Installing serial solar panels

[ramblinrod]

Originally Posted by hellosailor

Using multiple MPPT controllers in parallel to the same battery bank, with a heavily discharged battery (low voltage, high absorption rate) I would expect the real-world result to be fine. But as the battery voltage came up and the absorption rate went down, now the MPPT controllers will be seeing their mutual output voltage as "the" battery voltage, perhaps confusing the charge cycle state or costing efficiency, no?

Real world testing on that?

And if the controller is sampling battery voltage, by stopping output for a split second and reading battery voltage, then isn't there more potential for a problem as each controller goes to zero output, and reads another controller's output, more so than the real battery voltage?

There are some fairly respectable and highly regarded places, like Sandia National Laboratories, that have plenty of docs on solar installations. I'm betting there are papers up there discussing serial-vs-parallel and bypass diodes, and any "discrepancy" from them is very possibly from a panel manufacturer lying about what and how they've really connected any diodes. Wouldn't be the first time a vendor omitted parts--and then claimed they were still there. Like the no-name MPPT controllers that aren't MPPT at all.

From a fast peek at the Wynn's video (

) those Go-Power 160W panels are single series cell arrays. That is, 36 cells in series, on one panel, to provide a nominal 12v output. So shading ANY cell on one panel, turns one of the series cells into a roadblock, which can take down the whole panel. It won't matter which cell or how many more, when one panel goes passive, it blocks the whole array.

There would be no blocking diodes within the panel, and there might or might not be a blocking diode to prevent night time losses on the panel as a whole.

This is in contrast to other panels, which might have 72 cells, i.e. TWO series strings of cells, with a blocking diode across the output of each. In those panels, you could lose "half" a panel by shading one of the series strings. In these panels? Knock out any one, you knock 'em all out. Doesn't seem to be any issue there, just their confusion about what a "unit" is.

I hear you. I'm not familiar with the GO Sun panels, but according to some here "All modern panels have bypass diodes". So the Go Sun panels in the Wynn's video may not have bypass diodes. What about the Kyrocera panels they tested? Surely they have bypass diodes. But how could they, and yet the both panels were killed when one cell of one panel was shaded.

If not, it kinda seems to me like there may be lots of panels that don't work so well in series when shaded.

[s/v Jedi]

Here's a post with diagrams showing the advantages of series connection:

http://www.cruisersforum.com/forums/...ml#post1685893

[s/v Jedi]

And another similar thread; still not found the one I am looking for :-)

http://www.cruisersforum.com/forums/...ml#post1460791

[s/v Jedi]

And I finally found the thread:
http://www.cruisersforum.com/forums/...ml#post1249560
Read the end where the reports of how it works are posted

[Hkalan]

Hello,

The use of a combiner box at each array, on its own breaker (isolate negative and positive). To the controller, and a 200a on/off...

Take in the current available.

As they say... use the right tools, and do the job right the first time.

Only relying on a MPPT controller, is like only relying on the house bank to start your engine every time...

A few $$$ is worth it in the end. No guarantee when you are days from the next port.

1- Parrellel connections to a solar combiner w/main disconnect

2- a disconnect between the battery and the controller.

Absolute control !!!

Here is my Teak wood "costal Cruiser" restore project set up... worked for 3 years without a glitch !!

https://youtu.be/4nqk-zGayfw

Alan

[ramblinrod]

Originally Posted by DotDun

I believe what is happening in this thread is the difference between how engineers communicate and technicians speak using terms (very) loosely.

Are you freakin' kidding me? We all speak the same language. During the course of my career, I have worked with professionals in the electrical, electronic, and electro-mechanical industry in design, development, manufacturing, sales, marketing, quality control, test and repair.

I have developed and delivered technical training programs to engineers around the world.

When you make statements using terms like I pointed out in post 119, it gets engineers hackles up.

??????

Engineers don't question what happened in the video in post 30, there were no working bypass diodes, regardless if there was a 'pro techician' there nor 'Internet celebrities'. Those types of character associations don't contribute to engineering discussions and come across as babble.

I'm sorry, so how about this, "Kirchoff's First Law".

You don't seem to understand this, so how I am to know what you or others understand. This forum is not just for engineers you know. And lastly, I have worked with both good and bad engineers. Some don't know $#%^%$ from Shinola. Some don't even accept, wait for it....

"Kirchoff's First Law"

Terms and definitions deserve careful choosing in engineering discussions.[/QUOTE]
You mean like, wait for it...

"Kirchoff's First Law"

If you won't accept this very basic "technical" term and principle, how am I supposed to know how to communicate in a way that you will understand.

Regarding your post 139 that you referenced.

This is so offensive, AND incorrect, it deserves are direct response.

DOT: In your rhetoric, you stated voltage sources, which are devices, which means, in this context, solar panels.

RR: Correct. This is correct. It is Kirchoff's first law.

DOT: My point the whole time has been they don't have to be the same voltage, the lower voltage source may not participate in the application, but no police will show up due to breaking a law!

This has been your point. However your point IS, WAS, AND ALWAYS WILL BE INCORRECT. Because...2 VOLTAGE SOURCES IN PARALLEL DO HAVE TO BE THE SAME VALUE. What evidence do I have to prove it? Kirchoff's Law, and reality.

DOT, don't you find it odd, that not a single person in this thread, has come to your rescue, and confirmed your belief?

Whereas several, including Paul and TD have quietly stated to the effect that of course they have to be equal?

I would have let it go with barely a mention, until you attempted to insult me, and then when I tried to advise you of your error, you became increasingly insulting toward me.

Dude, if you're gonna dish it out, ya gotta be prepared to take it.

DOT: I went as far as to show you an example of an application where engineers play with voltage at a parallel source to provide varying results.

RR: I and many others have advised that this is incorrect. The open circuit voltage of two sources can be different, but as soon as they are connected by low impedance cable and conductors, they equalize, because they have to, as declared in... wait for it, technical term coming...

Kirchoffs' first law.

DOT: When you connect (2) sources together in parallel that have different voltages and measure with your DVM, you are reading the higher voltage source, not the lower voltage source, it is still the lower voltage source which reacts differently in the circuit from the higher voltage source until something changes.

RR: OK, let me get this straight DOT. You are claiming that two voltage sources in parallel do not have to be equal. And yet you declare that when one measures them with a DVM (digital voltmeter), they read the same.

Now isn't that interesting.

And why does the DVM read the same? Because the voltages are equal, because they have to be, as declared by , here it comes again...

Kirchoff's First Law.

DOT: The lower voltage source does not magically become an equal to the higher voltage source just because they are connected together.

RR: Nobody indicated there was any magic involved. It would be magic in your world if the voltmeter read voltage properly until two voltage sources were connected, and then it suddenly change behavior?

Of course not, the voltmeter reads the same, because the voltages are the same, and we know this because of...

Kirchoff's First Law.

DOT:So, my point the whole time has been you can connect sources with different voltages together if the application dictates doing so. No law broken!!!

RR: No DOT, this is not what you have been saying. Of course one can connect two sources of different open circuit voltages together. This is what every one in this thread has been says. It is done all the time. Every house bank that has two or more batteries in parallel. And what is the voltage across each battery after connection? EQUAL, has to be. Unless a cable or connector is "open circuit" which means, they are not connected in parallel.

Kirchoff's First Law.

Now one voltage source can present a higher of lower impedance to the other. And in the case of the batteries, the lower impedance battery will sink current from the higher impedance battery until the impedances equalize and current stops flowing between them. On initial connection, during, and after equalization, the voltage across the batteries may change, but at all times it is equal; it has to be...

Kirchoffs First Law.

In the case of identical solar panels in parallel, the shaded panel presents a high impedance to the unshaded panel. It's almost like an open circuit, like it wasn't even there. And the voltage across either panel will be equal, and why? You fill in the missing letters DOT...

K_ _ c _ _ f _ 's F _ r _ t L _ w

[Paul Elliott]

Originally Posted by ramblinrod

Now one voltage source can present a higher of lower impedance to the other. And in the case of the batteries, the lower impedance battery will sink current from the higher impedance battery until the impedances equalize and current stops flowing between them. On initial connection, during, and after equalization, the voltage across the batteries may change, but at all times it is equal; it has to be...

Kirchoffs First Law.[/COLOR][/B]

While I agree with the rest of your post, shouldn't the above read "voltage" rather than "impedance"???

As for paralleled voltage sources, I have to think that the confusion is because DotDun is considering the open circuit (unloaded) voltage of a particular source as that source's "voltage", regardless of what happens when it is connected to a load or other source.

When analyzing this type of electrical circuit, we often visualize a source as an "ideal voltage source" in series with a "source impedance". You can do anything you like to the output of this source, including shorting it or connecting it to a higher voltage, and regardless of the measured voltage at the output terminals, the internal "ideal voltage source" will not be affected.

This is useful in analysis, but perhaps not so much in the solar panel discussion. For one, the solar cell is more usefully modeled as a current source in parallel with a diode (and with additional series and parallel resistors). Regardless of the model, it is *not* a voltage source, and trying to analyze it as one is not really useful.

I hope that nobody actually thinks that there can be two different voltages existing at the same time across two nodes. I think we are disagreeing as to the point where the voltage is actually measured.

[ramblinrod]

Originally Posted by Paul Elliott

While I agree with the rest of your post, shouldn't the above read "voltage" rather than "impedance"???

As for paralleled voltage sources, I have to think that the confusion is because DotDun is considering the open circuit (unloaded) voltage of a particular source as that source's "voltage", regardless of what happens when it is connected to a load or other source.

When analyzing this type of electrical circuit, we often visualize a source as an "ideal voltage source" in series with a "source impedance". You can do anything you like to the output of this source, including shorting it or connecting it to a higher voltage, and regardless of the measured voltage at the output terminals, the internal "ideal voltage source" will not be affected.

This is useful in analysis, but perhaps not so much in the solar panel discussion. For one, the solar cell is more usefully modeled as a current source in parallel with a diode (and with additional series and parallel resistors). Regardless of the model, it is *not* a voltage source, and trying to analyze it as one is not really useful.

I hope that nobody actually thinks that there can be two different voltages existing at the same time across two nodes. I think we are disagreeing as to the point where the voltage is actually measured.

Hi Paul,

I applaud our moderation. ;-)

I have no trouble with considering a solar panel a current source, though I feel that for a discussion about what happens to the voltage of ANY electrical device (passive or active) when connected in parallel, it should be easier for anyone to
understand, if we call it a DC voltage source, but it really doesn't matter what we call it, we can leave it as just "2 solar panels", for this discussion.

I do have an issue with what you refer to as an "ideal" source. In my opinion an "ideal source" (voltage or current) has infinite impedance across it, and zero impedance through it. The "equivalent circuit" of a real world source, will not be ideal, but this is totally irrelevant to this topic under normal circumstances.

My belief is that Dot does not accept Kirchoff's Law with respect to a parallel circuit, and then attempted to ridicule me for resorting to non-technical communication, after a long bout of refusing to accept technical explanation.

I've got real thick skin (which enables me to persist in these situations) and am not afraid to stand up to someone who resorts to nastiness, rather than just admitting error and moving on.

It is in my nature to help people learn. I have been employed as a Technical Trainer for various corporations, and I conduct "Marine Electrical Sytem" seminars for various boating groups. I have been communicating with every level of electrical, electronic, and mechanical professional personally and professionally for over 35 years. If an individual has objections to the way I communicate technical principles, to help people learn, I believe the issue
is theirs, not mine.

In my professional capacity, I will bend over backwards to help someone learn, especially when it will help them in their work. I transfer this quality to help boaters learn about their boat systems, in order that may realize improved safety and enjoyment.

This is why, even when I am working on customer boats to the we hours of the morning, I take time out to contribute to this forum. When the thanks I get is insulting response I take serious issue with that. I understand that it is somewhat of a character flaw that I can't just let it roll off, but someone has to stand up for what is right, and true, and as I said, I have pretty thick skin.

L

[Paul Elliott]

Originally Posted by ramblinrod

I do have an issue with what you refer to as an "ideal" source. In my opinion an "ideal source" (voltage or current) has infinite impedance across it, and zero impedance through it.

I'm not following you here. Any two-node passive device has just one impedance, measured at the two nodes. There are not different "across" and "through" impedances.

An ideal voltage source has zero impedance.

An ideal current source has infinite impedance.

Real-world devices are not ideal. We add series resistance to an ideal voltage source for a first approximation of a real-world voltage source. We add parallel resistance to simulate a real-world current source.

Solar cells are even more complicated, but the usual model consists of a current source, a diode, and two resistors:

[hellosailor]

Dotdun's attachment back around #129 shows pretty clearly a bypass diode external to EACH PANEL in a series+parallel panel array. And yet, there's mention of bypass diodes as if they were on each CELL within a single panel. And at the same time, back in Dotdun's attachment, there's the statement that bypass diodes are not necessary for 12v panels, and blocking diodes are not necessary when modern controllers already do that job.

So there's room for plenty of confusion here. Whatever was a "bad diode" connection IN one panel, shouldn't or couldn't have been a bad diode unless that panel had 72 cells, or bypass diodes on each cell then, yes? And it the attachment was correct, then it would be a total waste of time, money, and reliability, for manufacturers to install ANY diodes AT ALL in panels that were going to be either used solo for 12v, or put into arrays with modern controllers.

Every diode is a voltage loss, and that's a power loss, and an expense, and a point of potential failure. Diodes are therefore like unicorns: Don't assume there's one behind the bush, because there's good reason it might not be.

[TeddyDiver]

Shaded panel in parallel is a source of nothing...

[Paul Elliott]

Originally Posted by hellosailor

Dotdun's attachment back around #129 shows pretty clearly a bypass diode external to EACH PANEL in a series+parallel panel array. And yet, there's mention of bypass diodes as if they were on each CELL within a single panel. And at the same time, back in Dotdun's attachment, there's the statement that bypass diodes are not necessary for 12v panels, and blocking diodes are not necessary when modern controllers already do that job.

Bypass diodes are a very good thing if you have a series string of multiple panels (12V or otherwise). They don't help you much if at all with a single panel, or paralleled panels, unless you have high-voltage panels with more than two bypass diodes.

Originally Posted by hellosailor

So there's room for plenty of confusion here. Whatever was a "bad diode" connection IN one panel, shouldn't or couldn't have been a bad diode unless that panel had 72 cells, or bypass diodes on each cell then, yes?

I don't understand this above. Can you re-phrase for me?

Originally Posted by hellosailor

And it the attachment was correct, then it would be a total waste of time, money, and reliability, for manufacturers to install ANY diodes AT ALL in panels that were going to be either used solo for 12v, or put into arrays with modern controllers.

A modern controller will not solve the problem of a shaded panel (or cell string) in an array of series-connected panels -- unless the panels have bypass diodes. Whether a single diode bypassing an entire panel, or multiple diodes bypassing panel sub-strings is better depends on the nature of the shading. As has been suggested, adding a panel-bypass diode will reduce the diode-drop loss in a complete shading situation.

Originally Posted by hellosailor

Every diode is a voltage loss, and that's a power loss, and an expense, and a point of potential failure. Diodes are therefore like unicorns: Don't assume there's one behind the bush, because there's good reason it might not be.

Bypass diodes do not introduce a voltage loss in an unshaded panel. They do introduce loss when they are actively bypassing a shaded section, but this loss is certainly an improvement over the virtual open-circuit of a fully-shaded cell.

Series diodes do introduce loss in any configuration. Modern controllers generally prevent the battery from draining into a dark panel, but I suppose there could still be loss through a dark panel that is in parallel with an illuminated panel. However, I didn't measure any appreciable leakage current when I tested that configuration a couple of days ago. My panels had no series diodes, just bypass.

[OceanSeaSpray]

I have taken the time to read the thread from start to finish... well!


I think one not-so-clear element in series configurations is that real shadows seldom affect one string only... in fact, a line of shadow across a panel can easily cut all strings. When this happens, the whole panel gets (must get) bypassed, i.e. no effective advantage on a parallel circuit (in fact slightly worse due to bypass diode losses). If this doesn't happen, then the Jedi's advantage can exist, but the bypass diodes become very important.
Some high-end HV Sunpower panels have one diode per cell from memory, and even then strings should still be bypassed and whole panels should be bypassed to eliminate the losses from cascaded diodes.
A lot of things need to be right for series arrays to operate well and I have seen many on-board installations where going from series to parallel resulted in a huge improvement because bypassing simply wasn't successful enough in practice.


It is a misconception to think that a shaded panel in parallel will contribute "nothing". Solar cells behave very much like current sources and their voltage rises very steeply with illumination. They reach the system charging voltage very quickly. The current they deliver then is dependent on light intensity.


The whole Kirchoff's Law argument wouldn't have taken place if panels were viewed as current sources in series with a blocking diode (which they are, one way or another). A fully shaded panel in parallel doesn't contribute to or influence the voltage at the junction because its cells are de-facto isolated. It drops out when its OCV would become smaller than the charging voltage.


My panels have no visible series diodes and I only measured 0.14mA of reverse current at night when I once checked.

[Paul Elliott]

Originally Posted by OceanSeaSpray

It is a misconception to think that a shaded panel in parallel will contribute "nothing". Solar cells behave very much like current sources and their voltage rises very steeply with illumination. They reach the system charging voltage very quickly. The current they deliver then is dependent on light intensity.

Agree. This of course depends on the degree of shading. In a clear sky the shadows can be pretty deep, and the current from a shaded panel will be very low. With haze or white clouds in the sky, a typical shadow becomes diffuse, and the current reduction will be less significant.

You have to completely and fully shade a cell or string or panel (as I and others have done in some of our testing) before the contribution from a shaded parallel panel drops to zero.

[noelex 77]

I think this is a very important topic. We should know which option is best, serial or parallel. If one method of connection gives, on average, the best results for a shaded panel then this method is likely to be the best system for nearly all sailboats. So the conclusions are important for everyone installing a solar system.

I think the thread lost its way a little, but my reading of the posts is that overall the consensus is that serial wins. This mostly stems from the theoretical calculations showing the output is better for serial connected panels if one string is shaded. These calculations are mostly correct, although some of the more minor effects such as the voltage drop over the blocking diodes have been ignored. However, I don't think this one isolated theoretical case is representative.

To mathematically model solar panel output with shade is incredibly complex and picking one isolated case and drawing conclusions from that is wrong. There are cases where the theoretical output of parallel connected panels beats series connected panels. There are even scenarios where wiring the panels in series with a shaded panel can drop the output of an unshaded panel, occasionally significantly.

If we want to theoretically model solar panel output we need to consider all these options. Diffuse shadows are particularly complex and the results are very different depending on the type of shadow chosen.

Let's take a simple example where parallel connection wins convincingly. One panel is in full sun and one has a diffuse shadow. The two panels can have their own controller or be connected in parallel or series.

The maximum output of the panel in full sun is 16v @ 5A = 80 w. The maximium output of the panel in diffuse shade is 15.5v @ 3 A = 46.5 w

Seperate controller for each panel. Result 111.1w going into the batteries:

Here we have the full 126.5w less the self consumption and inefficiencies in the controllers. As we have two controllers, this may amount to around 12w so the output is now 114.5w. Let's say 3% loss for wiring and resistance over circuit breakers and we have 111.1w going into the batteries.

Parallel connection. Result 114.1w going into the batteries.

Here we have a single MPPT controller. This controller must pick the optimum Vmp for the array as a whole. This would fall very close to the lower voltage so for simplicity let's assume it is 15.5v (the actual optimum Vmp will be slightly higher). The current output of the first panel will rise slightly at this lower voltage so panel one produces 15.5v @ 5.1A = 79.1w. Panel two produces 15.5v @ 3 A = 46.5 w. We only have a single controller so the self consumption and inefficiencies in the controller will be less, let's say 8w. The output is now 117.6w. Let's say 3% loss for wiring and resistance over circuit breakers and we have 114.1w going into the batteries.

Series connection. Result 67.7w going into the batteries.

We have a single MPPT controller. The optimum Vmp is 16 + 15.5 = 31.5 v. The current running through the two panels must be identical, so this is 2.5A. Total output is 31.5 x 2.5 = 78.8w (the other possible Vmp is close to 16v; the output here will be similar). The loss in the single controller will be slightly higher because of the greater voltage conversion. Say 9w. The output is now 69.8w. Let's say 3% loss for wiring and resistance over circuit breakers and we have 67.7w going into the batteries.


The above calculations are slightly simplified, but I have taken into account most of the factors for this one scenario. You can see parallel connection wins. It even beats separate controllers for each panel, but importantly series connection is by far the worst option in this example. In this series example the second panel contributes nothing. However, it would be silly to extrapolate this one example to all cases of shade. It is just one example. For other types of shade two controllers will be best and for others series connection will be best. So the above calculations while accurate are a waste of time unless we model all the possibilities and integrate them into a realistic model taking into account the likelihood of these options. This is incredibly complex.

Basically, I think we cannot use theoretical calculations to determine if series or parallel connection is best. The argument used in this thread by others is that a single example where series connection beats parallel connection means we should draw the conclusion that series connection is best, is a false conclusion. Just like the above example should not be used to conclude parallel connection is best.

So we need experimental data not theoretical calculations. So keep the data coming Paul . This is the only way we will answer the question. However, even here we need a sophisticated experimental protocol. Producing shadows by laying sheets of card on the panels is not the same as shadow from the boom where a reasonable amount of light is still falling on the shaded area.

My suspicion is that parallel wins, but not because of the above calculations. I have no vested interest in one method of connection. I would just like to know the right answer.