wiring different voltage panels in parallel - but why?

[SolarQueen]

I made a series of demo videos showing what happens when you wire mismatched solar panels in various configurations. I'm now trying to explain the "why" behind what we saw. I thought some of you folks here would be able to give me the mathematical or scientific reasoning. I'm cool with the bottle neck that was created with trying to wire two panels with different current in series. What I'm trying to explain is why when two different voltage solar panels are wired in parallel, the voltage from the higher voltage panel was pulled down to the lower voltage panel. I read one place that said the lower voltage panels' cells became reversed biased, and so basically the solar panel became a resistor. But is that the full story? When two batteries of different voltages are wired in parallel, the higher voltage charges the lower voltage one, equalizing them. Would it be the same reasoning, with current flowing from high to low voltage? I'm sure Ohm's Law is involved, but haven't nailed down what formula explains it. Anyone want to take a shot at explaining it in layman's terms? Thanks.

 

[GoldDigger]

I made a series of demo videos showing what happens when you wire mismatched solar panels in various configurations. I'm now trying to explain the "why" behind what we saw. I thought some of you folks here would be able to give me the mathematical or scientific reasoning. I'm cool with the bottle neck that was created with trying to wire two panels with different current in series. What I'm trying to explain is why when two different voltage solar panels are wired in parallel, the voltage from the higher voltage panel was pulled down to the lower voltage panel. I read one place that said the lower voltage panels' cells became reversed biased, and so basically the solar panel became a resistor. But is that the full story? When two batteries of different voltages are wired in parallel, the higher voltage charges the lower voltage one, equalizing them. Would it be the same reasoning, with current flowing from high to low voltage? I'm sure Ohm's Law is involved, but haven't nailed down what formula explains it. Anyone want to take a shot at explaining it in layman's terms? Thanks.

Here goes:

The solar cell (and the combination of cells in series to make a panel) can be modeled as a diode plus a voltage source (powered by sunlight) which has a maximum voltage almost independent of light level and a maximum current which is proportional to the light level.
With no load the panel produces the open circuit voltage, Voc, but no current and so no power.
With a short circuit the panel produces the short circuit current, Isc, but no voltage and so no power.
In between those two extremes is a point where the panel produces its maximum power, Vmp times Imp.

If you put two panels with different voltages in parallel and try to maximize the combined power output you will drag the voltage of the higher voltage panel down below Vmp and get less than its maximum contribution. if the two panels are close enough in voltage the operating point will be somewhere between the low Vmp and the high Vmp and the total power will be more than either panel alone but less than the sum of the two at their individual optimum.

When you put two panels with different current specs in series it is more complicated. Attempting to drive more than Isc through a panel will cause a resistive power loss in the cells and heat the panel. Too much forced current will damage the panel by heat or by exceeding the breakdown voltage of the diode. Either is bad, so all commercial panels also contain bypass diodes which are connected around subgroups of the series string. When you try to force current through the string the bypass diodes become forward biased and conduct, protecting the cells.
The power output from that panel then becomes zero (actually a small loss because of the voltage drop in the bypass diodes.)
So the device connected to the string (MPPT input) will end up drawing either the Imp of the higher current panel with no contribution from the other if the mismatch is large enough OR drawing a current between Imp of one panel and Imp of the other and seeing a lower voltage from the panel which is being overcurrented. Again the the total power MAY be more than either panel alone but less than the sum of the two at their individual optimum. But it also could be a little less than the output power of the higher current panel. Worse than if you used it alone.

I will not try to show the calculations, but the rule of thumb is that if you put two panels with less than 5% difference in voltage in parallel the output will be within a percent or so of the sum of the two rated outputs.
Same thing if you put two panels with current ratings within 5% of each other.
Outside the 5% range you need to look more closely at your requirements to decide whether to try to add mismatched panels or not.

A solar panel is not a voltage source like a battery is. It is a current source (with a non-flat current versus voltage curve). Any analogy to parallel batteries will just get you into trouble and prevent a proper understanding.
Ohms law is not a great help either since the solar cell is a non-linear device, where R is not constant but depends on the operating point.

 

[SolarQueen]

I knew I came to the right place. Thanks for your terrific answer. I did several different versions of this demo, with varying differences between the panels. The one I'm trying to explain is a 12V 100W and a 24V 200W, both with 5.56Imp. Wired in series with an MPPT was pretty much the sum of the individual outputs, but wired in parallel the voltage output was just about the same as the 12V alone (34.7V and 18.1V in parallel equaled 20.9V.) It was just as expected, the lower voltage dragged the higher voltage down, but the why was tricky. I think I will just stick with the "what" happens and not complicate it with the "why". The geek in me is trying to over complicate it.

I love your explanation of the series, I'm changing my explanation to use yours. Thanks again.

 

[jaggedben]

Well, I'm just a designer, not a physicist, but...

You can start with the I-V curves for each panel. When they are parallel the voltage has to be the same at the point of parallel connection. So when they are connected to a load that finds the maximum power point (such as an inverter or charge controller), the load will find the voltage where the two I-V curves result in the most power, which may well be closer to the lower voltage. Laying the two (properly scaled) graphs of the I-V curves over each other and seeing where the lines cross would probably be the most visually informative explanation of why you end up with a certain voltage.

...What I'm trying to explain is why when two different voltage solar panels are wired in parallel, the voltage from the higher voltage panel was pulled down to the lower voltage panel. I read one place that said the lower voltage panels' cells became reversed biased, and so basically the solar panel became a resistor. ...

To clarify, this happened when the solar panels were under load, or with an open circuit? I would expect voltages to be in-between with an open circuit. Under load is a different question though. Note that you're looking at two completely different spots of the I-V curves depending on the answer to that question.

I believe it's less accurate to say the solar panel 'became a resistor' than to say that its internal resistance is shifted higher by the other panel biasing the voltage. The internal resistance of the panel varies with the load and irradiance and that's why you have an I-V curve in the first place.

 

[jaggedben]

oops, posts crossed. :roll:

 

[GoldDigger]

I knew I came to the right place. Thanks for your terrific answer. I did several different versions of this demo, with varying differences between the panels. The one I'm trying to explain is a 12V 100W and a 24V 200W, both with 5.56Imp. Wired in series with an MPPT was pretty much the sum of the individual outputs, but wired in parallel the voltage output was just about the same as the 12V alone (34.7V and 18.1V in parallel equaled 20.9V.) It was just as expected, the lower voltage dragged the higher voltage down, but the why was tricky. I think I will just stick with the "what" happens and not complicate it with the "why". The geek in me is trying to over complicate it.

I love your explanation of the series, I'm changing my explanation to use yours. Thanks again.

Thanks for the comments. I have seen some of your videos and appreciate your contribution to consumer (and professional) education.
One thing that was implicit in your question and explicit in your videos is that the load on the panel combo is an MPPT input device. Without knowing that the whole exercise becomes meaningless. (For example two different voltage panels each with Vmp above the (battery voltage plus controller drop) threshold will combine into a PWM CC to produce almost exactly the sum of their individual currents into the same PWM CC.)

Even knowing that the load is MPPT leaves room for differences in results though. The parallel case, for example, may produce two downward opening curved sections of the power versus voltage curve. (Two local maxima, for the math geeks) And some MPPT CCs will find only one of them (typically the one at the highest voltage) and will not seek out the other, which may actually produce a higher total power. This is particularly likely when the Vmp of one panel is higher than the Voc of the lower voltage panel. If the MPPT algorithm stops as soon as the power starts to decrease (as it increases the current drain) then it will not find the lower voltage peak.
Similar double peaks can occur with a series string, and this behavior actually has real world effects on getting maximum power from a series string which is partially shaded.

 

[GoldDigger]

Laying the two (properly scaled) graphs of the I-V curves over each other and seeing where the lines cross would probably be the most visually informative explanation of why you end up with a certain voltage.

I find it difficult to estimate the MPP directly from looking at an I-V curve, although I can usually get within ten percent or so.

If you want to be able to find the joint MPP graphically you really need to plot the P-V or P-I curves instead. Even then the maximum will usually not occur where the two curves cross. You need to add them and look directly for the peak(s).

The point where one I-V curve is decreasing at the same rate that the other curve is increasing could be a good starting point for finding the joint MPP, rather than the point where they cross.

 

[jaggedben]

Good points, thanks.

 

[GoldDigger]

I find it difficult to estimate the MPP directly from looking at an I-V curve, although I can usually get within ten percent or so.

If you want to be able to find the joint MPP graphically you really need to plot the P-V or P-I curves instead. Even then the maximum will usually not occur where the two curves cross. You need to add them and look directly for the peak(s).

The point where one I-V curve is decreasing at the same rate that the other curve is increasing could be a good starting point for finding the joint MPP, rather than the point where they cross.

Typo there and too late to edit. The bold should have read P-V curve.

 

[SolarQueen]

You guys are terrific, thanks so much. And thanks for the compliment on the videos. We are trying to help spread solar education around the world (almost half of our audience is international), so I'm trying to be as accurate as possible, without getting too technical and scaring people off. You've given me a lot of material to incorporate. I'm thinking a follow-up video with PWM may be in order. It's a beautiful day out there today, I'm going to demonstrate partial shading effects, and anything else that can keep me outside as long as possible. Now where did I put my sunscreen? (I love my job)

 

[stephenpjohnson]

Thanks for your videos

Thanks for your videos

I have enjoyed watching your videos. I am an experienced system engineer with mostly a computer background and I'm trying to come up to speed on the practical aspects for a solar system design. Thanks much.