solar panels as sensor for tracker

[commanda]

Lets say I was building a 24 volt system. I have 2 x 12 volt solar panels (identical), wired in series, and mounted one East, one West, of the central axle of a solar tracker.

Could I use the mid-point voltage, fed into a window comparator, to drive my tracker motor?

Heightening the differential between the two cells, would be an inverted V reflector mounted between the 2 panels, similar to that done by Poulek.

http://www.solar-trackers.com/

Considered and expert opinion appreciated.

Amanda

[Flux]

Sounds perfectly sensible to me, not that I have ever been in a position to use a tracker so my opinion may not count for much.

Flux

[Opera House]

Pointing the panels in slightly different directions works, but defeats the purpose by lowering the output.  A leaf or bird dropping will send it pointing in the wrong direction.  I like their simple vertical panel idea.  A small micro could periodically search back and forth and find the best direction.  Any hunting looses a lot of power.  You are probably better off with a clock like drive that changes position no more than 5 or 6 times a day.  No one ever thinks about how much energy these systems use.

[wpowokal]

Amanda I prefer an indipendant circuit, I have 'e' mailed it too you, it returns the panels to the east at sunset.

If I understand you correctly you propose using each panel as a LDR (essentially), for my money a seperate circuit achieves the same result with less fuss.

Just my opnion.

allan on the other side

[WindChopper]

"Could I use the mid-point voltage, fed into a window comparator, to drive my tracker motor? "

Yes, No, Maybe, plus all of the above comments!

My Picaxe 40x1 Tracker/Logger is a work in progress and has been interesting for blasting some logical assumptions.

For reference - 90 watts of panels on a tracking mount on a 12vdc system.   Definitely in the toy class.

For the "sensors" to work properly, you need to sun block the East or West Sensor to get the differential voltage.   I tried a pair of the typical 2-3 v Garden Light cells with so-so results.   The best so far has been CdS resistive elements, as I can get a good differential reading at adequate level for the ADCs in the Picaxe.   Also being small ( .25" / 6.3mm ) the block and the package is small.

In addition, the panel voltage rises rapidly in anything that resembles light and just flattops against Battery Voltage, but produces no real charge current until you get good sun.   You would need to really block the non-favored panel to get any differential and that is self defeating.

You may want to look at the Redrock Green LED sensor units as an alternative or just go with a small CdS package and either a discrete component design or a PIC/Picaxe/whatever design.

I find that Gray-Days ( lots of those lately - Northern Illinois - US ) are the biggest problem and they are the days that you are trying to "Milk" the array for anything that it can produce.   As I said I changed to CdS differential detect, but the current re-write is going to look at the "Sun Density" and switch to time based if the levels are not good or just go flat South if it is really bad.

I reset my panes East at Midnight, but think that I am going to change that to due South and then go East in the AM if the effort is justified by the sensor readings.

Re: Power consumption - the reason that I am looking at flat due South on bad days is to avoid a net loss due to drive current consumption ( Linear Sat Dish Drive - 32vdc run at 12vdc - takes about 300ma ).   That's not much, but on some days I don't make enough charge current to justify the effort.

Over complicated - Yes - but its a test system - so now is the time to make the mistakes.

Not sure this helps.

Russ

[Opera House]

Instead of using voltage, sample the current from each panel.  Then make a decision to move based upon there being enough current.  I definitely think a micro based on time, position, and length of day beats any tracking system for least current used.

[wooferhound]

Wow I really like that Idea

a tracker that runs on a timer of some kind

[ghurd]

I know of a matched pair of PVs. A does better than B in the cool, vice versa when warmer.  Not sure if that is common.

There was a company making trackers with small PVs (yard light size, 5-cells?) as sensors and to power the logic circuits, set in an inverted V.  They didn't power the drive motor. Might have been a German brand.

It would keep the voltage below smoke point.

G-

[wooferhound]

Yeah current would probably be best to use as the sense mechanism

the voltage would be held down to the battery voltage

[TomW]

Woof;

Thats pretty much what I do. On a bit more elaborate type "timer".

I use the Linksys "Slug" and a USB driver module [thanks, Ibedonc] that runs via "crontab" on the Linux install on it.

This mini computer draws very little power, interfaces to my Outback Mate and an RS232 digital meter and gathers data from my system. It has plenty of horsepower to do lots more stuff like controlling a dump load and crunching numbers for real time data on the system.

Cheers.

TomW

[dr2pks]

Where does one find such a timer?  I mounted my 4 panels (24A) on an old C-band sat. mount and want to utilize the original linear motor to move them.  I've been moving them manually about 3x per day which seems adequate w/o getting too obsessive.  When I'm not around they're locked at about the 1pm position. An adjustable timer which would operate the linear motor for x minutes y times per day would be awesome.

The sat. mount with some minor modifications makes a free, perfect, windproof, bulletproof basis for panels.  I have owned Zomeworks Trackers, etc.  They're beautiful things to own and watch - when you don't need to spend your own cash on them.  I (and many other old pv heads) have come to believe in simple rigid mounts and buying more panels as a better alternative to expensive commercial trackers.  The free sat mounts make sense even if they're rarely or never swiveled.

The juice I'll spend on the linear motor may prove to be more than it's worth.  Going ouside to move the panels at 10:30 and 2 is the only exercise there is somedays.  

[commanda]

If I have 2 panels in series, the current through the 2 panels will be same.

And I forgot to point out that I have built and tested a tracker using green leds, similar to redrok's.

http://www.fieldlines.com/story/2007/10/23/3831/9944

This uses a timing circuit so it only samples briefly every so many minutes.

If someone has a 24 volt array on a tracker that could run a test for me, would be great.

2 resistors, say 100K each, in series across the 24 volt output. Digital meter from the junction of the 2 resistors to the junction of the 2 panels. Pointed at the sun should be zero volts. Deliberately mis-align it a bit, and see how much the voltage changes.

Might have to get a couple of small panels and build a model prototype.

Amanda

[ghurd]

Would you like the panels loaded (battery) or unloaded?

Mono or cryst?

G-

[commanda]

Loaded. Type of panel shouldn't matter.

If you're volunteering for the job, it's much appreciated.

Amanda

[ghurd]

Now it started snowing.  Calling for snow for 65 more hours, the some sun.

I can probably test about 4 configurations if all goes well.

The largest is only 140W rated. Next is 100W, then 10W and under.

Appreciated? It will never make up for all your help.

G-

[Tritium]

I have often wondered why no one uses a simple clock drive from a telescope for tracking. Granted it will not automatically stop at sundown or return to sunup position without modification but some of these telescope clock drives are pretty inexpensive.

Thurmond

[Volvo farmer]

You have everything you need to use a Redrok tracker. Seems like the simplest $35 solution to me. It will park in the East after dark and you'll never have to fiddle with your panels manually again.

http://www.redrok.com/led3xassm.htm#led3xforsale

[ghurd]

Hi Amanda,

Very strange results.

First I thought it had no hope. One panel was always different from the other.

Then, when tipped, the numbers went crazy.  I think thats good?  I now figure that's what you expected.

=


The best matched panels.  3W mono.

Lower light, the panels seemed closer.

As the light was better, the difference was greater.


Panels on the same plane, at the same time.

PV A, 12.65V.  PV B, 12.58V.  (lower light, A is higher)

A, 12.62V.  B, 12.75V.

A, 12.50V.  B, 12.95V.

A, 12.45V.  B, 13.00V.  (better light, B is higher)


Panels, with one tipped a hair, was a hard thing to do as numbers changed so fast.

Tipping one 15 degrees or so, would make the voltages very different.

The tipped away from the sun panel would drop voltage. The other voltage raised.  A LOT.

Best guess, 17V and 8.3V.

Even a 5 degree difference was drastic to the voltages, maybe 10V and 15V.

Then the window of reasonable light was gone.


=

The others were either not very matched, or cheap 5W thin-film (NOS- maybe Cronar / Chronar?)

The tipping had about the same results, however the voltage swings related to light level were worse.

ie:  PV A would be higher than PV B in good sun.  B higher than A with a slight cloud.

On the same plane, the thin-film had the worst swing of over 1V per panel, followed by 50W multi with far off serial numbers showing about 0.75V.

==

Did it with 2 meters at first.  Then added a SPDT and used 1 meter, just in case.

The resistors were 100K.

There was no controller, but a 10A 45V Schottky blocking diode.

The closer matched monos on the 24V system had ice on one.  They are not movable.

If that's not what you were looking for, let me know.

G-

[ghurd]

The more I think about this, the more I like it!

I believe the LED sensors sample a small piece of sky.  Here in the cloudy snow belt, that will get them hunting.

Using the whole panel should get less hunting, and better overall aiming.

If a pair of PVs were set 5 degrees different, higher inbetween them, then the output wouldn't be any worse than being off about 10 minutes from perfectly aimed.

The drastic voltage difference between the better and worse aimed PVs should make up for variations in individual panels.

Is that what you are thinking?

I don't think it needs the inverted V reflector.

Would be nice to have a sample every few minutes, like the other one.

Would be nice to have a 10 degree max rotation per cycle.  Not sure how it would reset itself every morning.

I didn't think to put a piece of tape (bird do-do) on one to see how that changed things.

I'll try it if the sun shows up.

I don't think that will be such a drastic change, if the panels themselves were a tiny inverted V.

Last stray thought.

If the panels were V, the light reflected off the misaligned lower V PV should increase the more properly aligned PV?

Sorry.

I see I did not do it like you asked, but the results are what you were looking for?

Each PV had a resistor in parallel, all 4 tied together at the center.

The voltage readings were across the resistor for each PV.

Your way would have been easier.

I can redo it if it helps.

No idea what happened with the text size up above.

G-

[commanda]

Thanks Glen,

It sounds like it will work, provided we have identical panels, and a fairly large (and adjustable) window for the window comparator. In my envisioned application, the inverted V concentrator between the 2 panels should hopefully help. The 2 digital outputs from the window comparator would feed this circuit;

http://www.fieldlines.com/story/2007/10/23/3831/9944

which only powers the motor briefly every few minutes.

Next step is to source a pair of small panels and build a model. Will also help me iron out the tilt adjustable traxle mechanics.

I think differences between the panels (low light to bright light as you found) will only result in the panels being very slightly off perfect aiming. Going by the large difference when deliberately mis-aligned.

But re-reading what you wrote several times, I suspect you didn't quite grasp exactly what I meant.

Imagine a bridge. The left arm is comprised of 2 panels in series. The right arm is comprised of 2 resistors of equal value in series. The meter is connected from the mid-point of the 2 panels, to the mid-point of the 2 resistors. When aimed perfectly at the sun, the voltage on the meter should be zero. Tilt the panels one way the voltage will become positive, tilt the other way the voltage will become negative.

But I put your numbers into a spreadsheet to derive the format I wanted.

cell A    cell B    Vbatt    Vmiddle    differntial

12. 65    12.58    25.23    12.62    0.035
    
12. 62    12.75    25.37    12.69    -0.065
    
12. 50    12.95    25.45    12.73    -0.225
    
12. 45    13.00    25.45    12.73    -0.275
    
    

Considering the differential when deliberately mis-aligned, I think it will work.

Amanda

[edy252]

hi....

i haven't seen the board for quite a long time...

i made a solar tracker once...here it is :

http://www.fieldlines.com/story/2007/6/17/65139/9125

hope this helps... if u cant figure out how it works i would be happy explaining it further more....in case i dont check out the board again for a long time, contact me on my email : edy_252 at hotmail.com