I have a 24 panel KC120-1 system currenly running an upgraded ST2500 inverter. Having read about 20% better power generation from Sunny Boy inverters, I'm thinking about switching. My house only has room to mount 16 of the panels facing south. 4 of the panels are facing southwest and 4 are facing west. The sma website discourages such installations. Does anyone have experience with a mixed orientation installation with Sunny Boy? How does the system perform?

Um , the 20% figure was for an Suntie or Suntie XR vs a sunnyboy NOT vs an Sunite UPG !!!


The Upgraded unit is within 0-2% of a Sunnyboy and if you have a single panel that is 2% or less than the others in the string , the entire sunnyboy yeild will drop since the panels are in series (limited by the weakest performing panel)

There is a list of reasons the Suntie is actually better than the Sunnyboy , have a read of

Suntie Upgrade White Paper

And to answer you question directly , if you mixed panels as you decribed the Sunnyboy will underperform the Sunite , most likey in the 30-50% range since the series current will be different due to the differnt isonaltions (sun angles) , maybe even more and will be limited by the lowest output panel (well at some point the bypass diode will conduct) but you get the idea

[ 09-21-2003: Message edited by: Solar Guppy ]

Thanks for the info.

Just to simplify this a bit for me, if I had ten panels in series and I nearly or completely covered one panel what would be the effect be to the system? What if 5 faced south and 5 west? Thanks in advance for your reply.

-John

To ask in a more direct way, if one of 10 panels is shaded, is the loss in output from the system total greater than just the 10% surface that is shaded? Thanks.

I have a SunTie inverter with 24 KC120 solar panels. They are installed with groups of 4 in series. I'm no electrical engineer, but I did some experimenting to see how shading affects the output of a group of 4 panels. This can easily be done by hooking a digital ammeter up to one of arrays of 4 and using a piece of cardboard to shade panels.

On the day I was doing the experiment, one group of 4 panels was putting out about 6 amps in full sun. If I covered just one cell of one KC120 panel (each panel has 36 cells) it would cut the output by an amp down to about 5 amps. I discovered that a single KC120 is divided into 2 independent strings of 18 cells, each of the 18 cell strings has independent bypass diodes. If you lay a KC120 with it's long side down, the top half (2 rows of 9 cells) and bottom half (2 rows of 9 cells) behave independently. When I shade just one cell, it knocks out an entire half of the output from one panel. This can easily be proved by shading all 18 cells in the bottom (or top) half of a panel. You still see 5 amps of current! As long as you're shading cells in top or bottom half of a KC120, once you shade one, it's the same as shading all 18 cells.

Now if you shade one cell in the top half of panel, and one cell from the bottom half of a panel, it's a whole different story. The cell shading on the bottom knocks out the bottom 18 cells and the cell shaded on the top knocks out the top 18 cells. Which means you lose the whole panel! The amperage for the whole string of 4 drops from 6 amps to less than 1 amp. That's with just a tiny amount of shade in the wrong place!

The fact that the inverter could even draw one amp out what has become 3 panels in series is only possible if the inverter drops the operating voltage of the entire array. This drop in voltage means your 4 panel strings are operating at non-optimal voltages which is going to lower the amount of power generated by the whole array. So a tiny amount of shade can knock out a disproportionate amount of power generating capability of your system - much more than the 1/24th it seems like you should lose!

In your case, with 10 panels in series you need to pay attention to what levels of voltage you need for your system to produce power. If you are near the lower voltage threshold for your inverter to generate power, losing just one out of 10 panels could have a disproportionate effect on system power output. Note, that these effects are worse early and late in the day when the sun is weak. Most certainly during the times of weak sun, shading just one panel could be enough to completely shut down power generation of your whole system!

Helios

If the panels are in series , and a panel is shaded , the strings combined voltage will drop by the panels voltage under no load.

What JV is saying is that just because the bypass diodes kick in, when hooked up to an inverter other issues come into play.

In JV (and my systems) we have other strings of panels in parallel and the inverter will be tracking to the MPPT of the other strings. Since only 3 of 4 panels in the shaded string are generating a voltage , the MPPT for the 3 remaining panels will be MUCH lower than what the inverter is tracking to.

The current measurements show this as the current drops significantly is the remaining panels in the shaded string are being sources much closer to VOC than VMP

In the SINGLE series string (YOUR QUESTION), a single panel dropping out will only lose that panels generation IF AN ONLY IF

1) The remaining panels combined VMP is above the minimum operational input for the inverter

2) The Inverter quickly re-tracks MPPT when the shading occurs

3) The String is a SINGLE series string

What typically happens in real world is the VMP goes below the minimum input of the inverter , the inverter maintains the minimum input by reducing the load as needed and more than the shaded panels output is lost

Hope this long explanation helps !

Thanks for the explanation!

I'd like to reopen this discussion, because I am designing a system that needs to use two different orientations, and I'm getting conflicting answers.

I'm talking about a setup with, say, one string with all the panels facing southeast, and another string with all the panels facing southwest, on the same inverter.

Fronius claims that with their inverters, this will result in a penalty of under 1%. (This is in a report at http://www.fronius.com/worldwide/usa.so ... itions.pdf .)

Sunny Boy's FAQ says not to do it. It makes no distinction between panels in the same string and different strings. (See http://www.sma-america.com/sysdesign.html question 3.) I called Sunny Boy, and a tech told me there would be about a 5% penalty for using differently oriented strings, with any manufacturer's inverter. I won't repeat his exact words regarding claims to the contrary, but he was vigorously dubious.

Xantrex customer support emailed me that "As long as each string is on the same side of the roof, you will get pretty good efficiency. You will get full production from one string and whatever is possible from the shaded one." This was in response to a question about this specific system orientation, so I take it he meant that all the panels in a string had to have the same orientation.

Does that mean that Xantrex and Fronius have a better algorithm? Or am I getting some incorrect guidance?

Re-read my explanation, that what happens in the real world.

I your case with parrallel strings, having 1/2 in a different orientations will have minimal losses, if any at all. It makes no difference if its a Xantrex, Fronius or SMA product, they all will track to find the vmp point and that is dominated by the tempature of the panels.

So if one sting is shaded, it isn't contributating to the harvest and has no effect on the other string as long as its of equal lenght. When the both get sun, even at different angles, its the tempature that make the largest effect of vmp, that the surounding air and local surfaces, like a roof

In real world system, panel mis-matchs and parallel mis-match stings have a much larger effect on harvest than the split angle issue your addressing. Even if your using panels that are all from the same manufacture and same rating, they are typically rated +5%-10% and this ends up with the parrallel strings having slightly different vmp boths as the voltage and current are within the +5/-10% window

The tech's that man the phones are well intended, but this level of detail on panels only a few people that actually work with and design with systems understand the details


Hope this helps