Answering some solar panel with MPPT controller ?

[ktmrfs]

Here is some real world data on solar panels when used with a MPPT controller to answer some question I, BFL and others have.
This report is focused on the questions we had regarding solar panel use with MPPT controllers.

AND FOR THOSE WHO MAY WANT TO START A MPPT VS. PWM CONTROLLER DISCUSSION PLEASE DO NOT TURN THIS INTO ANOTHER MPPT VS. PWM DISCUSSION, FLAME ETC. IF YOU WANT TO CUSS OR DISCUSS PWM VS. MPPT CONTROLLERS GO START ANOTHER THREAD.

OK, now to the questions I was looking to answer for:

1)BFL’s question; “Can you look for a peak in Voc or Isc to determine optimum panel orientation?”

2)Can one predict accurately output power by measuring MPPT input voltage and panel output current?

3)In late fall at high latitudes what % of nominal power can I expect to get from the panel under optimal light?

4)Can one predict power loss between the panel and controller and predict the effect on output power (charging current and watts) with reasonable accuracy?

5)If one knows the resistance and current between the panel and controller can one predict the power to the battery with Series vs. parallel panels and longer or higher resistance paths between the panel and controller? (really an expansion of the above question)

6)Can one measure Temperature effects on Panel output power?

7)What is typical MPPT controller efficiency, or for this paper, the efficiency of a Morningstar sunsaver 15 MPPT controller operating in the 130 W region???

8)What if I just lay the panels flat, how far from peak power will I be?

9)With series connected panels what is the effect of shading on one of the panels?

10)Since I also have another trailer with PWM controller and use the same portable panels with this trailer, I was curious what effect a long (90ft) run between the panels and PWM controller has when used with a PWM controller. In this case I don’t have the choice of going series connected panels, that’s counterproductive. So staying with parallel connected panels is the only choice. Can I predict the results?

Rather than one LONG……. Post, I will break the post into sections. The next section is focused on assumptions, test conditions, test equipment, and test methodologies.

Then there are individual posts for each of the 10 questions.

And for those reading now, please delay any responses till you see a “That’s All folks” in my last post so my posts can be in sequence uninterrupted. I should have them all posted within an hour.

[MEXICOWANDERER]

Remember, a dime an hour income difference ain't squat.

A dime for 10,000 hours is......squat...

[Canadian_Rainbi]

I've got a headache! 😞 I did manage to follow it all though! 🙂

Thanks for all the work. It answered some of the puzzling aspects of my system. Or rather "collection of parts strung together and hoping for the best". :S

I really have to re-think and modify my setup, I have way too much voltage drop. . . on the output. I guess that is the "exercise for the student". :h

I already made my panels tiltable 4 ways and shortened the input cables. The output though has serious issues. I'm about to post those on a new thread.

[ktmrfs]

Salvo wrote:
You may have a measurement error. Where are you measuring Vmp? It should be directly at the panel, not at the controller.

In any event, the controller is starved for volts. This system should show better results if battery voltage is dropped to 12.5V.

ktmrfs wrote:
Post #5


Parallel connection with 90ft of cable
Vmpptin=14.60V (TX3DMM)
Impptin=8.3A (DM330)
MPPT power in=14.60*8.3=121W
Vbat= 13.50V (TX3DMM)
Ibat=9.0A
Power to battery= 121.5W

No signficant measurement error. Math and calculations are very close.
As noted in the writeup I'm NOT measuring voltage at the panel, I'm measuring the voltage and current AT the MPPT controller input (Vmpptin, Impptin) to determine available power to the controller. Yes, it is getting starved for voltage, the 90ft of approx #8 cable has a 1.3V drop across it, so parallel connected panel output voltage is 14.6+1.3=15.9V. That 90 ft of cable in parallel connection is about a 1A disadvantage in charging current compared to a series connected set of panels with 90ft of cable. Driven by system losses between the panel and controller that are noticeably higher for the parallel case than the series case with the same cable length.


Below spec'd peak Vmp & Imp, yes, but it's fall, low sun angle, high latitude.

And battery voltage during charging is determined by the battery state of charge, and charge current. In this case with batteries at near 80% SOC 9.0A battery input current, battery terminal voltage rises to 13.5V. Using a 12.5V voltage source, battery acceptance current would likely go down, not up.

Now if the batteries were near 50% SOC, I suspect I would see higher current, because I would also see a lower battery voltage, but the power (watts) to the battery wouldn't appreciably change, it would still come out near 120W.

As to lowering the voltage to say 12.5V, sorry, when charging batteries you can choose the charging current or the charging voltage you CANNOT set both of them. The battery has some internal impedance. Solve the loop equation, since you can't change battery impedance, if you SET the voltage, the current will be determined by the loop equation solution. If you SET the current, the voltage will be set by the loop equation. You CANNOT set both arbirarily. Setting the voltage to 12.5 V would LOWER the charging current, not raise it.

think about it. If I "lowered" the converter voltage to 12V I wouldn't get 10A going into the battery, The Battery would likely be supplying current TO the controller and probably blow the fuse inline with the converter!!

The controller is behaving during bulk charging mode as a constant current device.

[Salvo]

You may have a measurement error. Where are you measuring Vmp? It should be directly at the panel, not at the controller.

In any event, the controller is starved for volts. This system should show better results if battery voltage is dropped to 12.5V.

ktmrfs wrote:
Post #5


Parallel connection with 90ft of cable
Vmpptin=14.60V (TX3DMM)
Impptin=8.3A (DM330)
MPPT power in=14.60*8.3=121W
Vbat= 13.50V (TX3DMM)
Ibat=9.0A
Power to battery= 121.5W

[12thgenusa]

Salvo wrote:
WTF?

12thgenusa wrote:

:R

It means, give it a rest. You made your point in the other thread. Let this thread be what it is.

[Salvo]

WTF?

12thgenusa wrote:

:R

[CA_Traveler]

Thanks for the further clarification.

I'm sure everyone that reads your report will appreciate the effort you put into it, the questions you answered and the homework exercises for the rest of us!

[ktmrfs]

CA Traveler wrote:
In Post #3 and Post #9 ktmrfs observed that serial panels produce the current set by the highest output current while the second shaded panel has reduced voltage.

I expected the current to limited by the shaded panel. Ie At 1000W a 8A panel would produce 4A at 500W based on the typical IV curves.

So how did ktmrfs see 4.1A in direct sun and 4.1A with one panel in the shade and reduced voltage for the shaded panel?

CA Travel. Great questions.

In retrospect, I probably should have made some MPPT voltage measurements, it would have helped explain what is likely going on with series panels that are shaded. sorry for the oversight.

And at the same time had I done a parallel shaded output, it also in retrospect would have given better insight in how panels operate in series vs. parallel.

Here is what I believe is happening.

If it was a SINGLE panel, or two panels in parallel with one completely shaded, then yes, I would have expected panel output current to drop roughly in half, just as you outlined.
Effect would be the same, roughly 1/2 the current to the battery. That's what I've qualitatively observed in use with parallel panels.

I didn't directly measure the second shaded panel voltage, but the voltage set by the MPPT controller for max power dropped. I suspect it dropped from about 30V down to 15V or so into the MPPT controller. Since the panels are in series the current remained 4.1A, but the panel combined voltaged dropped probably to what a single panel would output so power dropped roughly in half. Output impedance of the shaded panel is still low enough to not impact the current output capability.

So series or parallel, shading 50% of the total panel area cuts charging current in half. In the parallel case, or a single panel case, current to the contoller will drop by 50%, but panel output voltage remains almost constant. in the series case panel ouput voltage drops by 50% but current remains almost constant and ends with a roughly 50% drop in charging current. In either case panel output power drops by about 50%.

Note, this is when a whole series panel is shaded, or how one of two parallel panels is completely shaded. how partial shading of a panel or panels effects output I didn't test.

Interesting note. If instead of shading a panel, you completely cover it with say a thick blanket, that is another story entirely, at least with the series case. in that case power and charging current drops to almost zero. Even in shade the panel will get some solar energy, enough to keep the panel output impedance "low enough" but not supply much current. When NO solar energy reaches the panel it's output impedance must go quite high. My first attempt at shading was t throw a big thick towel over the panel, and current went to less than 1A. removed the towel, rotated the panel 180 degrees so it was shaded, then current dropped by 50%. Picked the panel up and placed it in the shade between trailers, current stayed at the 50% level.

However, in a parallel panel case I suspect putting a towel over one panel will just drop the current in half. The covered panel has high output impedance but is in parallel rather than series and so it's effect is minimal.

But this kind of shading isn't real life either, so in a practical sense, it's a moot point.

[CA_Traveler]

In Post #3 and Post #9 ktmrfs observed that serial panels produce the current set by the highest output current while the second shaded panel has reduced voltage.

I expected the current to limited by the shaded panel. Ie At 1000W a 8A panel would produce 4A at 500W based on the typical IV curves.

So how did ktmrfs see 4.1A in direct sun and 4.1A with one panel in the shade and reduced voltage for the shaded panel?

[12thgenusa]

Salvo wrote:

ktmrfs wrote:
Post #5

For series, cable power loss is I^2R, or V*I or V^2/R. To satisfy Salvo I’ll use V*I to calculate the power loss.

You haven't understood the previous discussion. I don't care how you calculate cable power loss. That's never been an issue.

Note that the parallel case with double the current has 4x the power loss, not double the power loss. (11.6W vs. 2.9W)

Abain, that's never been the issue. If you double current, cable power loss will be 4 times has high; no matter what configuration you use.

The intelligent way to figure out what size cable is needed is to use a 1% or 2% voltage drop rule of thumb. Whatever percent voltage drop is used, battery charging current will drop by same amount, not 4 times that amount.

Sal

:R

[Salvo]

ktmrfs wrote:
Post #5

For series, cable power loss is I^2R, or V*I or V^2/R. To satisfy Salvo I’ll use V*I to calculate the power loss.

You haven't understood the previous discussion. I don't care how you calculate cable power loss. That's never been an issue.

Note that the parallel case with double the current has 4x the power loss, not double the power loss. (11.6W vs. 2.9W)

Abain, that's never been the issue. If you double current, cable power loss will be 4 times has high; no matter what configuration you use.

The intelligent way to figure out what size cable is needed is to use a 1% or 2% voltage drop rule of thumb. Whatever percent voltage drop is used, battery charging current will drop by same amount, not 4 times that amount.

Sal

[BFL13]

Yes, only took a minute or two. As soon as I got it out in the sun I measured Voc and it was 37 but while I still had the meter on, it started to drop in voltage. I wondered how low it would go before it stopped. Got to 36. Definite temp effect as also felt panel getting quite warm.

[NinerBikes]

ktmrfs wrote:

NinerBikes wrote:

ktmrfs wrote:
Post #10

10)
This is for the case of resistance between the panel and controller. The effects of a long run with PWM controller between the controller and battery bank is not part of this investigation. That experiment is “left as an exercise to the reader”. Any volunteers???

How would this vary between MPPT and PWM, assuming same input voltage and same input amperage coming out of the devices to the battery? Obviously, the thicker the wire and shorter the run, being DC current, the lower the losses of watts/voltage and amperage. This is of concern if you use long runs of wire between portable solar cells and your batteries, I would assume?

As you noted, "short and fat wins over long and skinny every time."

However there are several things that come into play that could have different consequences on MPPT vs. PWM with a higher resistance between the controller and battery.

I gave it signficant thought, and really couldn't remotely convince myself what the effects would be on each system, so rather than speculate, I left it as "an exercise to the reader".

similar to some of the textbooks I had in college where MOST answers were in the back. the ones without an answer often said. "left as an excercise to the student" I quickly figured out that the prof that wrote the book didn't give the grad student solving the problems the solution, and the grad student finally gave up trying to solve the problem and "left it as an excercise to the student".


If anyone has some actual comparison data, or even data for increased resistance effects between the controller and battery for EITHER controller type, please post.

It's good to listen to EE's post up, when not having idea's squashed by bean counters and the bottom line. Although, at 140 watts and 14.7v coming out at around 8 Amps, 10 gauge wire probably does not suffer too much loss with a 10 to 15 foot run... % wise, at least in my application. That might change if I decide to park near the aspens or pine trees in the shade and need to reach out a bit.

[ktmrfs]

BFL13 wrote:

harold1946 wrote:
My question is how were you able to maintain 70 degrees panel temperature throughout the testing?

He said 70F was ambient throughout and panels heated up right away so presume also maintained steady temp at 28C above ambient as he reported.

In my recent test of a panel, taking it out from the cool temp garage, it showed 37v Voc immediately, as rated, but very quickly that dropped to 36v as panel heated up. You could feel the glass was now warmer than it was at first. Happened fast.

BFL, I didn't pay lots of attention to time, but I'll bet my panels took less than 5 minutes to warm up in direct sun. Is that consistent with what you observed??

And, yes, one nice thing is that the daytime temps from about 11AM PDT through about 3PM PDT were very constant, right in the 70F range. Much better than trying to do experiments in mid summer when daytime temps are constantly changing.

And I did take panel temps with my IR thermometer every half hour or so and they stayed pretty constant. IR thermometers,(at least mine) is notorious for reading variability. 10F or so on repeated measurements isn't uncommon on mine, but they all were in the same range.