proper string inverter sizing

[charles2]

For my 2 Renesola subarrays composed of 310W modules, I want to use a single SMA Sunny Boy TL-US-22 series inverter that has the "Secure Power Supply" feature. I will have 13 modules facing mostly south and 10 modules facing mostly west. What is the proper size for the inverter?

[sensij]

SMA has an online design tool and a downloadable spreadsheet that may help.

If I were designing it, I would do something like this, consistent with what you'd get from one of the design tools (I checked).

If your panels are these, look at the following:

@STC
Isc = 8.80 A
Voc = 45.0 V
Vmp = 37.0 V
Temp coef of Voc = -0.30%/deg C

A string of 13 will have a Voc of 585 Vdc and a Vmp of 481 Vdc @ STC.
A string of 10 will have a Voc of 450 Vdc and a Vmp of 370 Vdc @ STC.

Athens, GA should use -11 deg C as the minimum design temp. [(-11) - 25] * -0.3%/deg C = 10.8%.

Temp corrected Voc for a string of 13 = 585 * 1.108 = 648 V.
Temp corrected Voc for a string of 10 = 450 * 1.108 = 499 V.

This series of inverter has a maximum Voc of 600 V. So, 13 panel string is no good, but the 10 panel string is OK. 12 panel string would be OK too, if you can drop a panel from the big array.

If you can't drop a panel, splitting it into a 6 and 7 panel string is a bad idea, they need to be equal to minimize losses. Therefore, add a panel and make it two strings of 7. VOC @ STC of a 7 panel string is 350 V, safely above the minimum string voltage of 125 V.

So, now you either have 22 or 24 panels. 24 * 310 = 7440 W, 22 * 310 = 6820 W. The SB6000TL-US-22 would be a good choice for either of those combinations, especially since the two arrays won't hit peak power at the same time due to the difference in their orientation. It has two MPPT inputs so put each sub-array on its own input and you'll be fine.

[charles2]

Thanks, that's good info.

Tell me, what happens when an inverter is oversized or undersized? I'm guessing that an oversized inverter clips array output when it is weak, and an undersized inverter either is destroyed or clips excessive array output. Am I correct?

[sensij]

Thanks, that's good info.

Tell me, what happens when an inverter is oversized or undersized? I'm guessing that an oversized inverter clips array output when it is weak, and an undersized inverter either is destroyed or clips excessive array output. Am I correct?

If undersized relative to the DC power available from the panels, the inverter will limit its output to something close to its AC rating. On the other end, except at the extremes, for the most part having an inverter too big is just money that you could have spent on something else.

[charles2]

So there is no clipping on the low end, and you cannot destroy an undersized inverter?

[sensij]

So there is no clipping on the low end, and you cannot destroy an undersized inverter?

I'm sure you can come up with extreme cases that can have a bad result. However, when the inverter is modestly oversized or undersized... for example, when the ratio of the DC panel rating to AC inverter output rating is between 0.67 - 1.35, different ratios are just going to have small differences in efficiency.

You can see a typical efficiency curve for an SMA inverter in their datasheet. As long as the inverter is putting out at least 20% of its rating, the efficiency is pretty good, and it doesn't really fall off the cliff until you get down to the 5% range.

[charles2]

Would you run thru the minimum and maximum string sizing calculations for Suniva OPT325-72-4-100 modules and the SMA Sunny Boy SB 7700TL-US-22 inverter? I'm hoping it's OK to put 13 of those modules on one string and 14 on the other.

[sensij]

They have a 46.3 Voc. (Datasheet)

14 * 46.3 = 648 V. Way too much.
13 * 46.3 = 602 V. Too much.

Assuming the temp correction used in the previous post was correct, the maximum number of those panels allowed in a string on that inverter is 11.

[charles2]

I guess I don't understand why SMA even makes a 7.7kW inverter if it can only handle 22 panels?

[ButchDeal]

I guess I don't understand why SMA even makes a 7.7kW inverter if it can only handle 22 panels?

It can handle more modules, you just need a combiner box with fuses.

[sensij]

It can handle more modules, you just need a combiner box with fuses.

You don't even need the fuses or a J-box, in some cases. For example, a 28 panel system using the 72 cell panels mentioned above might have two strings of 9 combined in parallel on one MPPT input (using MC4 Y-connectors), and a 10 panel string on the other input. Fuses aren't required until at least three strings are being combined on a single input.

The design constraints for strings are tighter than they are for microinverters or optimizer systems, which is one of their disadvantages, but workable designs often exist.

[charles2]

The design constraints for strings are tighter than they are for microinverters or optimizer systems, which is one of their disadvantages, but workable designs often exist.

So what's the best way to set up these 27 Suniva panels with the 7700TL-US-22? Should I drop a panel to get an even number, or what?

[inetdog]

So what's the best way to set up these 27 Suniva panels with the 7700TL-US-22? Should I drop a panel to get an even number, or what?

You do not need an even number. You just make one string with an odd number (like 9) leaving an even number (18) which you then divide by two.
The -22 model has two MPPT inputs, so you do not need a fused combiner. Just put one string on one input and the other two strings in parallel on the other input.
If you had 26 or 28 the situation would not be any better or worse. With 26 you can go for one string of 8 and two strings of 9 or one string of 10 and two strings of 8 if you want to distribute the power more equally between the two MPPT inputs.

[sensij]

I think the problem is 13 panels are on one roof face, and 14 are on the other. Splitting them into parallel strings doesn't work great. A string of 6 * 37.0 Vmp = 222 Vmp, which is too low for this inverter (low MPPT range is 270 V). With the 72 cell panels the OP picked out, it looks like strings should be in lengths of 8 to 11. 60 cell panels would have more flexibility, but may not use the available space on the roof as efficiently. The inverter operates below the MPPT range down to 125 Vdc, but I'm not sure of the consequences for being outside of the range.

[ButchDeal]

I think the problem is 13 panels are on one roof face, and 14 are on the other. Splitting them into parallel strings doesn't work great. A string of 6 * 37.0 Vmp = 222 Vmp, which is too low for this inverter (low MPPT range is 270 V). With the 72 cell panels the OP picked out, it looks like strings should be in lengths of 8 to 11. 60 cell panels would have more flexibility, but may not use the available space on the roof as efficiently. The inverter operates below the MPPT range down to 125 Vdc, but I'm not sure of the consequences for being outside of the range.

One of the many reasons why I have change our company base inverter to SolarEdge. This design is trivial with SolarEdge and optimal performance for the layout.