diy solar

diy solar

Grouping Solar Panels

TechRider

New Member
Joined
Feb 29, 2024
Messages
35
Location
Ohio
Hi Everyone,

Getting into the weeds on my new solar build and I've got a question. I'm going to end up with 36 solar panels on my roof. They'll be grouped into 4 strings of 9 panels in series. This complies with my requirements from the panels and sol-ark 15k. My question is, does it matter how I group them?

Assuming I'll have 3 rows of panels, do I just group them by the rows and a couple groups span multiple rows? Or - Do I do more of a grid grouping, IE the first 9 panels to the east are grouped, then the next set of 9, etc. So it would be my like a 3x3 array.

I think there will be a little shading to deal with on the lower row from a tree. The tree may get a trimming as a result.

Here's a rough sketch of what I'm looking at: The white lines are my "safety zone" of where I can place the panels in the roof. There's a few inches of roof space on each side. The brick looking thing is my chimney.
1710429850777.png

Yes - There are also only 34 panels on that drawing. I have a hipped roof so the roof is broken up a bit so I'm going to have to put 2 panels on the little side sections. One faces S-SE and the other S-SW.

I've got 340W panels on order and should be here in a couple weeks so the panels are already set.

Thanks!
 
Not sure how the sun travels, but if I / we assume your roof drawing is oriented so top is north, bottom is south, etc, there are two general scenarios

First, there are no trees or chance of shade, then it does not matter.

Second, if there is a chance of shade say on east side, then you want to setup as 3x3 on west side, 3x3 next to it which is about middle, 3x3 on east side. Shade on one panel will severely hinder your solar collection for that entire 9 panel array. That's why you cannot run a long array going east to west.

More fringe scenarios for those with OCD.

Third, it's always easier to setup long single line arrays but depending on the array orientation (east to west vs north to south), early morning and late evenings get marginally impacted as sun gradually rises and sets. Also, will need to factor in ease of install vs arrays that zig zag rather than align left to right.
 
Hi Everyone,

Getting into the weeds on my new solar build and I've got a question. I'm going to end up with 36 solar panels on my roof. They'll be grouped into 4 strings of 9 panels in series. This complies with my requirements from the panels and sol-ark 15k. My question is, does it matter how I group them?

Assuming I'll have 3 rows of panels, do I just group them by the rows and a couple groups span multiple rows? Or - Do I do more of a grid grouping, IE the first 9 panels to the east are grouped, then the next set of 9, etc. So it would be my like a 3x3 array.

I think there will be a little shading to deal with on the lower row from a tree. The tree may get a trimming as a result.

Here's a rough sketch of what I'm looking at: The white lines are my "safety zone" of where I can place the panels in the roof. There's a few inches of roof space on each side. The brick looking thing is my chimney.
View attachment 202046

Yes - There are also only 34 panels on that drawing. I have a hipped roof so the roof is broken up a bit so I'm going to have to put 2 panels on the little side sections. One faces S-SE and the other S-SW.

I've got 340W panels on order and should be here in a couple weeks so the panels are already set.

Thanks!
If you have panels in series where part of string is on main house, and part are on gables at a different angle, they will not produce significant power ever...
 
I would not bother with the extra 2 panels as they will drag the strings down they are connected to. If you do mount them then use them on their own with controller and battery to power wifi/lighting.
 
Probably easier to sacrifice one 9 panel array and use it to setup those 2 straggler panels and any other panel that does not align well. You just will need some long PV wires.

That will give you 3 good arrays and 1 not as good array.
 
Thanks! Your assumption about the top is north is correct.

There's a little chance of shade across everything. The bottom row of panels will get hit the worst. The tree that's in question will most likely get a hair cut if that happens.

My gut feeling was the 3x3 array.
 
Well, truth be told, I left out part of the build.

I've got 54 panels in total which maxes out my inverter for my panels. The last set will be an E/W string. I guess I could just add an extra panel to that side and steal one from the other array.

Here's more of a full picture:
1710436960811.png

I'm going to number them all with the panel and string so I have a reference. The triangles face S with the main section of panels and the far ends are E/W facing.

Here's the house:
1710437076919.png

The one tree is going to get trimmed a bit.
 
Not sure if this is diy or you're going to pay someone. Also, not sure if this is pure business in the sense that you are watching the ROI numbers or a hobby / partial hobby where the joy of solar can offset some of the costs.


When setting up a string, it's best to keep them as close to each other since a 2ft square piece of cardboard (or anything equivalent) can drop an entire string's solar collection from say 50% to 5% (don't quote me in actual drop though) .

Any string in series should be grouped together where performance / solar collection is similar. Stings in parallel can have difference performance.

Another way to say it, is group panels where they are located e.g. if you have a small set of panels pointing say east, then those can be connected together in series. They then can be connected in parallel as long as you keep the panel count the same for all strings (e.g. 9 is your number).

Another option is to include another MPPT that may fit a string of say 6 panels. You can use this MPPT to connect in parallel any other 6 panel string. Obviously you must stay within the specs of your MPPT.
 
You have multiple surfaces with same orientation, so OK to have panels of a string scattered.
Shadows kill output of a given panel, but rest of string will push current past (if half-cut, its internal parallel-series can confuse MPPT.)
Don't allow a shadow on one panel when others in same string get full sun; excess current through diode can cause failure. OK if off-angle illumination like 45 degrees.

If each string on its own MPPT, they will deal with shade individually.
If two strings in parallel, try to avoid shading differences greater than about 10%. On string heavily shaded in parallel with another unshaded, the shade panel won't deliver any power.

Your 4 panels E and 4 panels W each want to be on their own MPPT, or those two in parallel 4s2p into one MPPT.
Wouldn't be good 8s. And don't parallel strings of different lengths onto a single MPPT.
 
Thanks for the responses!

I'm 100% DIY on this. Doing it for the fun, cut my electric bill, and provide some grid independence if needed. That's why I was wondering about the grouping. The 2 sets of sides will be on 1 MPPT then the main roof section on the other MTTPs.

I'm also looking at putting an optimizer like a Tigo on all the panels as well. I'm not sure if that changes much.

When you're saying don't parallel strings of different lengths, you're talking series panel lengths, not cable lengths, correct?
 
I just had a thought and it's not something I've researched.

Would it be better to have 9 panels with one being partially shaded occasionally or 8 panels?
 
I think optimizers won't help. Might confuse something, whether MPPT or AFCI.
RSD may be required.

If mostly in full sun, 9s should produce more power than 8s.

Cable lengths don't matter for PV strings, even in parallel. I/V curve dominated by PV cells.
(does matter for batteries, trying to match current exactly. Also for parallel AC wires, like pass-through of grid through parallel inverters.)

PV strings in parallel, e.g. 8s2p, need to be same number of panels. Not 4s || 8s. Unless paralleling "12V" with "24V" or "36V" (which I have done.)
Approximately same Voc and Vmp per string is what matters. Doesn't have to be exact, just have W/V curve rounded peak somewhat overlapped. That is why off-angle is OK in parallel, peak voltage only shifts a few percent.
 
Back
Top