Answering some solar panel with MPPT controller ?

Post #6

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

Our ambient temp remained about 70F during the whole test. If there is a temperature effect, the thermal time constant on the panels is short enough that by the time the panels were set up and ready for charging measurements, they had reached equilibrium Using an IR thermometer I measured the panel front and back temperatures. Readings were consistently in the 115F range. That’s a 45F (28C) rise.

Conclusion: Predicting or monitoring temperature effects on a panel certainly is not trivial, and could end up being an exercise in futility. Solar power is changing during the day so to separate out changes in solar power vs. changes in ambient would be tough. And then even if you know the effect, not a lot you can do about it anyway. However, panel temperatures will rise noticeably especially in clear sunshine. And most panels are spec’d at 25C (72F), certainly below typical actual panel temperatures.

Post #5

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)

So with a MPPT controller, how does a long extension cord (cable power loss) affect the output power? In my case the 90ft of cable has an R=0.015 calculated, so using this number, I calculated the expected output power loss for a series and parallel panel configuration.
At Imp (4.5A) I calculated the power loss to be 3W (1/4A) for series connected panels and 12 W (1A) for parallel connected panels. However, for this measurement I repeated the calculation using actual measured panel current.

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. R=0.15ohms, I= 4.4A, therefore the voltage drop across the cable would be 4.4A*0.15ohms=0.660V. Therefore the power loss in the cable would be V*I= 0.660*4.4=2.90watts. The expected drop in output current (charging current) would be about 2.90W/13V= 0.2A. Not much, worth the loss if I can get the panels in the sun.

For parallel configuration, same calculations, but I am now estimated to be 8.8A. Therefore voltage drop= 8.8*0.15=1.32V, power loss = 1.32*8.8=11.6W. Or about 11.6/13=0.9 A. Ugg…

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

So now with the panels and cable ready for quick changes, I checked the input and output conditions once more, everything consistent with the initial measurements, and then added 90ft of my extension cord and redid the measurements for series and parallel configurations.

Series connection with 90ft of cable.
Vmpptin=30.5V (TX3DMM)
Impptin=4.5A (DCM330)
MPPT power in= 30.5*4.5=137W
Vbat=13.39 (TX3DMM)
Ibat=10.1A (Trimetric)
Power to the battery =13.39*10.1=135.2

Conclusion:
Measurement uncertainty and repeatability is probably compromising results. Couple that with possible slight variation in solar radiation. Trying to resolve a 3 watt change with a shift in time is suspect. About all I can conclude is that observed results are consistent with the calculated analysis, a few % change in output power and that with the low current and low resistance, the extension cord isn’t making a noticeable (to me at least) change in available power to the batteries.


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

Output power dropped about 12 Watts compared to earlier measurements, and output current by about 1.1A. When I quickly unhooked the cable and configured in series with short cable, output current jumped right back up to 10A. I then quickly when back to 90 ft of cable and current dropped again to 9.1A So, while a change of a few watts is hard to observe, bigger expected changes are consistent with actual measured results.

Conclusion:
1)If you have an estimate of panel output current and cable resistance you can do a reasonable accurate estimate of cable power loss and estimate loss of output current.

2)Even with what one would often consider low current, (8A) and low resistance (0.15ohm) noticeable charge current and charging power loss in watts can occur. Most noticeable with portable panels with long runs. If I can increase charging current by .75A -1.0A by going series vs. parallel, why wouldn’t I??

3)If you’re looking at power losses in a few watts range and even in the 10W range, unless you have very constant and clear skies, and measurement equipment with high resolution, it could be very hard to verify the expected results.

In any event, doing a power loss calculation for your setup certainly makes sense.

Now for those who might say, “Well, a 100ft run to a portable panel is really not a realistic user situation anyway”. What I will say is for ME, it IS. When we dry camp, often it is in the trees, and if I set panels close to the trailer, to get decent sun they need to be moved very often, sometimes every hour or so, and still only gives me sun for 6 hours.
If I can get 50-100 ft away I can get the panels in direct sun from sunup to sundown, and not need to move them, that’s what we do.

And if I can do it with very minimal power loss with a series configuration so much the better.
Kinda like having your cake and eating it to!

Post #4

2)Can one predict accurately output power by measuring MPPT input voltage and panel output current?
3)In late fall at high latitudes how much power can I expect to get from the panel under optimal light?

I configured the panels in series with the shortest connection path to the controller possible and started taking measurements. With the panels in series I was able to measure panel output current with the DMM. In my case, there is not an easy way to measure the output voltage of the panel easily other than at the MPPT controller input. And its panel input power vs. output power we care about anyway. Vout (and battery voltage) measured with the DMM at the controller output terminals. With the configuration I have, the voltage drop between the controller out and battery terminals is only a few millivolts for the currents of interest. So rather than using the trimetric voltmeter with 0.1V resolution, I choose to use the Tek TX3 DMM with 1mv resolution. Battery current measured with the Trimetric with its 0.1A resolution.

So here are results for series connected panels:

Vmppt in= 30.74V (TX3DMM)
Impptin= 4.442A (TX3DMM)
Panel output power= 30.74V*4.442A= 136.5W
Vbat= 13.36 (TX3DMM)
Ibat= 10.3A (Trimetric)
Controller output power = 13.16*10.3= 135.5
Controller efficiency = 135.5/136.5=99% I don’t think SO! I suspect that the trimetric current resolution of 0.1A and trimetric accuracy is the issue. 10.2A for current yields 97% efficiency. So even 0.1A error is significant here.
And the 135W panel output power seems (at least to me) reasonable given the latitude and time of year. 135W is 83% of rated.

I repeated the measurements for parallel connected panels. In theory with a very short low resistance interconnect results should match very closely a series connection.

In the case of parallel panels I could not use the TX3DMM to measure current, but had to resort to the clamp on current probe, an unwelcome but unavoidable additional variable.

And here are the results for parallel connected panels.
Vmpptin= 15.43V (TX3DMM)
Impptin= 8.8A (DCM330)
Panel output power = 15.43*8.4= 135.8W
Vbat= 13.30V (TX3)
Ibat= 10.3 (Trimetric)
Controller output power = 13.30*10.3= 136.99W
Controller efficiency= 137/135.8 >100%

NOTE: I’m nailed by two low resolution current measurement tools, the DCM330 and the trimetric. So efficiency measurement is meaningless. However, output power and current are close enough to “identical” to the short series configuration that I’ll call them equal, which is what one would expect.

I’m not going to “dry lab” the numbers to try to correct the efficiency.

Conclusion: yes, by measuring controller input voltage and current you can get pretty close on predicting output power. Close enough for me anyway.

And looks like in the fall at 46N latitude, I’m getting about 135 Watts from a nominal rated 160W panel, or about 85%. Sounds pretty reasonable to me, but would be interested to know what others are getting under similar conditions.

Post #3

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

I’ve broke this into two conditions. One with the panels not connected to the controller and another with the panels connected to the controller supplying a load to the battery. Basically unloaded and loaded panel conditions.

Condition 1: NO LOAD CONNECTED TO THE PANEL!!!

With no load I measured Voc with the panel vertical and horizontal to see how much Voc varied. With the panel vertical Voc = 20.85V, Horizontal, Voc=20.52, and it peaked at 20.90 at an estimated 60 degrees. Humm….. Might work.

Then with the panel at best orientation by using Voc, I measured Isc to see if I was close to optimum. Isc= 4.94A, and further adjustment showed it was at or very close to optimum position. I couldn’t increase Isc, and much movement caused Isc to drop.

Repeating the alignment, starting by monitoring Isc, got me to essentially the same orientation.

And once connected to the controller, I wasn’t able to get any noticeable improvement in output power. But I only fussed around with it for a few minutes and decided it was “good enough” after all the optimum position is varying all day long.

Conclusion:
Maximizing Isc or Voc seems to get one “good enough” to optimum angle. At least it is “good enough” for me. However, Isc seems to be more sensitive to position than Voc.

Since optimum orientation changes continually during the day, I figured using either of these methods is probably good enough. Once you get max Voc or Max Isc, it’s up to the user to decide how to orient the panel angle to maximize power captured during the entire sunshine period if you choose a fixed angle.

Condition 2: PANELS UNDER LOAD.

This test was done on a different day and different time. 10AM PDT on October 16th, with a few wispy high clouds.

In this case, once the MPPT controller finds it peak power conditions, it will wait some period of time before searching again. I wondered how this might affect trying to find the optimal angle. Also, one the panels are connected to the MPPT controller, it is not easy for me and probably others, to monitor panel or MPPT voltage while adjusting the panel. So that leaves monitoring current.

So, with my panels connected in series via a short run I hooked my clamp on DC current probe onto one output lead and watched current as moved the panels from horizontal through vertical. I did see about 0.75A change in current, and when the current peaked it did so over a fairly narrow angle. I then repeated with the second panel. However, with panels in series what happens is the current is set by the panel with the highest output current, and the second panel will increase/decrease output voltage as it is moved. Good news/bad news. Good news is that the second panel would be expected to be in the same orientation as the first. Common sense. So with that as a starting point, I tried adjusting the second panel, then the first. Careful adjustments gave me about another 0.1A in peak current. But my final orientation was darn close to the optimal position with one panel.

Then I went in and looked at the trimetric. 9.1A. given the wispy clouds and early morning, not bad. I then waited a few minutes to see if the controller was able to get any more power once it did another search. A little later another 0.1A, but don’t know if that was the controller or a little more sun.

Conclusion:

I think it is easier to find optimal angle with unloaded panels, but it certainly is doable with loaded panels. So if in the AM you adjust for optimum, and later in the day want to adjust, using the clamp on meter to optimize current does seem to work and is a lot easier than asking DW to monitor the trimetric as you tilt the panels.!!

But WAIT there appears to be a better method:

One proposed by Jim Denver and expanded by CA Traveler.

Put some form of cylinder on the panel. When the sun is perpendicular to the panel (the optimum position), there should be no shadow. Jim Denver suggested a TP tube, CA Traveler a dart with a suction cup. I’ve got a spare Garmin suction cup mounting system I’m tempted to put a plastic tube on and use that. I tried a TP tube. Very quick, very accurate.

So after all those current/voltage measurements, SIMPLE is better, thanks Jim and CA Traveler!!

Post #2

Background:
During the testing the maximum battery acceptance current was higher than the maximum panel output current. I.E., the batteries were discharged enough that the solar panel can’t output the max current the battery can accept at its state of charge.
Once the battery acceptance current drops below what the panel and controller can put out, most of the configurations I have analyzed should converge and all should give virtually identical output currents to the battery. But I’ve done enough for the last few days, and we are heading to some partial clouds, so as one of my favorite professors used to say. “The proof is left to the reader”.

Measurement resolution:
Due to measurement tools resolution and accuracy, some of the end results may seem and are somewhat suspect for absolute accuracy. MPPT controller efficiency is one example. Please don’t get mired down in the numbers, but look at the end results for what the comparison is saying. E.G MPPT conversion efficiency seems to be very good, rather than It can’t possibly be 98%, it can’t be better than 97%. Sorry, but my measurement tools aren’t good enough to get close than probably 2-4%. All I can say from my measurements is that conversion efficiency appears to be very high.
Unfortunately the resolution of the trimetric and my DC current probe is really not sufficient when one is making comparisons between two large numbers that are expected to give only slightly different values for current, voltage or power. The trimetric has 0.1V and 0.1A resolution and the current probe 0.1A resolution. And while I know the accuracy of my DMM and current probe, I don’t know the accuracy of the trimetric for current and voltage measurements. This really showed up in trying to calculate converter efficiency.
Since data without test conditions, test equipment, and actual configuration are pretty meaningless, I’ll start with outlining the setup my trailer has and the equipment I used to make measurements.

Trailer, solar Panels, and solar panel extension cables:
4 T-125 Golf cart batteries
Morningstar Sunsaver 15A MPPT controller
Morningstar 10A PWM controller
Battery monitored with a Trimetric battery monitor
All removable interconnect between the Solar panel and controller is via Anderson 50A connectors. Very low contact resistance and very repeatable.
Trailer connector for the MPPT controller is a 50A Anderson connector with 5’ of #10AWG wire between the connector and controller
Controller output is 2’ of #10 wire to a junction box and from there 8ft of #3/0 AWG wire to the +and – battery terminals.
Each Solar panel has 5’ of #10 wire to solar MC4 connectors
Panel output is combined (either series or parallel) with a combiner consisting of MC4 connectors on one end and an Anderson 50A connector on the other. Wire is 4’ of #10
Panel extension cords consist of sections of 30A RV cord with the ground soldered to one of the other wires at both ends. This translates into a cord with one conductor #7AWG, the other #10AWG. Cables are 50’, 25’ and 15’ in length. 50A Anderson connectors on each end.
Calculated combined resistance at 90ft is 0.15ohm

Solar panels are two each 80w panels, Voc 21V, Vmp 17.5V, Isc 4.98A, Imp 4.58A
During the test, the trailer batteries were at 80%SOC per the trimetric. The goal was to get the batteries discharged enough to allow the batteries to easily accept whatever current the controller could output.

Solar Conditions:

Completely clear skies with not even a jet contrail!!! Very unusual for western Oregon, but it does happen. And it stayed that way during the several hours of test both days.
Temperature 70F
Latitude 46N
Altitude: 175ft (basically sea level)
Date of Tests: October 14th and 15th, 2013

Test equipment:
Tektronix TX3 DMM NIST calibrated, Resolution: Volts 0.001V, Current 0.001A
Tektronix DCM330 AC/DC True RMS clamp on current meter, resolution 0.1A, Max current 1000A.
Trimetric battery monitor system. Resolution; volts 0.1V, current 0.1A

Measurement uncertainty:
The trimetric voltage resolution is 0.1V, current resolution 0.1A. Accuracy is unknown.
The clamp on current meter also has 0.1A resolution and accuracy at low current levels is not specified. So, the first order of business was to compare the current measured with the clamp on meter to that of the TX3DMM at low currents. I measured Isc of a single panel, and with panels in parallel and with the TX3 and DCM330. In both cases, the DCM330 matched the TX3 to the 0.1A resolution of the current probe. I repeated the measurements several times and go consistent results. So it gave me confidence that using the current probe would be reliable and accurate enough.

The posts that follow present my results and conclusions for each of the questions.

And to repeat: FOR THOSE WHO SEEM TO 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.


So now off to some measurements.