Installing serial solar panels

[DotDun]

Originally Posted by downunder

Seems a good summary.

Agree! Good summary, JEDI!

[noelex 77]

Originally Posted by s/v Jedi

On the panels in series where was has a diffused light that reduces output. Current must be equal but the panels are a current source meaning that the current is equal even if one panel only generates 1/3rd of rated power. I believe the example was 5A and both panels will deliver that so array becomes 5A with reduced power output caused by lower voltage. We saw specs where a panel that delivers 5.25A at Vmp still delivers 5A at Vsc, a shorted panel, i.e. zero Volt. That is less than 5% difference which is spectacular for practical implementation of a current source.

Jedi, perhaps I am not understanding what you are saying correctly, but have a look at a typical IV curve for the sort of panel I was using as an example. If one panel is receiving good illumination around 900-1000 Wm2 it would be generating about 5A. The second panel has a diffuse shadow and, as an example, is only receiving around 200-300 Wm2. You can see the from the IV curve that the second panel cannot produce more than 2A. No matter what voltage the second panel is operating at it cannot produce 5A at the reduced illumination.

It would be nice, for series connection, if in this sort of example the second panel could produce 5A at a lower voltage so instead of 2A @ 16v the panel could produce the same number of watts at a higher current but a lower voltage so say 5A @ 6.4V, but solar panels dont work like this as the IV curve shows. Even at 0v the panel is still producing almost the same current of less than 2A.

[Paul Elliott]

Originally Posted by noelex 77

It would be nice, for series connection, if in this sort of example the second panel could produce 5A at a lower voltage so instead of 2A @ 16v the panel could produce the same number of watts at a higher current but a lower voltage so say 5A @ 6.4V, but solar panels dont work like this as the IV curve shows. Even at 0v the panel is still producing almost the same current of less than 2A.

But when the shaded panel is driven to reverse-bias by the full-current panel, the bypass diode conducts and carries the higher current. This does not show up on the single-panel load curve, but you can see it in my simulation (a 6A panel, and a 3A panel in series). I doubt that a typical MPPT controller could find the lower-voltage, higher-power MPP, but it's there.

[noelex 77]

Thanks Paul. I agree with everything you have written but I am bit confused by your 6A/3A model that you have posted. Is this an example of a localised shadow or a diffuse shadow (where the illumination on the second panel is reduced but is uniform across the panel)?

If it is an example of a diffuse shadow how can the current at lower voltages be 12A as shown on the graph for this example when the panels are connected in parallel?

If it is a localised shadow could you please post the plot of the models prediction for 6A /3A output when the 3A is produced by a diffuse shadow as I think the performance of series and parallel needs to be considered for these cases.

[Paul Elliott]

Originally Posted by noelex 77

Thanks Paul. I agree with everything you have written but I am bit confused by your 6A/3A model that you have posted. Is this an example of a localised shadow or a diffuse shadow (where the illumination on the second panel is reduced but is uniform across the panel)?

If it is an example of a diffuse shadow how can the current at lower voltages be 12A as shown on the graph for this example when the panels are connected in parallel?

If it is a localised shadow could you please post the plot of the models prediction for 6A /3A output when the 3A is produced by a diffuse shadow as I think the performance of series and parallel needs to be considered for these cases.

What I posted (or tried to post) was the case for two 36-cell panels in series. Each panel consists of two 18-cell strings, each string having a bypass diode.

In this example, one string is shaded so it produces only 3A, and the remaining strings are unshaded, having a potential 6A current.

Again, in this example only one 18-cell string out of four total is shaded. The maximum short-circuit current here is 6A. In the graph I posted above, I don't see a 12A current, just a 6A current with a heavy load and a 3A current with a light load.

The type of shading isn't modeled in my circuit -- it could be either a diffuse shadow, or perhaps a hard shadow that covers half of a cell. I believe the panel performance will be similar in either case, but I haven't verified this in the real world. I can modify my simulation circuit to only "shade" a single cell in the string, but I would want to do testing with actual panels to see if the model truly matches the real thing. The general behavior of the model for series and parallel connections does match my rudimentary 2-panel real-world testing.

[TeddyDiver]

Diffuse shadow shouldn't be a problem on a boat, unless you tieup in the mangroves or such..

[newhaul]

Originally Posted by TeddyDiver

Diffuse shadow shouldn't be a problem on a boat, unless you tieup in the mangroves or such..

Ocasional cirrus clouds.

[Paul Elliott]

Originally Posted by TeddyDiver

Diffuse shadow shouldn't be a problem on a boat, unless you tieup in the mangroves or such..

During the 2010 Pacific Cup we had about two days of full sun during a two week race. The rest of the time it was moderate cloud cover. The shadows from the mast, main, and boom were quite diffuse. We still had less solar power than we would have seen under clear skies, but I do believe the weak shadows kept the situation from being as bad as it might have otherwise been.

[noelex 77]

Originally Posted by Paul Elliott

Again, in this example only one 18-cell string out of four total is shaded.

Thanks Paul. I was discussing a different case where 2 strings (both in the same panel) were equally shaded, so your graph did not make sense. This, of course, is just one other possibility from a multitude of possible shading options. Considering how the solar panels will behave with all the options is incredibly complicated, but to theoretically predict which is better, series, or parallel, we need to consider each of the different scenarios.

[ramblinrod]

Originally Posted by s/v Jedi

On what I saw about different voltage panels in parallel: assuming they have no blocking diodes, the higher voltage panel will start dumping energy into the lower voltage panel as well as into the controller.

Theoretically possible, but most modern panels, consisting of P/N junctions that are forward biased when producing current (ie. light is shining on them), will naturally be reversed biased when a higher voltage (open circuit) panel is placed across it in parallel, and therefore not permit current draw from the higher voltage (open circuit) panel.

Some older panels required blocking diodes.

On the panels in series where was has a diffused light that reduces output. Current must be equal but the panels are a current source meaning that the current is equal even if one panel only generates 1/3rd of rated power.

Depending on the bypass diode configuration, if any. With no bypass diodes, (such as the Gosun and Kyrocera panels tested in the Wynn's video), if the current produced by a single cell is reduced, it's impedance increases, and the current through the entire string of panels is reduced. If the cell is completely shaded, reducing the current to zero, the current through the entire series string is zero, and with no bypass diodes, the current through every connected serial string is zero.

I believe the example was 5A and both panels will deliver that so array becomes 5A with reduced power output caused by lower voltage.

Incorrect. As solar panels are also a pretty good voltage source, as shown by all the V/I curves posted, unless the panel is completely short circuited the voltage is quite high.

The real situation is that as shading occurs, and current through a cell is stopped, all current through that string is stopped. If there is no bypass diode, all current in the extended string is stopped, as it is essentially open circuit. If there are bypass diodes, the diodes act as a load, while passing current produced by other cells in the series string.

It isn't that the voltage across the cells has necessarily fallen that much, it is that the cells are simply producing less or no power, because of the lower current produced in the cells.[/QUOTE]

[Paul Elliott]

Originally Posted by noelex 77

Thanks Paul. I was discussing a different case where 2 strings (both in the same panel) were equally shaded, so your graph did not make sense. This, of course, is just one other possibility from a multitude of possible shading options. Considering how the solar panels will behave with all the options is incredibly complicated, but to theoretically predict which is better, series, or parallel, we need to consider each of the different scenarios.

Here's your scenario, with one series panel in full sun (6A) and the other partially shaded (both strings in that panel limited to 3A):

The green trace is current, the lighter trace is power.

I am planning to make some more measurements on real panels to see the amount of leakage current when voltage is placed across a shaded panel. This matters in the parallel configuration, and at night where it might put a drain on the battery. My quick tests the other day didn't show much if any leakage (even without series diodes), but I will do more tests with better metering. I think ramblinrod is correct about the newer panels not having much reverse current.

[noelex 77]

Originally Posted by Paul Elliott

Here's your scenario, with one series panel in full sun (6A) and the other partially shaded (both strings in that panel limited to 3A):

Thanks Paul. That graph is more representative of what I was expecting. Peak power in series occurs at 3A and parallel connection would produce considerably more power. I hope people do not generalise from this case that parallel is necessarily better. This is one specific, but interesting, example.

Originally Posted by Paul Elliott

I am planning to make some more measurements on real panels to see the amount of leakage current when voltage is placed across a shaded panel. This matters in the parallel configuration, and at night where it might put a drain on the battery.

It varies considerably from panel to panel, but I did some measurements a long time ago and in most cases the leakage current is very small. For most practical purposes, but not quite all (such as the unusual case where the solar panel is covered in snow for 24x7 and no controller is fitted) the leakage can be ignored.

Note: nearly all solar controllers will disconnect the panels at night. So the case of leakage from a panel in sun to heavily shaded panel connected in parallel is the only practical problem. The leakage is normally so small that this is of no practical consequence.

Fitting a blocking diode is rarely a good idea.

[foggysail]

Installing serial solar panels! I am changing my solar system and am struggling to find what I need. Sure, Alibaba has plenty of deals but then if you can get a company to supply a small quantity there is the shipping and huge tariffs.

So---- where are the places to purchase reasonably priced panels? And don't waste time with Google unless you want to get involved with home estimates and all the other crap that I personally am not interested in.

So where am I. I just tossed out my StarPro30 controller for a lot of reasons. For me, its time to go MPPT and in that quest, I just purchased and installed a Victron 150-70 controller and wonder if I messed up by not purchasing a 150-85 instead.

Panels! My current 4 polys which where advertised as 140W/panel and the best I have seen to date is about 100-110W will stay put until I get better panels. What I want are 4 each 200W mono, 36V panels. I thought of keeping my current panels but I need the real estate to increase my capacity. And in increasing my capacity, I have to be vigilant to avoid exceeding the 70 ampere limit on my new controller.

No way will I use lower voltage stuff although I would consider higher voltage panels. If I find what I am looking for, at a reasonable cost, I will operate them with two strings of series/parallel raising the operating voltage to 72V. And yes, I understand the hazards of higher voltage, I also understand the advantages of higher voltages.

[Paul Elliott]

Originally Posted by noelex 77

It varies considerably from panel to panel, but I did some measurements a long time ago and in most cases the leakage current is very small. For most practical purposes, but not quite all (such as the unusual case where the solar panel is covered in snow for 24x7 and no controller is fitted) the leakage can be ignored.

Note: nearly all solar controllers will disconnect the panels at night. So the case of leakage from a panel in sun to heavily shaded panel connected in parallel is the only practical problem. The leakage is normally so small that this is of no practical consequence.

Fitting a blocking diode is rarely a good idea.

I did some leakage measurements on my Renogy 100W panel, using a variable power supply to apply a voltage to a completely dark panel. I measured leakage current vs voltage across the panel. This is similar to the case where a completely shaded panel is in parallel with an unshaded one, or at night where the panel is connected directly to the battery without a blocking diode or a leakage-blocking controller.

What I found was that the cells behave in a similar manner to the spice simulation model: a cell is a diode. With my 36-cell panel, here's the leakage current:


Note that if this dark panel were connected directly to a battery at 12.5V (for example), the leakage current would be about 15mA. Not enough for me to worry about, but some folks might be concerned.

If this dark panel was in parallel with an operational panel at the Vmp of 18.9V, the leakage current would be about 105mA. This is about 2% of the active panel current. We do start worrying about 2% voltage drop in our wiring, so perhaps a blocking diode is worth thinking about.

But, that Schottky blocking diode wastes about 0.2V, or 1% of the Vmp, and this drop is *always* there, so we would probably lose more power, on average, with the blocking diodes. The leakage current only happens when the panel is heavily shaded.

We have occasionally referred to this leakage current as a reverse bias current. This actually is not correct -- the solar cell diodes are forward biased in this case. Cells may become reverse biased in a series-panel configuration, and that's when the bypass diodes kick in.

[OceanSeaSpray]

Originally Posted by Paul Elliott

I did some leakage measurements on my Renogy 100W panel, using a variable power supply to apply a voltage to a completely dark panel. I measured leakage current vs voltage across the panel. This is similar to the case where a completely shaded panel is in parallel with an unshaded one, or at night where the panel is connected directly to the battery without a blocking diode or a leakage-blocking controller.

What I found was that the cells behave in a similar manner to the spice simulation model: a cell is a diode. With my 36-cell panel, here's the leakage current:

Note that if this dark panel were connected directly to a battery at 12.5V (for example), the leakage current would be about 15mA. Not enough for me to worry about, but some folks might be concerned.

If this dark panel was in parallel with an operational panel at the Vmp of 18.9V, the leakage current would be about 105mA. This is about 2% of the active panel current. We do start worrying about 2% voltage drop in our wiring, so perhaps a blocking diode is worth thinking about.

But, that Schottky blocking diode wastes about 0.2V, or 1% of the Vmp, and this drop is *always* there, so we would probably lose more power, on average, with the blocking diodes. The leakage current only happens when the panel is heavily shaded.

We have occasionally referred to this leakage current as a reverse bias current. This actually is not correct -- the solar cell diodes are forward biased in this case. Cells may become reverse biased in a series-panel configuration, and that's when the bypass diodes kick in.

Indeed. I find your reverse currents surprisingly high overall. I have measured 0.14mA flowing back into a 36-cell 95W monocrystalline panel from the battery at night.

Schottkys are not THAT good, unless the current is tiny. They commonly drop a good 0.4V under load. It is very easy to replace them with an ideal diode made from a MOSFET and a small IC like a LTC4412 or LTC4357 and this is really the way to do it. There are also fully packaged ideal diodes like this one if the voltage is ok.