estimated 18,566. I'm looking at the hourly report that seems to go on forever, seeing some DC array output numbers over 10,000w but all AC outputs under 10,000w, which would seem to make sense with the inverter max. I'm just not sure what numbers to subtract from what.

Understood.
1.) I got 19,768 kWh/yr. using your #'s except for 0.975 for the inverter efficiency as the published CEC efficiency on the spec sheet. Looks like your company didn't underestimate as much as some others, ~ 1- (18,566/19768) = ~ 6.1% low.
2.) Apologies, as there's a couple of fine points I left out on using PVWatts:
- Again, using this method is a loose approximation of possible clipping and may indicate when an inverter may be undersized. It's actually a sort of bastardization of the purpose of the model but can be useful. Just don't take any hour's output as gospel. Sum all the hourly values. Any hour's output by itself is pretty useless, particularly for this purpose.
- Yes, the spreadsheet has 8,760 + rows, one for each hour of a 24 hour * 365 day year.
- If all the hourly DC outputs from the modeled array are < (10 kW/inverter eff.) = 10,000/0.975 = 10,256 W, that would point to (but no smoking gun) the inverter, apart from any shortcomings, maybe not being undersized. If any of the DC hourly outputs are greater than that 10,256 W then, then, by the model (only), any array DC output that is > 10,256 W for any hour will be clipped. If so, then to the amount of the total annual clipping, the inverter is undersized.

So, for each hour, subtract that hour's DC output from 10,256 W. Then, take the sum of all such negative values for each hour. That sum (of all the negative values) will be a VERY ROUGH indication that clipping is going on at some point and a probably a rougher yet SWAG of an average of how much clipping is likely to occur per year over many years within maybe a +/- 10% - 20% window. BTW, as the array gets older, the clipping will tend to be less as the array's efficiency drops off a bit. How much ??? Only time will tell. FWIW, early on (<6 yrs.), I'm losing ~ 0.3% /yr. Your results will be different.

I've done a PVWatts run based on your supplied #'s but with an inverter eff. of 0.975.

From that, it looks to me that perhaps the model (and I stress the MODEL only) over a "typical meteorological year ", thinks the inverter on your system will clip some of the array's output approx. 200 hrs. The clipping might amount to something like 143 kWh/yr. or so total over a "typical meteorological year" (out of that 19,768 kWh/yr. of a "typical meteorological year's" output). The model thinks your array's max. hourly output will be 13,230 kW. That may be possible or a model burp. If t's a burp, the model thinks there are that hour's and two other hour's outputs (out of what the model thinks are 4,319 positive hourly outputs) that are greater than 12.96 kW. Possible I suppose and not a big concern in the larger sense.

To SWAG the cost of that 143 kWh/yr. for a real rough estimate, multiply 143 by your ave. cost/kWh.

To perhaps overstress, these results are not much more than a dart throw based on what's actually a misapplication of the PVWatts model. The model's output is also f(input) which is seldom absolutely accurate. But given what seems to be the model's general usefulness as a design tool, some general direction may be gleaned by using it in the way described above. Still, any system's annual output for any 365 day period will most certainly be different than the model's, if for no other reason than the weather the array sees will be different from the weather the model uses. Over many years, and if the input to the model is a reasonable representation of the array, the long term array output will probably not be too much different than the model's output.

[discodanman45]

I've done a PVWatts run based on your supplied #'s but with an inverter eff. of 0.975.

I believe the older SE10000A Inverters were 97.5% CEC Weighted Efficiency. The SE10000H inverters are 99% efficient.