Mounting inverter in Phoenix: Garage or Outside?

[inetdog]

...because the way you numbered your panels is mostly to group those with the same orientations together....

Unless you are using DC optimizers, there is absolutely no sense to putting panels with different orientations into any one series string.
You will not be able to use any of the output of the lower-producing panels at the same time as the full output of the better insolated panels.
Putting different orientations (complete strings) in parallel into one MPPT input, on the other hand, can often work out just fine.

[HX_Guy]

The advantage of going with 2 smaller inverters is to have 4 strings of 12 panels on 4 separate MPPT (assuming each inverter can give you 2 dual MPPTs each), where you can string the panels with the same orientation up on the same string for optimum performance. That way, your strings are optimized for their different orientations.

I notice that you do have south, east and west facing panels on your roof, so there may be some advantage to have 4 strings to optimize to your panels' different orientations. I copied the pic you use of your layout to show my point. Here you actually have 5 sets of 10 panels, and we're talking about just 48 panels above, so it's not apples to oranges, but you get the idea, because the way you numbered your panels is mostly to group those with the same orientations together.

[ATTACH=CONFIG]4282[/ATTACH]

I know I have a million threads going but the final layout has been changed to this:



I actually have two orientations, South (3 strings of 12 - yellow, red, green) and West (1 string of 12 - blue). I've read that there shouldn't be a problem with different orientations into one MPPT as long as it's a complete string, which is what inetdog just said as well.

My biggest unknown is the shade factor and how if one string is shaded, how does it affect the other strings? According to this site, it doesn't...but is that accurate?

http://www.solarchoice.net.au/blog/partial-shading-is-bad-for-solar-panels-power-systems/

"If, for example, shade from a tree or a chimney is cast on even one of the panels in the string, the output of the entire string will be reduced to virtually zero for as long as the shadow sits there. If there is a separate, unshaded string, however, this string will continue to produce power as per usual."

[inetdog]

OK, I can go into more detail on any part of this if you like, but here is the bottom line executive summary:

1. Different orientations but with the same number and type of panels and no partial shading which affects one string and not another will work well with one MPPT input and may even let you use a smaller inverter while spreading the high power production hours out over a longer period of time each day.
2. Partial shade which reduces the Vmp of a string by more than 5% compared to unshaded strings, or anything which causes a difference of more than 5% in string Vmp can cause a significant loss in power from that string. If the effective Voc of that string drops below the Vmp of the other strings you may get zero contribution. I am introducing the term effective Voc because even a shaded panel will produce full voltage as long as no current is drawn. So instead of an open circuit measure, you would measure effective Vocas the voltage you see when you draw less than 10% of the current Vmp of the rest of the panels in the string.

[HX_Guy]

You rock inetdog! Any way I can get a phone or chat consultation?

Ok so help me understand this...talk to me like a child because I know nothing about voltages and electricity and such.

1. So in the months where there is zero shading issues throughout the whole layout, the different orientations are no problems. I got that part.

2. Would you mind explaining this to me using actual numbers from my setup?



This is what I understood, correct me where I'm wrong...

If the vmp of 1 strings drops below 258.875 (based on reg. vmp of 272.5), then that whole string can have significant loss in power, correct? Now does it matter where that loss comes from...one whole shaded panel vs say 1/3 of multiple panels? And you also mentioned Voc...what is Voc vs vmp?

Lastly, how will the other 3 completely unshaded strings be affected if string #4 is shaded? Let's say full shaded and completely out of operation. Will the other 3 strings still work as normal? What if 2 of the 4 strings are significantly shaded and off line (say over 50% shaded)?

[inetdog]

You rock inetdog! Any way I can get a phone or chat consultation?

Ok so help me understand this...talk to me like a child because I know nothing about voltages and electricity and such.

1. So in the months where there is zero shading issues throughout the whole layout, the different orientations are no problems. I got that part.

2. Would you mind explaining this to me using actual numbers from my setup?



This is what I understood, correct me where I'm wrong...

If the vmp of 1 strings drops below 258.875 (based on reg. vmp of 272.5), then that whole string can have significant loss in power, correct? Now does it matter where that loss comes from...one whole shaded panel vs say 1/3 of multiple panels? And you also mentioned Voc...what is Voc vs vmp?

Lastly, how will the other 3 completely unshaded strings be affected if string #4 is shaded? Let's say full shaded and completely out of operation. Will the other 3 strings still work as normal? What if 2 of the 4 strings are significantly shaded and off line (say over 50% shaded)?

Consultation: phone probably not, but chat maybe. You only get what you pay for and I would not charge you anything.

1. Yup, correctly stated.
2. The current versus voltage curve of a solar cell (or panel or string) has a cutoff voltage at the high end which is temperature dependent but otherwise is almost completely independent of the amount of light hitting the cell, from candlelight to full sun. This is call the Open Circuit Voltage or V_OC for short.
You measure that by putting a high input impedance voltmeter across the output with no other load connected.
The voltage drops slowly as you start to draw current, then the voltage drops off very quickly to zero as you get to the Short Circuit Current, I_SC. The value of I_SC is roughly proportional to the intensity of the light hitting the cell/panel/string.
Since both zero voltage and zero current will produce zero power, the maximum output of the panel will come at a point somewhere in between, called the Maximum Power Point or MPP. The corresponding voltage and current values are V_MP and I_MP respectively.
To get a string to deliver its maximum power, the CC or GTI must adjust its input impedance so that it is drawing exactly I_MP amps at V_MP volts. That process is called Maximum Power Point Tracking, or MPPT.

<Pause for breath.>

If the V_MP value of a string drops by no more than 5% relative to the other strings, the result of the MPPT seeking and finding a voltage somewhere in between the two will be close enough for government work to getting full power from both strings. The output power changes fairly slowly with voltage on either side of the MPP. (Just because it is a maximum.) It does not matter whether that drop in voltage comes from a single panel or a small part of lots of panels. But the smallest part of one panel that you can add up this way, because of the way the bypass diodes divide the panel into sections, will be 1/2 or maybe 1/3 of the panel. One totally shaded cell will cost you 1/2 or 1/3 of the output voltage at V_MP. Two totally shaded cells in different parts of the panel could cost you the entire output.
If the MPPT control unit sees the output power starting to decrease as it pulls more current, then it might not keep on pulling even more current to get to the voltage point where the shaded string will start to contribute at all. That is what happens when the effective V_OC of the shaded string is lower than the V_MP of the other string(s).

If one of the strings is so shaded that it will not contribute anything, it still will not interfere with the operation of the other strings, since its V_OC will not change noticeably. So it just will not deliver any current, but will not suck any either.

[HX_Guy]

Wow that's a ton of mind boggling info, thank you though. Can what you said there be put into practical example with the setup I'm running?
Can the VOC of each string be calculated, what would it be? Or not really because as you said, it totally depends on the temperature outside?
I mean I see in the software I'm using it says "Max System VOC 494.9", so is that it? Then it says "Max Short Circuit Current 43.2" You said the maximum output will come at a mid point, so would it be 269.05 (494.9 + 43.2 / 2)? Or it's not quite that cut and dry because

It seems the array Vmp changes with temperature...but at the inverter stage, does the inverter try to keep it at a certain point? (Not sure if that question actually makes sense). Basically, would I be able to figure out at different temperatures when the string would drop out of production (drop of 5% or more Vmp)?

If a panel's voltage is 23.575 (based on 282.9v @ 107º divided by 12 panels = 23.575v) and it has 60 cells, each cell equals 0.393v? But with 60 cells, they are divided into groups of 20 with each group having a bypass diode...so if one or two cells in a group of 20 is shaded, you lose 7.86v from that panel? And if you can only lose up to 5% before the production drops significantly or totally, then 2 groups of 20 shaded cells (either on one panel or on 2 panels) would basically knock out the whole string? Am I understanding that right?

So in this example from SketchUp, at noon on Nov 1st, since a full panel of the yellow string plus the bottom 1/3 of about 4 panels on the green string would be shaded, both of those strings would produce most likely zero, correct? But the red (and blue, not shown) would continue to produce 100%?

[inetdog]

Wow that's a ton of mind boggling info, thank you though. Can what you said there be put into practical example with the setup I'm running?
Can the VOC of each string be calculated, what would it be? Or not really because as you said, it totally depends on the temperature outside?
It can be approximated by looking at the V_OC spec on the panel label, and then corrected for temperature (if necessary, I would not worry because V_OC and V_MP change together in a close to constant ratio.
I mean I see in the software I'm using it says "Max System VOC 494.9", so is that it? Then it says "Max Short Circuit Current 43.2" You said the maximum output will come at a mid point, so would it be 269.05 (494.9 + 43.2 / 2)? Or it's not quite that cut and dry because
You are taking the average of a voltage and a current. Stop that.
As a very broad rule of thumb, V_MP will be 80% of V_OC, regardless of the amount of current the panel is producing at the time.

It seems the array Vmp changes with temperature...but at the inverter stage, does the inverter try to keep it at a certain point? (Not sure if that question actually makes sense). Basically, would I be able to figure out at different temperatures when the string would drop out of production (drop of 5% or more Vmp)?
Do not try to bring temperature into it. It does not really matter for this calculation. (shading). It can effect the maximum number of panels you can safely put in a string if you get very cold winter temperatures.


If a panel's voltage is 23.575 (based on 282.9v @ 107º divided by 12 panels = 23.575v) and it has 60 cells, each cell equals 0.393v? But with 60 cells, they are divided into groups of 20 with each group having a bypass diode...so if one or two cells in a group of 20 is shaded, you lose 7.86v from that panel? And if you can only lose up to 5% before the production drops significantly or totally, then 2 groups of 20 shaded cells (either on one panel or on 2 panels) would basically knock out the whole string? Am I understanding that right?
Pretty close, yes. We generally approximate the voltage per cell as .5V, but that is V_OC, so .393V for V_{MP is quite credible}

I put my answers in bold above....

[HX_Guy]

One other thing, you said that as long as the Voc between strings stays pretty much the same, then shading will not affect the other strings, correct?

But what if the area is so shaded that it affects the Voc of those panels/string? Since Voc is related to temperature, I assume the Voc of panels getting direct sunlight will be hotter than ones in the shade, no?

I'm talking about an example like here:

[inetdog]

One other thing, you said that as long as the Voc between strings stays pretty much the same, then shading will not affect the other strings, correct?

But what if the area is so shaded that it affects the Voc of those panels/string? Since Voc is related to temperature, I assume the Voc of panels getting direct sunlight will be hotter than ones in the shade, no?

I'm talking about an example like here:

Shading does not affect the V_OCof a cell unless the resulting light is much less than moonlight or maybe even starlight intensity. It does not happen during the day with any kind of normal shade. You are not planning to put a blackout cover over one of the panels are you?

PS: your image link comes up 404....

[HX_Guy]

I waas planning to but I guess I won't.

Ok, so then the Voc is not something I should worry about when it comes to shading.

Back to the Vmp and the 5% different between one string and the others...basically as I already said earlier, 40 cells = 5% of the string's Vmp, which is very easy to reach and will practically always happen as soon as the area is shaded. Is it a pretty safe bet to figure that any shade on any string in my setup = no production out of that string? Which means that...

Yellow String

October: No production before 2:00pm
November: No production before 2:45pm
December: No production before 3:00pm
January: No production before 3:00pm
February: No production before 2:00pm

Green String

October: No production
November: No production
December: No production
January: No production
February: No production after 4:15pm

Kind of a bummer, huh?

Here are the links to the images, not sure why it's not showing up, it works on my screen:

http://www.nitnelav.com/shadedstrings2.jpg

http://www.nitnelav.com/shadedstrings.jpg

Did the other images earlier show up?

[inetdog]

http://www.nitnelav.com/shadedstrings2.jpg

http://www.nitnelav.com/shadedstrings.jpg

Did the other images earlier show up?

Not Found

The requested URL /shadedstrings2.jpg was not found on this server.

Additionally, a 404 Not Found error was encountered while trying to use an ErrorDocument to handle the request.

Does that site require a login?

[HX_Guy]

Very weird, I'll have to look into that. Anyway, I uploaded the images elsewhere, should work now.

http://s27.postimg.org/rrqgrjp8z/shadedstrings.jpg

http://s9.postimg.org/434vg4cvz/shadedstrings2.jpg

[inetdog]

Very weird, I'll have to look into that. Anyway, I uploaded the images elsewhere, should work now.

Much better now....

Yes, the situation where you have one full panel shaded or four or five 1/3 panels shaded will likely give you zero output from both strings.
But if you connect the two of them to their own shared MPPT input they both should be able to produce.

[HX_Guy]

Eh? Run that by me again?

What do you mean if I connect both of them to a shared MPPT they would still produce? Is it because they would have a similar Vmp? I guess that makes sense but only when both are shaded equally. In a situation such as this, it wouldn't help I don't think, right?



I guess this is where having two inverters that each have dual MPPTs would come in useful...each string could be on it's own MPPT. But even then, once the Vmp drops below the inverters minimum Vmp, that string would still drop out. Though there would be a lot more leeway...with something like the Power One 6000US, it has a minimum Vmp of 180 Vmp and 12 panels in the winter should be about 300 Vmp meaning I could have 240 total cells shaded (4 full panels or 14 1/3 panels) and it would still produce. Pretty good.

[HX_Guy]

Giving it a little more thought, I still don't quite get how my installer came up wtih +$4000 for a 2 inverter setup. We didn't talk specifics on which inverters, maybe he assumed two of the Fronius, but why would he?

Looking online, the Fronius 11.4 can be had for $3219 while the Power One PVI-6000US is $2,277, or $4554 for two. He said a combiner box is about $400 of the price plus additional labor and such...but still, it shouldn't be more than an additional $2,000...say $2500, but $4000?