Installing serial solar panels

[TeddyDiver]

Originally Posted by ramblinrod

OMG.
2 solar panels connected in parallel create 2 common nodes.
One at the positive connection, one at the negative connection.

The panel presents a high impedance to the circuit. If there is
sunlight on the panel it becomes a voltage source and
develops a voltage across it which passes current through it directly proportional to the amount of sunlight up to the panel limit (which is sorta a fixed value but not really).

The voltage between the common nodes is one and the same. Has to
be; it is a parallel circuit. If one panel is shaded, current passed (in the positive direction only) drops. The voltage between the common nodes may change slightly, but it is still a common node. One voltage. A common node can't be two different voltages at the same time, nor can current travel in opposite directions at the same time. They either contribute or negate. Because a panel that isn't contributing current presents a high impedance to one that is, the shaded panel can't negate (unless there is something wrong with it)

This is solar panels 101, first lesson. If after 3 days of discussion you still don't get it, I have to accept you may never.

Paragraph one and two agree, but in paragraph three the thruth is quite well encrypted between the lines. And that's what I've been trying to say all the time..

So you finally admit there's nothing wrong in post 39?

BR Teddy

[ramblinrod]

Originally Posted by ramblinrod

The reference statement is 100% correct if there are no bypass diodes. If there at bypass diodes, each panel in series can deliver different current.

Actually, I have to correct my own response second sentence, which was not 100% correct.

If the bypass diodes are external to the panel, each panel in series can have different current passing through it. If the diodes are internal, (in relation to the measurement points) each panel must pass the same current.

In either case, if a panel cell string is shaded and its impedance increases, the current will bypass the shaded cell string through the diode (internal or external) to the next cell string.

However, I am still perplexed by the Wynn's testing as linked to the YouTube video in post 30.

I have performed testing with small semi-flexible panels in series and parallel which produced identical results as shown in the video. I assumed this was because the panels I was using did not have bypass diodes.

However, the panels in the Wynn's test (actually connected and under the supervision of a pro solar panel tech, produced the same results, shading even one cell of a parallel panel killed the output of that panel, whilst shading the a single cell of a panel connected in series, killed the output of all panels connected in series.

First thought is that the solar panels under test did not have bypass diodes, but that apparently is not so. It was repeated with two different brands of panels, that apparently do have bypass diodes.

So that leaves me with 2 possibilities:

1. A huge coincidence that the bypass diodes of the 2 sets of different brand panels under test, apparently randomly selected from a pile, were defective.

2. That bypass diodes do not function in practice as expected in theory.

While I do not necessarily consider the test of a pair of You Tube sailing celebrities to be 100% definitive; the methodology, that was reasonably evident to the viewer and appears sound, and supervised by a tech, is pretty compelling.

Yet we have the test results provided by Paul, and the testimony of some boat owners on this forum, that indicate differently.

To me, it remains a mystery, for which I do not have a reasonable explanation, based on anything more than conjecture.

And then we heard from Paul, that a portion of his testing didn't go as expected, due to a bypass diode problem.

So my position, based on this evidence is, until I have sufficient further evidence available to me to conclude otherwise...

"Series shaded panels may not always function as expected; it is possible that a series shaded panel could kill the output of other panels in the string."

I s'pose this may be due to missing, improperly connected, or defective bypass diodes, or some other logical reason why the bypass diodes don't function as expected, that I just haven't considered...yet.

[ramblinrod]

Originally Posted by TeddyDiver

Paragraph one and two agree, but in paragraph three the thruth is quite well encrypted between the lines. And that's what I've been trying to say all the time..

So you finally admit there's nothing wrong in post 39?

BR Teddy

As mentioned in a prior post, we have conflicting reports about what happens with shaded panels in series, and at the very least, we have reliance on bypass diodes that have been reported in several cases to be unreliable.

I do concur that in theory, unless there is something I am not considering, that a series connected panel with a shaded string equipped with a bypass diode, and additional unshaded strings, should contribute current. This is confirmed in by Paul's testing and reports by some others. However it is in direct conflict with the testing performed on the YouTube video in post #30.

It's certainly a mystery.

What is troubling, is that I recall another post around that time, declaring that in parallel connection, the shaded panel would kill the output of the unshaded panel. This was the post that I had the greatest objection to. That post is no longer present and the pagination of posts has changed, indicating at least one post has been deleted.

[Cotemar]

This is why a lot of cruisers use a separate MPPT controller for each solar panel.

A lot of cruisers are using multiple Genasun GV-10 MPPT controllers.

1) No Serial or Parallel wiring issues.

2) System has redundancy.
.
.

[Paul Elliott]

Originally Posted by ramblinrod

And then we heard from Paul, that a portion of his testing didn't go as expected, due to a bypass diode problem.

The bad bypass diode connection did not affect my testing. I discovered the problem during the tests, and so used the "broken" panel only in the non-shaded and completely shaded states. I used the properly bypassed panel as my partially-shaded guinea pig.

I haven't watched that video, but will check it out. My testing shows that the current from an unshaded string will pass through the bypass diode of a shaded string.

Some testing I plan to do will look at the effects of (say) 50% shading on a string, and the effects on total current in the series and parallel connected situations. I expect to find some cases where paralleled panels deliver more power than series-connected ones, but I need to look closely at when the bypass diodes start to conduct. It's going to take me at least a week before I can start those experiments. In the meantime, I'm going to look into cell simulation models so I can run the tests on my computer.

Of course delivering power into a resistive load is not the same as delivering power through an MPPT controller -- that depends on the controller's input range.

All this testing is mostly for my amusement, since I'm already switching my modest three-panel setup from parallel to per-panel MPPT regulators.

[Paul Elliott]

Originally Posted by Cotemar

This is why a lot of cruisers use a separate MPPT controller for each solar panel.

A lot of cruisers are using multiple Genasun GV-10 MPPT controller

1) No serial or parallel wiring issues.

2) System has redundancy.
.
.

This is what I'm doing, using the Genasun controllers. I'm also including V/A monitoring of the panel outputs and V/A/Wh monitoring of the MPPT output, just because I want to see what's happening.

[ramblinrod]

Originally Posted by Cotemar

This is why a lot of cruisers use a separate MPPT controller for each solar panel.

A lot of cruisers are using multiple Genasun GV-10 MPPT controllers.

1) No Serial or Parallel wiring issues.

2) System has redundancy.
.
.

This is certainly a design consideration but as indicated in many other posts, there are pros and cons to every design consideration.

Pros:

1.Optimal MPPT tracking (assuming the controller MPPT tracking capability
good. (Multiples of poor is still poor.)

2. Redundancy, if one controller goes down the rest may operate. (This is dependant on the reliability of the controllers. If the small
Individual controllers fail frequently, and the large
Central controller never fails, one is worse
Off with multiple unreliable controllers.

3. Smaller gauge cable and lower cost components.

Cons:

1. Lots more cabling, especially if the distance between the panels and
Controllers is significant.

2. Possible higher total controller costs

3. Possible lower functionality than a larger
Controller.

4. Need for additional equipment to determine the total array performance.

So for some, individual controllers may be the "best" set of
Compromises. For others no so much.

[Vladis]

Hello,

We are talking about electric circuits and I only see words and words. Not a single circuit diagram. Electric circuits are not subjective. They follow exact physical principles and most of the discussions I found here a consequence of arguments not being properly explained. Diagrams please!

Regards and sorry if it is only me that have difficulty in following the discussions, Vladis

[DotDun]

Originally Posted by ramblinrod

As mentioned in a prior post, we have conflicting reports about what happens with shaded panels in series, and at the very least, we have reliance on bypass diodes that have been reported in several cases to be unreliable.

I do concur that in theory, unless there is something I am not considering, that a series connected panel with a shaded string equipped with a bypass diode, and additional unshaded strings, should contribute current. This is confirmed in by Paul's testing and reports by some others. However it is in direct conflict with the testing performed on the YouTube video in post #30.

It's certainly a mystery.

What is troubling, is that I recall another post around that time, declaring that in parallel connection, the shaded panel would kill the output of the unshaded panel. This was the post that I had the greatest objection to. That post is no longer present and the pagination of posts has changed, indicating at least one post has been deleted.

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

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

Engineers have no problem with post 39, it's well written and based on engineering principle.

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 engineeering discussions and come across as babble.

Terms and definitions deserve careful choosing in engineering discussions.

[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 (

http://www.youtube.com/watch?v=1qD3mN8VotQ

) 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.

[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.

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

Engineers have no problem with post 39, it's well written and based on engineering principle.

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 engineeering discussions and come across as babble.

Terms and definitions deserve careful choosing in engineering discussions.

Gimme a break.

This is the most offensive thing you've said to date. You refuse to accept Kirchoffs first law, you continue to suggest that 2 voltage sources in parallel must be the same voltage (under normal circumstances)?

Do you have an engineering degree from an accredited teaching facility? If so, you have been schooled in this thread by an Elecronic Engineering Technician, and it certainly wouldn't be the first time I've schooled people with an Electrical or Electronic degree, up to the PHD level.

[DotDun]

Originally Posted by ramblinrod

Gimme a break.

This is the most offensive thing you've said to date. You refuse to accept Kirchoffs first law, you continue to suggest that 2 voltage sources in parallel must be the same voltage (under normal circumstances)?

Do you have an engineering degree from an accredited teaching facility? If so, you have been schooled in this thread by an Elecronic Engineering Technician, and it certainly wouldn't be the first time I've schooled people with an Electrical or Electronic degree, up to the PHD level.

Thank you for confirming for all who read it what I've claimed is the problem with having a conversation with you.

[ramblinrod]

Originally Posted by ramblinrod

Gimme a break.

This is the most offensive thing you've said to date. You refuse to accept Kirchoffs first law, you continue to suggest that 2 voltage sources in parallel must be the same voltage (under normal circumstances)?

Do you have an engineering degree from an accredited teaching facility? If so, you have been schooled in this thread by an Elecronic Engineering Technician, and it certainly wouldn't be the first time I've schooled people with an Electrical or Electronic degree, up to the PHD level.

Whoops, my mistake in this post. Dot keeps insisting that two voltage sources in parallel can be different voltages. I guess the notion is so absolutely absurd
to me that I have difficulty writing it.

[ramblinrod]

Originally Posted by DotDun

Thank you for confirming for all who read it what I've claimed is the problem with having a conversation with you.

Dot, two voltage sources in parallel, low impedance cables, connections, and load. One is at 12 Vdc. What is the voltage across the other one?

Call in some help if you need it, because any technician, technologist, or engineer (you see, we all speak the same language) worth his salt will tell you, 12 Vdc.

[s/v Jedi]

So what have we learned in this thread:

1. The Jedi is right: series outperforms parallel in a 1-cell shading test with panels that have bypass diodes like every panel had for the past 25 years.

2. #1 is not only scientifically supported as explained multiple times, but also demonstrated by practical test, right here by a member in this thread, as well as many times before over the years.

3. Rambler denies the above points, but has shown that he just doesn't fully grasp what he is rambling about. He will keep denying that partial shading of a panel in parallel setup kills it's output power, even going so far as claiming that the bypass diodes will work in a parallel setup.

4. It was claimed that the load in the test was not high enough, which keeps the voltage too high but that is just a load of cr@p because with one string killed by shading, the Vopen is only 11V so even a direct connection to the battery will make it discharge the battery instead of charging it (if no blocking diode like claimed). On the other hand, in a series setup we still have 33V (yeah yeah minus diode loss) with three strings pumping juice into the controller.

5. The separate Genasun controller per panel: I think I was the first one recommending this to somebody that got a whole bunch of people claiming it was wrong. The guy actually did buy the Genasun's, connected everything and then actually reported back the results which was even better than what I had claimed. I tried to find the thread but my Internet is flaky at the moment...

I think it's up to Rambling now to create a test setup with two parallel panels, shading a cell and ending up with still 75% + output or so he claims, then repeating the test with the panels in series and demonstrate the 0% output like he claims. For now, the test proved that parallel lost 50% of array output and series only 25% of array output. The difference with my example was that I used 3-string panels which improves the results for series even more