Anyone got some good Sun tracker code or concepts for making a sun tracker?

Ralph

Ralph

I hope you/we find code or help. I am in the process of going total solar for my home and shop. I want to use an ATMEGA to track the sun and move the solar panels.

Jim

I have thought about making one for a long time as well. (Should increase the overall output of my solar panels about 25% from what I understand...)

Technically, the position of the sun for a specific lattitude year-over-year is quite predictable so tracking is not really necessary. (Since my new-found-use-of-Python, I ran across a couple Python Modules that will do sunrise/sunset/position calculations by location...)

That being said, something that tracks the position by light I find is still of interest.

My idea of a Sun Tracker, although never/not yet implemented, uses 5 light sensors. 2 sensors with little domes on either side in order to get a good average for the ambient light. These readings also are intended to be used to make sure there isn't any motion due to cloud movement/shadows.

The other 3 sensors are grouped together in a plastic unit. Kind of a fan shaped thing with 3 shafts. The outside shafts are wider, capturing more light, the middle shaft is more narrow. The plastic would be sprayed, pretty much all over, with Krylon/Rustoleum silver paint to reflect light. There would be a couple coats of high-gloss to cover the silver coating to protect it a little from oxidation.

Sensor Array

    ___________________________
   |          |     |          |
   |          |     |          |
    |         |     |         |
     |        |     |        | 
      |       |     |       |
       |   _  |  _  |  _   |
        | / \ | / \ | / \ |
        | | | | | | | | | |         
        ___________________

           1     2     3

Only an "ASCII-ART" drawing, but maybe the point gets across?

The microcontroller will use the "average ambient light from the one sensor pair" as the "goal" for sensor-2 of the 3 sensor array.

The basic algorithm would be:

1 - if sensor-2 = goal no change
(It may need to be some "percentage of the goal" either fractional or multiple...
That is where some trial and error would come into play during testing...) 2 - if sensor-1 > sensor-2 track left.
3 - if sensor-3 > sensor-2 track right.

All light measurements are done off of averages over 1 minutes.

That was what I was thinking about. Keep in mind, nothing was done actually done except some thought.

We almost went solar this fall but the size of the array killed the idea.

I have seen claims from tracking array manufacturers claim 50% improvement.

Any improvement I would think would pay but I went to a local University's solar seminar and the leading professor insisted that tracking mechanisms were just not worth the maintenance troubles.

I actually have three different solar projects running through my mind.

A passive computer controlled system would alleviate clouds but if it was a cloudy day a couple of my ideas would not work anyway so I like the idea of active tracking.

Ralph

Apparently, solar photo photovoltaic cells don't need to be facing the sun, as they can capture photons from scattered (non perpendicular) light. Only solar concentrators need to track the sun.

Here is a lecture from Stanford University which explains.... the point is made at time 1 minute 45 seconds,,,,

http://class.stanford.edu/solar/Fall2012/videos/u1airmass1

Apparently, solar photo photovoltaic cells don't need to be facing the sun, as they can capture photons from scattered (non perpendicular) light. Only solar concentrators need to track the sun.

Here is a lecture from Stanford University which explains.... the point is made at time 1 minute 45 seconds,,,,

http://class.stanford.edu/solar/Fall2012/videos/u1airmass1

What I presume is a link to a video does not play. A note at the top of the web page says that "you must be logged-in to view the content you chose"...

So I cannot comment directly on the content of that video.

While it is true that solar photo photovoltaic cells don't need to be facing the sun" in order to produce power to produce "optimum power" orientation is quite critical. The orientation of the panel can have a HUGE AFFECT on overall power output. (Try facing a panel north and see how much power you get then...)

All you have to do is signup to create an account to view the videos.

New world record for flexible thin film solar cell efficiency of 20,4 percent The article claims that these solar cells have the potential to enable low-cost solar electricity in the near future.

Link to related article: http://www.sciencedaily.com/releases/2013/01/130118064733.htm

More useful solar data:

http://pveducation.org/pvcdrom

Insolation with all angles.

http://pveducation.org/pvcdrom/properties-of-sunlight/making-use-of-TMY

Cloud cover data adjusted insolation Flat plate optimum tilt insolation 2 axis tracking insolation. European insolation

http://pveducation.org/pvcdrom/properties-of-sunlight/cloud-cover-data

Superb insolation by lattitude, date, tilt, cloud, ...

http://pvcdrom.pveducation.org/SUNLIGHT/SHCALC.HTM http://pvcdrom.pveducation.org/SUNLIGHT/SHCALC.HTM

Calculator

http://pveducation.org/pvcdrom/properties-of-sunlight/calculation-fo-solar-insolation

USA Insolation data 1991 - 2005 Hour by hour !!! Many sites Immense amount of data.

http://rredc.nrel.gov/solar/old_data/nsrdb/1991-2005/tmy3/by_state_and_city.html

Local solar time correction factor.

http://pvcdrom.pveducation.org/SUNLIGHT/SUNPOS.HTM

European Solar Test Installation - ESTI

http://re.jrc.ec.europa.eu/esti/index_en.htm

European PV potential

http://sunbird.jrc.it/pvgis/cmaps/eu_cmsaf_opt/PVGIS_EU_201204_presentation.png

http://sunbird.jrc.it/pvgis/cmaps/eu_cmsaf_opt/PVGIS-EuropeSolarPotential.pdf

UK :-(

http://sunbird.jrc.it/pvgis/cmaps/eu_cmsaf_opt/G_opt_UK.png

ISOFlux plots, world, by quarter

http://pveducation.org/pvcdrom/properties-of-sunlight/isoflux-contour-plots

Thanks everyone, I saw a headline for a study that claimed maximizing the gross output of a fixed angle solar array by heading it to the west instead of south. I cannot find the article now.

But speaking of future efficiencies check this article out.

Now that's what I need, we had looked into going solar this fall but the size of the array (65' x 12') killed the idea.

tpdcoyne, thanks good references.

Ralph

I've been combing through the Cornell ECE 4760 "Design with Microcontrollers" course SITE...and came across a final project built by some of the students HERE. It may be more involved than what you are looking for, but it might provide some ideas.

Thanks pcbolt, yeah that is a great project. It covers a lot of what I want to do.

Some of their reasoning is strange. Why would you need a active GPS module, well I suppose if you wanted to move the array you'd have the GPS built in but I would just take a GPS reading with a external device and plug that in. And then since you have a GPS why do a RTC you can get a time stamp from the GPS, plus they think the RTC running off the micro will last a long time whereas the Nerdkit RTC is only good for a couple of hundred days.

Oh well it is a great project and there is definitely a lot that I can use, thank you.

Ralph

Heya Ralph!

I have been working on various optical sensors lately, and ran across something that might interest you. No good code or anything, but a thought on sensors. My apologies if this is old news or a rehash of earlier posts, I haven't been keeping up around here as well as I want to lately.

As stupid as it sounds, the best light sensors I have come across recently is just a group of plain old LEDs(I used yellow-green). LED's can both be emitters and detectors just depending on how you use them, but the cool part is that they are sensitive to the light they emit(and ignore other wavelengths for the most part). You put each one or more LED's in parallel set up "backwards" with a big resistor(say 1+MOhm) and connect your voltmeter or ADC inputs across the resistor. Each LED will give a few 10-20's of mV's but with only a few <<yA's of current, so the resistors need to be big.

When testing 4 standard yellow-green LED's in parallel with a 10MOhm resistor I was seeing ~335-500mV in "good" sunlight fading fairly nicely to about 125mV in partly cloudy and then 1-2mV in very cloud covered sky or dusk. It would be best to have your ADC set up to measure a differential voltage, since it gets kind of messy sending several clusters to a common ground. (I put an extra 1Mohm resistor between each "cluster ground" and the common ground to keep the clusters from overriding each other). So if you were to you have a few of these "clusters" pointing in different directions/angles facing South, they might work well at tracking the sun/light source. Depending on which light band your solar panels pull the most power, you can use that LED for your light sensors. I have tried putting a couple different LED colors together in one cluster and have gotten mixed results, but honestly haven't played around enough with the mixed clusters to have a good handle on it yet. There are a decent amount of articles on the web about this, and evidently this is pretty old use for LED's. A few articles I saw mentioned doing this in the 70's. Anyhow, just a thought that might help. After all, if $5 of LED's and $2 of resistors can work better than $20-$50 solar sensors, why not use them?? hope that helps some...

Hey 6ofhalfdozen, I have missed you. Thanks, the college paper pcbolt link to used LEDs I was wondering how that worked. I will play with some to see what I get. I wonder what would happen if you surrounded your perimeter with LEDs you probable would need a external ADC but I wonder what you would get?

Ralph