Solar Panels - Series or Parallel ?

[prroots]

Originally Posted by twistedtree

For every complex problem there is a simple answer - and it's usually wrong.

I think the reality here is that it's complex, and the best answer depends on a number of factors. Things like:

- Does the charge controller accept high input voltages, and if so, what's the max allowed?

- How many panels can be wired in series without exceeding the max input voltage. This needs to be a worst case calculation using high light and low temp. There's a limit to how many panels can be in series, and if you have more then some degree of parallelism is required.

- If you are forced to parallel, the strings need to be the same. For example, you might have 4 panels, but your charger can only handle 3 in series without exceeding it's max input voltage. This will force you to wire 2 strings of 2 panels each.

- Is shading more probable in one area over another? If you have a series/parallel combination, you can pick which panels are in series and which are in parallel. This can let you optimize for the most likely shading scenarios.

No matter how you slice it, some degree of engineering is needed to get an optimal configuration, or to at least confirm that you intuition (or buddy down the dock) is right.

Thanks. I notice that 3 out of your 4 points reference the maximum voltage of controller. To be fair when comparing parallel to series, I believe we should assume the controller is capable of handling all possibilities within reason. Our conclusions should not be constrained by limitations of the controller. Once we understand the fundamental behavior of series vs parallel we may decide to install a lesser controller (eg, to save money), but at least that decision would be based upon a knowledge of the pros and cons of each configuration.
Pete

[prroots]

Originally Posted by 44'cruisingcat

I have 6 panels connected to an Outback power systems MPPT charge controller.

Used to have all 6 in series, but read in the Outback manual that for a 12 volt system, having 24 volts from the solar panels was the most efficient.

So I've gone to a series/parallel setup, but actually THINK(1) I'm getting less power from the system on cloudy days. Probably try all in series again, and see how it looks.

1: (It's difficult to know for sure - on good days the system goes into absorb/float modes, so you don't get the full input, and comparing cloudy days is guesswork - how cloudy is cloudy?)

Based on your testing it would seem all in series would be a little better.

Do what I do; turn on every DC load you can think of to keep it from going into absorption mode. I agree, it's near impossible to make valid comparisons when overcast. I think the Outback manual was referring to one aspect of the issue ie, the efficiency of controller itself, however, there are other issues which must be weighed in the balance and which could change your overall results. How many bypass diodes does each panel have? I believe that could significantly affect the results under shading.
Pete

[s/v Jedi]

Originally Posted by prroots

I'll make a first attempt at repeating your theoretical analysis under the the case of two 18 Volt panels with 4 bypass diodes/panel (ie, my case).

Overall, it still seems to favor the series configuration. If one assumes that each of the above cases were equally probable, the average for series is 4.7 watts and for parallel is 3.8 watts. Comments welcome.
Pete

The only remarkable thing about it is that testing was actually needed to get to this point (theory is solid, no magic involved), but now that Pete got that done, we can point all the naysayers to this post from now on. From 3.8 to 4.7W means that series connection is 25% more efficient.

ciao!
Nick.

[hellosailor]

noelex-
Yes, it should be easy by now, but have you ever tried getting a straight answer from a mattress salesman?

Every panel maker apparently uses some arbitrary logic to install bypass diodes, so every brand or model of panel may behave differently in "shade" of different types.

Then, there are also 3 different logic schemes used by MPPT controllers. Each maker apparently considers their own choice of logic to be proprietary information and prefers not to discuss it, making that hard to compare. But each of the three schemes performs differently under changing light conditions, i.e. passing clouds.

And it has been suggested to me that since MPPT controllers aren't putting out DC, but are using sometimes complex pulsed DC, that even an RMS-reading meter isn't necessarily going to be reading "true" power, the power may have to be considered as an AC signal since it may be a complex DC one that fools some meters.

I would suggest that the correct answer (series or parallel) is going to vary with the panels (bypass diodes where?) and the controller brand and model. Finding the pony in this pile may require putting hands into it.<G>

[ChuckSK]

Here are some graphs showing power vs voltage curves for various connections of 4 Kyocera 130 panels. Also included are curves for 1 shaded cell for each configuration. These were developed using AIM-Spice and a model I designed to match the Kyocera 130 specifications.



These should help inform a decision on series vs. parallel.

Chuck

[prroots]

Originally Posted by ChuckSK

Here are some graphs showing power vs voltage curves for various connections of 4 Kyocera 130 panels. Also included are curves for 1 shaded cell for each configuration. These were developed using AIM-Spice and a model I designed to match the Kyocera 130 specifications.
These should help inform a decision on series vs. parallel.

Chuck

Thanks; very interesting. Again, series has the highest performance at least with one cell shaded. Also, shading in both parallel and parallel/series configurations causes a bump in the IV curve which could fool some MPPT controllers. Our Morningstar TriStar MPPT 45 controller does a complete scan from 0 to Voc every few seconds and even reports Voc along with Pmp and Vmp after each scan. By comparing Vmp and Imp with the characteristic IV curves for my panels I can determine if it's operating at the characteristic maximum power point.

It appears that you lost approximately 65 watts (ie, 50% of one panel's output) with one cell shaded in series mode. I'm a bit surprised is was that high unless you modeled it with only 2 bypass diodes/panel which would suggest a shaded cell causes loss of about 50% plus the diode.

I see that you lost approximately 130 watts (ie, 100% of one panel's output) with one cell shaded in parallel mode. This is expected since the bypass diodes (as I understand it) do not help when panels are in parallel mode with the effect that the entire panel dropped out.

Wonder if you can model the case with shading of both one and two cells in different strings of the same panel with, let's say, 4 bypass diodes/panel?
Pete

[GeoPowers]

Originally Posted by s/v Jedi

The only remarkable thing about it is that testing was actually needed to get to this point (theory is solid, no magic involved), but now that Pete got that done, we can point all the naysayers to this post from now on. From 3.8 to 4.7W means that series connection is 25% more efficient.

ciao!
Nick.

Unfortunately I don't think it is as simple as was demonstrated here. For example, I read this interesting article by the Canadian Center for Science and Education that pointed to the superior performance of parallel vs. series or series-parallel systems.

See here: Effect of Shading on Series and Parallel Connected Solar PV Modules | Ramabadran | Modern Applied Science

I for one would like to go series for the cost savings in wiring and fuses even if there was no other benefit, but I am still on the fence due to the shading issue. If I run wiring for series, I would have to replace it all with heavier wire if I changed my mind and went to a parallel configuration.

Frank

[ChuckSK]

Frank....

That study did not include bypass diodes in the model. All modern panels use a bypass diode for about every 18 cells or 9 volts. Thus limiting the loss during partial shading.

prroots....

The Kyocera 130 watt panel has 36 cells and 2 bypass diodes. Adding more bypass diodes to these panels would be unrealistic. I will run additional studies with more cells shaded.

Chuck

[prroots]

Originally Posted by ChuckSK

The Kyocera 130 watt panel has 36 cells and 2 bypass diodes. Adding more bypass diodes to these panels would be unrealistic. I will run additional studies with more cells shaded.

Chuck

Thanks. Personally, I don't think we'll learn much of interest by adding additional shading with only 2 bypass diodes/panel ie, shading of two cells in separate strings will shut the panel down. If I'm correct, we don't need a computer model to tell us that. As suggested, it would be of interest to see the effect of increased shading with panels having 4 or more bypass diodes/panel. My panels have 54 cells with 4 bypass/diodes which gives either 13 or 14 cells per diode.
Pete

[ChuckSK]

Hi Pete

The panels you describe seem unusual. What type are they?

[prroots]

Originally Posted by ChuckSK

Hi Pete

The panels you describe seem unusual. What type are they?

They are Photowatt model PW2050-210 panels rated at 210 watts each:

http://www.photowatt.com/documents/p.../PW2050_EN.pdf

I'm not sure why you say they are unusual. I read somewhere that some of the Kyocera's have 8 bypass diodes/panel

BTW, if I had your panels and decided to configure them in series, I would be very tempted to put bypass diodes around each panel. That way any one or two panels could shut down and, If I'm correct, you would still be generating power. I have read that this is sometimes worthwhile.
Pete

[s/v Jedi]

Originally Posted by GeoPowers

Unfortunately I don't think it is as simple as was demonstrated here. For example, I read this interesting article by the Canadian Center for Science and Education that pointed to the superior performance of parallel vs. series or series-parallel systems.

Frank

Originally Posted by ChuckSK

Frank....

That study did not include bypass diodes in the model. All modern panels use a bypass diode for about every 18 cells or 9 volts. Thus limiting the loss during partial shading.

Chuck

Exactly this is the reason that the myth of parallel vs series connection stays alive... there is tests and research out there that must be from decades ago because that is how long all the panels have bypass diodes.

Just to make sure: yes, it is as simple as demonstrated and as simple as calculated before demonstrating it: series outperforms parallel by such a huge amount that it is absolutely weird to even contemplate parallel connection. A 25% difference... !!

ciao!
Nick.

[ChuckSK]

Hi Pete

I checked the other products produced by Photowatt and all other panels had 1 bypass diode for each 18 cells. I suggest that your panel of 54 cells has 3 bypass diodes and that the documentation has a misprint.

If that is the case then 2 Photowatt 210 watt panels in series would be equal to 3 Kyocera 130 watt panels in series. Both configurations would have 6 sub-panels.

Chuck

[Cotemar]

Originally Posted by s/v Jedi

Just to make sure: yes, it is as simple as demonstrated and as simple as calculated before demonstrating it: series outperforms parallel by such a huge amount that it is absolutely weird to even contemplate parallel connection. A 25% difference... !!
ciao!
Nick.

Its random statements like this that continue to keep the parallel vs series connection a myth.

As most solar setups are using a MPPT controllers and thats what handles both parallel & series wiring scenarios with ease.

[prroots]

Originally Posted by ChuckSK

Hi Pete

I checked the other products produced by Photowatt and all other panels had 1 bypass diode for each 18 cells. I suggest that your panel of 54 cells has 3 bypass diodes and that the documentation has a misprint.

If that is the case then 2 Photowatt 210 watt panels in series would be equal to 3 Kyocera 130 watt panels in series. Both configurations would have 6 sub-panels.

Chuck

Interesting! I'll drop them an email. I've been thinking of asking for a schematic showing where the diodes are located with respect to rows and columns
Pete