Help Needed - 327/SMA vs 250/Enphase for small roof w/ shadow elements

[kny]

A solar coop in my neighborhood put out an rfp for a bulk bid and ultimately a vendor was chosen based on the following bid:

1 - Sunpower 327 panels + SMA string inverter for $3.80/watt
2 - CanadianSolar 250 panels + Enphase multi-inverters for $3.15/watt

Here's my situation. I have a pretty small roof and want to maximize the kw as we currently have one electric car and will someday surely have a second. And, if I'm going to do this I want to do it right. We're looking about at 22-24 panels that can fit, so ballpark 5.5kw for the 250w system to 7.5kw for the 327w system. That's a big difference and all things being equal I would definitely go for the 327 panels and the larger system despite the $0.65/watt premium. Money is a factor, but not the key factor. Again, if I'm going to do this, I'm going to do it right, which means getting the most kw out of my roof as possible.

My roof is on two planes: one south facing and one west facing. Each plane has most panels that are perfect and unobstructed all day and some panels that will see passing shade throughout the day from a chimney, adjacent neighbor peak roof, or a/c units.

Questions:
1 - With the 327/SMA system I would have two strings, one for each roof plane. But, some panels on each of those planes get shaded modules at some point throughout the day. I understand that a single shaded panel brings down the performance of all panels on the string, but don't fully understand how much (and why). Does the entire system (string) perform as if all panels were shaded like the single shaded panel? If I had 10 panels on a string and 9 were operating at 100% and I threw a blanket on the 10th, what would I see from the string? 0%, 50%, 90%? I know this is complex and there is no simple answer, but I just can't get a sense of whether these temporary shadow elements from chimney, a/c units will bring down the performance of a 327w system on a string inverter to the level of a 250w system on a multi-inverter. This is really the crux of my decision; how badly does a partially shaded single panel impact the performance of the entire string.

2 - Are there other upside/downsides to these two options that I am not considering. I'm a data junkie and cannot imagine not being able to monitor my system on a per-panel level, so that's a big ding against the 327/SMA system. Other considerations?

3 - While it's not on the menu of options in the winning bid, if I were to convince the company to use the 327 panels with SolarEdge optimizers and SolarEdge inverter, would that be my ideal setup? Best of all worlds? How much premium over the 327/SMA offer would be reasonable? This video (https://www.youtube.com/watch?v=E8SiDDysDlU) mentions a 5200w cap on the SolarEdge inverter; that seems small and limiting. Is that per string, or per inverter?

Thanks for humoring this newcomer to the game.

[rwb1921]

This may not help much, but I noticed today that just a little shade on one panel (1 and 2 cells at the most shaded caused a drop of about 50% on that one panel (using Enphase and Solar World 285). Not sure how that would affect a string. I am sure someone here would know. I was surprised that just a little corner shaded would cause that big a drop. Mine is from a chimney.

[solar pete]

A solar coop in my neighborhood put out an rfp for a bulk bid and ultimately a vendor was chosen based on the following bid:

1 - Sunpower 327 panels + SMA string inverter for $3.80/watt
2 - CanadianSolar 250 panels + Enphase multi-inverters for $3.15/watt

Here's my situation. I have a pretty small roof and want to maximize the kw as we currently have one electric car and will someday surely have a second. And, if I'm going to do this I want to do it right. We're looking about at 22-24 panels that can fit, so ballpark 5.5kw for the 250w system to 7.5kw for the 327w system. That's a big difference and all things being equal I would definitely go for the 327 panels and the larger system despite the $0.65/watt premium. Money is a factor, but not the key factor. Again, if I'm going to do this, I'm going to do it right, which means getting the most kw out of my roof as possible.

My roof is on two planes: one south facing and one west facing. Each plane has most panels that are perfect and unobstructed all day and some panels that will see passing shade throughout the day from a chimney, adjacent neighbor peak roof, or a/c units.

Questions:
1 - With the 327/SMA system I would have two strings, one for each roof plane. But, some panels on each of those planes get shaded modules at some point throughout the day. I understand that a single shaded panel brings down the performance of all panels on the string, but don't fully understand how much (and why). Does the entire system (string) perform as if all panels were shaded like the single shaded panel? If I had 10 panels on a string and 9 were operating at 100% and I threw a blanket on the 10th, what would I see from the string? 0%, 50%, 90%? I know this is complex and there is no simple answer, but I just can't get a sense of whether these temporary shadow elements from chimney, a/c units will bring down the performance of a 327w system on a string inverter to the level of a 250w system on a multi-inverter. This is really the crux of my decision; how badly does a partially shaded single panel impact the performance of the entire string. Thereis no simple answer but it depends on how much shade on how many panels and for how long, and you cant answer that either can anybody else. Its no secret to the regulars here that I am not a big fan or micros but we do and will use them in situations like yours

2 - Are there other upside/downsides to these two options that I am not considering. I'm a data junkie and cannot imagine not being able to monitor my system on a per-panel level, so that's a big ding against the 327/SMA system. Other considerations? Sounds like you have answered your own question, if you want micro's go with micro's

3 - While it's not on the menu of options in the winning bid, if I were to convince the company to use the 327 panels with SolarEdge optimizers and SolarEdge inverter, would that be my ideal setup? Best of all worlds? How much premium over the 327/SMA offer would be reasonable? This video (https://www.youtube.com/watch?v=E8SiDDysDlU) mentions a 5200w cap on the SolarEdge inverter; that seems small and limiting. Is that per string, or per inverter? Tuff one, have no experience with SolarEdge but a 5000watt inverter generaly means thats it maximum capacity

Thanks for humoring this newcomer to the game.

Hello kny and welcome to the worlds best solar forum.

I have made some comments above, cheers.

One thing to remember, try not to put panels in the shade cause if they are on a string or a micro it dont really matter cause a panel in the shade dont make nothin

[bcroe]

This may not help much, but I noticed today that just a little shade on one panel (1 and 2 cells at the most shaded caused a drop of about 50% on that one panel (using Enphase and Solar World 285). Not sure how that would affect a string. I am sure someone here would know. I was surprised that just a little corner shaded would cause that big a drop. Mine is from a chimney.

My panels are divided into 3 sections that can be bypassed in shade, etc. Even one cell not
keeping up will cause the whole section to be bypassed with no production. In a string, all
the rest of the string cells will shift to a less efficient operating point (higher voltage), and
the entire string output will be affected. Bruce Roe

[kny]

Here is the roof layout (more or less; more panels will fit than shown):

http://i.imgur.com/fyxQfBW.png

Here is morning shadow (9am) from chimney on south facing roof (also can see higher neighbor peak that casts evening shadow on same roof):

https://lh4.googleusercontent.com/-tWXNslNg07g/VbY7A_XI-PI/AAAAAAAA3rA/Ebh2ZYmpf4Y/w1114-h630-no/IMAG0906.jpg

Here is morning shadow (9am) from vent and satellite dish, plus a/c units on west facing roof.

https://lh4.googleusercontent.com/-8TLk939Va-0/VbY7Iwr-lXI/AAAAAAAA3sI/JGzmVlki83c/w1114-h630-no/IMAG0915.jpg

Here is late afternoon / evening shadow (5:30pm) starting to encroach on south facing roof:

https://lh6.googleusercontent.com/-wRNV49TL7tg/VbafHn5kR-I/AAAAAAAA3tQ/oNGp7b0f7ng/w1114-h630-no/IMAG0921.jpg

Here is late afternoon / evening shadow (5:30pm) from a/c unit on west facing roof (ignore roof hatch cover):

https://lh4.googleusercontent.com/-vjnvRV-lB1w/VbafIhNUBpI/AAAAAAAA3tU/2pmb9mJ0x0w/w1114-h630-no/IMAG0922.jpg

I guess there's two shade factors:
1 - shading a few cells on a panel impacts the production of an entire panel by how much?
2 - lessened production on a single panel impacts production of an entire string by how much?

Based on Sunpower product video, one claim they make is that their panels mitigate 1 better than others. True or not, I do not know.
To best mitigate 2, the options are micro or power optimizers. My option, based on the winning vendors bid, is 250 w/ micro or 327 w/ string inverter. What I can't figure out is whether small bits of shade at various times throughout the day will make the 327 w/ string produce net less than 250 w/ micro. It does seem that 327 w/ SE power optimizers and SE inverter is the best scenario, but that is not necessarily an option without going outside this coop's bulk bid winner.

[sensij]

I guess there's two shade factors:
1 - shading a few cells on a panel impacts the production of an entire panel by how much?
2 - lessened production on a single panel impacts production of an entire string by how much?

You may have already picked up your answers from other threads, but just in case:

(1) Talking non-sunpower here... It depends on which cells. There are three bypass diodes that would each allow 20 cells to be bypassed (assuming 60 cell panel), so shading a single cell can be enough to lose 1/3rd the panel. If the next cell shaded is part of that 1/3rd, no harm done. If it is part of the panel that was still working, you can lose another 1/3rd. Shade the entire row of cells along the short edge of the panel, and you can wipe out the entire panel.

Sunpower handles shade differently, and doesn't use bypass diodes like other panel companies do. They have literature that explains it better, it sounds like you have seen some of it.

(2) On a single string, a panel that is shaded will not meaningfully hurt the rest of the panels, if the bypass diodes are working as designed. If you have two strings in parallel, and only one is shaded, you will end up with some collateral losses because the voltage of the good string will get pulled away from the maximum power point as some middle ground is found between the shaded and unshaded string.

FWIW, if it was me, and you can get enough of the 250 W panels on the roof to meet your goals, I would choose the Enphase system (assuming they are using new M215's or M250's not old M190's). The failure risk is higher than with the string inverter, but your roof looks flat enough to work on and none of the panels appear to be boxed in, so they will be accessible when service is eventually required.

[kny]

You may have already picked up your answers from other threads, but just in case:

(1) Talking non-sunpower here... It depends on which cells. There are three bypass diodes that would each allow 20 cells to be bypassed (assuming 60 cell panel), so shading a single cell can be enough to lose 1/3rd the panel. If the next cell shaded is part of that 1/3rd, no harm done. If it is part of the panel that was still working, you can lose another 1/3rd. Shade the entire row of cells along the short edge of the panel, and you can wipe out the entire panel.

Sunpower handles shade differently, and doesn't use bypass diodes like other panel companies do. They have literature that explains it better, it sounds like you have seen some of it.

(2) On a single string, a panel that is shaded will not meaningfully hurt the rest of the panels, if the bypass diodes are working as designed. If you have two strings in parallel, and only one is shaded, you will end up with some collateral losses because the voltage of the good string will get pulled away from the maximum power point as some middle ground is found between the shaded and unshaded string.

FWIW, if it was me, and you can get enough of the 250 W panels on the roof to meet your goals, I would choose the Enphase system (assuming they are using new M215's or M250's not old M190's). The failure risk is higher than with the string inverter, but your roof looks flat enough to work on and none of the panels appear to be boxed in, so they will be accessible when service is eventually required.

Thank you for your response. The bypass diode info on non-SP panels was new information for me, so thank you. On #2 above, when you say that a single shaded panel on a string will not impact the remainder of the string, does that assume multi-inverters or power optimizers? Because my understanding was that with a standard string inverter system (like the SP + SMA option I've been presented) when one panel on the string was compromised that this impacted the production of the entire string, the justification/technology behind this I still do not understand.

My goal is to maximize kw on my roof, within reason. And my small roof just got smaller as I've just now been told the fire code requires 3' from ridges and edges, and as my roof is already broken into two separate small roofs, this restriction just made everything quite a bit smaller. I had thought we would get about 22 panels in place, but now it looks like 15 is gonna be about max. So, likely winning configurations will be either SP-327 + SE400 optimizers + SE5200 inverter or SP-327 + SMA.

Again, thanks for your input.

[kny]

Ok, I'm going with 22 Sunpower E-327 panels with SolarEdge P400 optimizers for a 7.194kw system. They are estimating 9100kwh annually from this. Does that seem reasonable? How can I model/predict this to confirm based on the specifics of my location, roof, etc....?

And, they have me specced for a SolarEdge SE6000A inverter, which has a max input of 6000 watts. Given my system is a 7.2kw system, am I being set up for clipping? My roof is on two planes (see photos above), so at no point in the day will all 22 panels be at full capacity at the same time, so maybe a 6000 watt max input inverter is fine?

Sunpower E-327 + SolarEdge P400 optimizers + SolarEdge SE6000A at $3.90/watt. Washington, DC. Good?

[sensij]

Ok, I'm going with 22 Sunpower E-327 panels with SolarEdge P400 optimizers for a 7.194kw system. They are estimating 9100kwh annually from this. Does that seem reasonable? How can I model/predict this to confirm based on the specifics of my location, roof, etc....?

And, they have me specced for a SolarEdge SE6000A inverter, which has a max input of 6000 watts. Given my system is a 7.2kw system, am I being set up for clipping? My roof is on two planes (see photos above), so at no point in the day will all 22 panels be at full capacity at the same time, so maybe a 6000 watt max input inverter is fine?

Sunpower E-327 + SolarEdge P400 optimizers + SolarEdge SE6000A at $3.90/watt. Washington, DC. Good?

7194 / 6000 = 1.20, which suggests a very small amount a clipping in a south facing array, maybe on the order of 1% of annual production or less. On more than one roof, you might get a decent idea of how much is clipping by running PVWatts for each orientation, then adding the hourly output values. For example, if 11 panels face south and 11 panels face west, run PVWatts for each orientation with a 3.597 kW array, and then add the hourly output together. Correct for clipping by including another column that looks for any values over 6000 W AC in the summed output and sets those values to 6000 W. Check the difference in annual output between the corrected and uncorrected columns... it will probably be a small number.

When running PVWatts, selecting "Premium", "Roof mount", and 8% loss should be a decent starting point. Your shade will affect the totals a bit, so don't read too much into that, but the model of midday power (when you are most likely to clip) should be better since you will hopefully be shade free.

[kny]

7194 / 6000 = 1.20, which suggests a very small amount a clipping in a south facing array, maybe on the order of 1% of annual production or less. On more than one roof, you might get a decent idea of how much is clipping by running PVWatts for each orientation, then adding the hourly output values. For example, if 11 panels face south and 11 panels face west, run PVWatts for each orientation with a 3.597 kW array, and then add the hourly output together. Correct for clipping by including another column that looks for any values over 6000 W AC in the summed output and sets those values to 6000 W. Check the difference in annual output between the corrected and uncorrected columns... it will probably be a small number.

When running PVWatts, selecting "Premium", "Roof mount", and 8% loss should be a decent starting point. Your shade will affect the totals a bit, so don't read too much into that, but the model of midday power (when you are most likely to clip) should be better since you will hopefully be shade free.

Thanks much. I did this as precisely as possible, and indeed, out of the entire year there are a whopping 4 hours where I will exceed the 6kw AC output max of the SE6000 with the max of 6.18kw.

Raw data here, with separate tabs for analysis by day, month, and hour of day.

https://docs.google.com/spreadsheets/d/1NUhrGHMVcRo-PErVbRc3-Mro8bQl-mnNZeQ7sCcKyeA/edit?usp=sharing

Questions on pvwatts usage:
1. Does it just randomly assign shady days based on historical weather?
2. Please explain DC to AC ratio parameter. I entered 1.35 due to the 8100w Max DC Power (STC) to 6000w AC output max, but that resulted in pvwatts capping the calculated AC output at 5333.33w many hours. Putting the ratio back to the default of 1.1 and my max AC output values were not capped. So, not sure I understand this parameter .

Thanks again.

[sensij]

It looks like you got the idea, and your results sound right.

Questions on pvwatts usage:
1. Does it just randomly assign shady days based on historical weather?

The weather used is based on what is known as a TMY (typical meteorological year). Gory details here, page 1 has a nice overview though.

2. Please explain DC to AC ratio parameter. I entered 1.35 due to the 8100w Max DC Power (STC) to 6000w AC output max, but that resulted in pvwatts capping the calculated AC output at 5333.33w many hours. Putting the ratio back to the default of 1.1 and my max AC output values were not capped. So, not sure I understand this parameter .
Thanks again.

This is actually a way to test clipping loss if your entire array could be simulated at one orientation. For example, if you have a 7000 W of panels and a 6000 W inverter, you would enter 7000/6000 = 1.167. This will limit the AC output to 6000 W, and you can see the energy produced with this constraint. Then, you could enter 1.0 and see how much energy would have been produced with an inverter that could have handled 7000 W. In your example, with a 7194 W array, 7194 / 1.35 = 5329, which is roughly the AC output cap you observed.

[kny]

It looks like you got the idea, and your results sound right.



The weather used is based on what is known as a TMY (typical meteorological year). Gory details here, page 1 has a nice overview though.



This is actually a way to test clipping loss if your entire array could be simulated at one orientation. For example, if you have a 7000 W of panels and a 6000 W inverter, you would enter 7000/6000 = 1.167. This will limit the AC output to 6000 W, and you can see the energy produced with this constraint. Then, you could enter 1.0 and see how much energy would have been produced with an inverter that could have handled 7000 W. In your example, with a 7194 W array, 7194 / 1.35 = 5329, which is roughly the AC output cap you observed.

Got it. Much appreciated. pvwatts is an impressive modeling tool, especially for a site that ends in .gov.

[J.P.M.]

Got it. Much appreciated. pvwatts is an impressive modeling tool, especially for a site that ends in .gov.

Easy to learn and gets what most folks are looking for with min. of input. Wait til you get to SAM. Suggest remembering it's all a SWAG. Read the help info screens if you have not yet done so for limitations, or ways to tweak the input.