Tracking the Sun
william chan
Posts: 1,326
Hi,
Anybody got any experience setting up small mirrors (1 sq foot each) that track the sun to heat up a small boiler or oven?
What's the best actuator to use. Servo motors?
Which is the best sensor to use to sense the exact location of the sun, track and reflect the heat to the oven.
Is the SX suitable for this job?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my
Anybody got any experience setting up small mirrors (1 sq foot each) that track the sun to heat up a small boiler or oven?
What's the best actuator to use. Servo motors?
Which is the best sensor to use to sense the exact location of the sun, track and reflect the heat to the oven.
Is the SX suitable for this job?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my
Comments
Check out this thread for some suggestions.
http://forums.parallax.com/showthread.php?p=584537
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe
IC Layout Engineer
Parallax, Inc.
Thanks for the reply.
Tracking the sun for a solar panel is the easier part, because it allows for inaccuracies.
Tracking the sun and directing the energy to an oven seems to be more difficult,
especially if you have many individual mirrors.
There needs to be much higher accuracy, or the sunlight will miss the oven.
Is there an easy, accurate and reliable way for directing sunlight to an oven?
Can 2 parallax servos be used for a 2-axis controller to position the mirror?
Is there enough torque and accuracy?
Thanks.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my
Parallax servos are nothing more than Futaba servos "in disguise". I believe it's the S-148, but I stand to be corrected. The real question then is, how heavy is the mirror?
Regards,
Bruce Bates
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
regarding tracking the sun: given a specific location and date & time, the
position of the sun should be known. So it is rather a matter of an ongoing calculation
i.s.o. find and track.
Ed
here a link to a document about the calculations needed. Looks "do-able" BUT what about the floating point calculations?
The BS2 series hardware is not capable of handling this with enough accuracy. You might need a floating point co-processor.
Anybody any thoughts ?
Thx, Ed
http://www.stargazing.net/kepler/sun.html
Thank you for the reply. The calculations looks quite complex.
Maybe we will need a Propeller instead of SX.
Now how to add in the calculations to direct the light to an oven?
I think we may need some photoresistors to fine tune the positioning, just to make sure the
reflected light does not miss the oven.
I was thinking of a module with a 1'x1' mirror, each powered by 2 servos controlled by one SX processor.
Then I can add as many modules as needed and wire them all using RS485 networking.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my
agreed that you need a sensor and a feedback mechanism for the final adjustment because, although the calculations
are more than accurate enough for this project, the mechanics that control the mirrors will never be. So measure
and adjust is a must.
The calculation itself is what worries me most now. As far as I understand the capabilities of the microcontroller
offered by Parallax AND others, they do not handle floating point calculations. And again that seems to be a must
for thsi project.
Adding a math co-processor seems to be the ultimate solution.
Ed
Gee. Why track the sun? Don't we know--or can easily calculate its position--with something like a PC for the next jillion years or until it becomes a red giant, or something?
Seriously, given your known position, why couldn't you use a PC to calculate the sun's position, say every 5 minutes . . . or maybe even every minute and store the data on something like the memory card that is compatible with the SD Data Logger and have THAT device simply tell your Stamps where in the world (well, where in the sky) the sun is and point your mirror(s?) that way?
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
see previous messages in this thread. Yes, the position can be calculated and some "horse-power" is needed to do the calculations.
Good idea to use a PC and put the results on a memory card !!!!
But next step is to have a position system + feedback because I don't believe it is affordable to build a mechanical set-up that can be
positioned with enough accuracy. Measure and feedback is a very good alternative.
Ed
Seems like we need photoresistors near the oven itself to check whether the light from the mirror are getiing to the oven.
But if the oven is so hot, I think the photoresistors will melt.
If we keep moving other mirrors away to get feedback, by the time we move them back, the sun may have moved on a little bit.
Seems like its a difficult problem.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my
Technically all you need is to set the position of only one mirror as a reference. Based on where you want your focal point, and following Pythagoreans theorem in at least 2 dimensions, the position of the other mirrors could be calculated.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe
IC Layout Engineer
Parallax, Inc.
Great idea and drawing you have there.
But shouldn't the oven be shaped like a cylinder that has a flat face facing upwards so that mirror modules can be placed all
around it?
When the photoresistor array is swung down, will the sensor touch the oven and get burnt?
Maybe we should use the SX instead of BS2 since we can use the built in comparator to implement a multidrop RS485 networking.
I have another idea, instead of using complex calculations, why don't we use a joystick to manually move mirror to reflect light to the oven and then store the position, date and time?
I also think the position data should be stored in the PC, not on the mirror module itself.
Another problem could be the accuracy and repeatability of the servo motors.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my
If you want any real accuracy for this project, you'll need steppers. Preferably with worm-drives or other gearing.
Also, if you add a RTC-chip(some have serial IO and are easy to use from a BS2), and have the BS2 periodically query it while waiting for the next time to move the mirrors...
Third, this project CAN be done with a single BS2 as the mirrors doesn't have to move simultaneously. (assuming that each mirror is only moved every minute or so, there's loads of time to move them all. )
Just use a couple of IO-pins for 'motor-select' and you won't even run out of pins...
Will the mirrors be flat or curved?
I suggest calculating the positions based on the mirrors and oven all having known positions, and storing thosse values in tables. Then set the kit up using a GOOD compass with declination-adjustment, good measuring tape and levels.
(Note, declination can change with more than a .1 degree/year, so check a good, NEW, local survey map for information)
If your maths is OK, most of the errors will be from inaccuracies in positioning, and can be adjusted by observation. The rest should be easy enough to correct in the tables.
Also, the mirrors should be mounted on an axle that is parallell to the Earth's axis, to make the movements easier to calculate.(Take a look at the mount used for motor-powered satellite dishes)
Then you'd only need to adjust the vertical position once every day.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...
Here is a web page with all sorts of interesting information on solar trackers and heliostats:
http://www.redrok.com/main.htm
Regards,
Bruce Bates
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
when you say mirrors? are you planning on using parabolic mirrors? I see parabolic mirrors being much more efficient and easier to "aim"
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
A complex design is the sign of an inferior designer. - Jamie Hyneman, Myth Buster
1. Make a prism shaped object with two equal sides (the sensor sides). The third side is the base and is attached to the same plane as the mirror.
2. Attach a photo sensor to each equal side about half-way between the point where the two equal sides meet and the base.
3. Buy a 4-in-1 op amp chip and wire an op amp to each sensor and configure the two to work as a voltage comparator. The other two op amps are wired as voltages references.
4. Illuminate the device uniformly (straight on) and set the op amps for a null output.
5. Interface the op amp output to a STAMP. The STAMP will activate a turning device whenever the op amps are not equal and based upon the output when they are not equal, will know what direction to turn the turning device.
You can make the prism device by making a simple form and filling with automotive Bondo.
If you need to track in two axis, make the object in the form of a pyramid, use a sensor on each face, connect pairs of opposing sensors to the op amp circuit above.
See attachment.
1.·Mount·a mirror so that a pivot axis is parallel to·Earth's axis.· The sun 'travels' across the sky at·1 rev per day = 360 degrees per 24 hours.· Because of the geometry of reflected light (angle of incidence = angle of reflection etc), the mirror has to pivot half that speed, or 180 degrees per 24 hours.· Synchronize the mirror with the sun, then move the mirror at that·rate.· (Because the mirror isn't at the center of the earth, there will be some error but maybe it is small enough).· To account for seasonal variation in the height of the sun, just go out and manually nudge the mirrors up or down every few days.
2. Build a sun tracking device·to measure the angle of the sun.· The sun position is sent to the mirror positioning device (a multidrop network).· The mirror can then calculate the correct mirror position.· The tracking device is kind of like a sundial, or inverted 'T',·with a light sensor mounted on either side of the 'blade'.· If the blade is pointing right at the sun, each sensor·receives the same·illumination.· If the blade points to the side, then one sensor will be in the blade's shadow.· So the difference between the sensor signals is an error signal indicating which direction to tilt the sundial.· Encoders on the sensor positioning mechanism indicate the sun's position.
3.· The mirror will reflect a "beam" of sunlight toward the oven.· Position sensors between the mirror and oven·near the edge of the beam.· When the sun moves the reflected beam will change direction slightly causing one sensor to get more light than the sensor on the opposite side of the beam.· Use the difference as an error signal.
#2 is essentially the method I used here...
http://forums.parallax.com/showthread.php?p=584537
...but instead of an inverted T, you have a rim or hollow tube. The tube needs to point directly at the sun, otherwise a shadow from the tube will fall on any of the 4 sensors located at the back of the tube and cause an imbalance.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe
IC Layout Engineer
Parallax, Inc.
Baking bread is a good idea.
Another idea could be to try to generate some electricity from the heat.
I don't know where to buy parabolic mirrors, but I can buy flat mirror tiles from IKEA.
If we plan to make a lot or mirror modules, we have to use flat mirrors because some mirrors are near to the oven and some mirrors are far from the oven.
How to design an oven that absorbs the most heat and keeps it in?
Gadgetman,
I think you are right about using stepper motors. The servo motor's repeatability had me worried for a while.
We can't use one BS2 only b'cos we want a scalable system that can potentially scale up to a hundred mirrors.
Hmm, maybe we need RF networking as well.
BubbleHead.
Can you draw a picture of the sensors all around the cylindrical oven that faces upwards?
Seems like we need a lot of sensors.
I am also afraid that light from other mirrors might also fall on sensors meant to calibrate other mirrors.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my
But you'll only be able to bend it one way, so that it makes a narror vertical or horizontal beam on the target.
This may or may not be useful, depending on the target-area of the oven.
As for how many mirrors you'll be able to control with a BS2, who knows...
One thing you will have to do is to protect the oven against wind.
Placing it inside a half-sphere may be enough.
(You may also have to place a piece of glass in front of the opening)
As for the design of the oven itself, that depends entirely on what you'll use it for.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...
Mirror(s):
You can cut "strips" of the mirror tiles (1" should do it·using a glass cutter and a straight-edge) and arrange them in a curved frame to concentrate the beam on your oven. The radius of the frame would determine the focal length of the beam (distance from mirrors to oven). Also, if the mirror glass is not tempered or too thick, you can build the frame to allow you to bend the strips to make a vertically concave shape and concentrate the beam even more.
Positioning:
A simple box with a vertical slit·opening positioned at the base of the mirror,·three phototransistors placed horizontally inside the box opposite the slit·and a simple comparator circuit would do the trick. Facing south as the sun moves from left to right, the narrow band of sunlight entering the box through the slit would move right to left across the photoresistor array and indicate the direction that you need to slew the mirror to keep it pointed at the sun. A second box, rotated 90 degrees, could be used to control the vertical positioning. Just aim the mirror(s), position the box(es) so that the center photoresistor is "lit" and you're calabrated.
Maximum heat:
Google DIY solar collectors and look for references to special 90+ % black paint. This type of paint is very high-temp and can absorb up to 98% of the light energy hitting it. This stuff is really cool. It can actually cause wood to catch fire in strong sunlight.
This thing could make a heck of a hotdog cooker! Good luck,
Steve
cutting strips of the mirrors sounds like a good idea IF the strips aren't too narrow, and the frame can be built accurately enough. If the frame is inaccurate you may end up wasting more of the light than if using a flat mirror.
That paint sounds interesting...
It also sounds like the stuff that was on the roof of my first car.
Now THAT was a solar oven in the summer...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...
My dad has looked into doing something similar to what you are doing, He found a flexible plastic mirror product that came in 4x8 sheets which you can cut into the shape you want. I'll see if I can't find out a product name and post it.
http://quicksilverscreen.com/watch/1182
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
A complex design is the sign of an inferior designer. - Jamie Hyneman, Myth Buster
---this project can easily be accomplished with a couple of SX's, or with whatever little motor-turning computer you are comfortable with...
--- most of your work will be MECHANICAL.· Buy a 3-foot plastic picnic table - mount nine foot-square mirrors, and then get it so it rotates left/right and up/down·with any motors that work for you...··· this is fun stuff - you get to build things that move.·· Make one mirror-unit to start - add more later.
--- construct a target - bricks are good, like a free standing barbecue, but with the sunlight entrance from the side...
--- make a motorized gizmo to follow the sun and to measure an "angle".· It won't be a real angle - but maybe just one rotation of your motor's shaft.·· Get the SX's to talk a few numbers with each other... connect them with (gasp !) wires...
--- the math is simple (though on two axis) - a picture is worth a thousand words:· three sets of mirrors are shown in their "home positions"...
·
--- oh, I wish I had a big back yard...
tommy
·--- ps - that redrok site mentioned earlier·is priceless ! ! ---> http://www.redrok.com/
Post Edited (tommy) : 6/22/2007 5:09:37 PM GMT
I was assuming that this would be a "simple" project with the oven (collector) and mirror(s) as a single dual-axis arrangement. If the project is being designed to aim multiple mirrors at a fixed but seperate target, it gets A LOT more complicated. It begins to look like a "solar farm" I saw once out in Californis years ago. 150+ mirrors all focused on a collector at the top of a 200' tower. When the beams were focused away from the collector, during maintenance on the tower, the beams would ionize the air where they intersected (and vaporize birds that flew through it). When focused on the collector, it was dangerous to look at.
I wasn't able to find a supplier for the 98% black paint, but several sites that I found (DIY solar panel) talked about using a flat black paint that's used to paint automotive exhaust systems that could be purchased in many auto parts stores. It must work pretty well, because they cautioned against using it on wood or light guage metal collectors.
I look forward to seeing what becomes of this project.
Steve
Your ideas are brilliant !
If you don't mind can you also draw a picture of the picnic table to show how to mount 2 worm drives?
Thanks a million.
William Chan
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my
Although many solar arrays are controlled by simple optical sensors, they do have a drawback. Just watch one try to track while clouds are drifting by!
I worked many years ago at Sandia Labs in Albuquerque, and they have done much research in this area. I suggest you google Heliostats and Sandia Labs.
I designed a system that controlled a large array of mirror panels, and obtained an accuracy of better than 1 degree. Ifound the error was mostly in the mechanics of each individual unit.
I admit that the system was complicated, but the concept can be scaled down.
It is like trying to march a bunch of soldiers down a parade field, so I first designed the "Sergeant".
The Sergeant was a PC/104 platform, with a cheap·OEM GPS receiver and a cheap 900Mhz RF module. The software was based on a Basic program I found on the internet that accurately· calculates the position of the sun from anywhere on the globe, but you do need floating point! (Many Solar websites have references to code written to calculate the suns position.)
I then designed the "Soldiers"........a small PIC with a receive only RF module (Soldiers do not talk to the Sergeant while marching) and it controlled the stepper motors to set the proper azimuth and elevation for each panel.
Because of the GPS, the system was capable of being moved anywhere.....but it does require some mechanical calibration of the mounts on the initial installation.
Also, the recommended mirror geometry is an elongated, parabolic trough!
If you have any questions....please feel free to ask.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I would never advocate drugs, alcohol, violence or insanity: But they have worked well for me.....Hunter Thompson
-- I wasn't advocating a 'tower' in Mr. Chan's back yard...·· This is a hobby - (yes, certainly - it can be a serious hobby) at which we get to build things...·· from programming SX's in order to get a stepper-motor to work, to building sun-sensors, and targets - then figuring out how to use the heat (or the light) to do something practical...· the most important word is "learning".
-- sometimes the hardest things are the most mudane -- how to make a bracket to hold a wheel, or how to connect a 1/4" motor shaft to a·3/16" one that will fit inside a certain worm gear.· This is the most fascinating, informative hobby - ever:· building stuff you think of ! - things that "think" and move.
--·I was mostly criticizing the many calls for complicated math.· Below is a left-right mirror pointer using no math at all ! --- as the single motor is caused to point the sensor at the sun, it also causes the mirror to point correctly to aim the sunlight straight ahead... (up)· -- this is a "top view".
·
-- and I suggested nine mirrors on a 3' diameter plastic table top.· The table-top is weather proof, and will last for a couple of season's worth of experimentation...· I don't think (I don't know) that nine suns will·ignite a crumpled newspaper in open air -- but single-sun hot water systems are just so common nowadays, that nine suns opens up other paths.· Nine suns will tan your skin and hurt your eyes - but like owning a swimming pool, responsibility comes with this activity - children and alcohol come to mind immediately,·pets and fire hazards, too.
-- common sense, horse sense and nonsense - all play a part.· But it's·so much fun to improve upon one's project as one goes along - and to then see it·finally, "work by itself".
·--- ps - that redrok site mentioned earlier·is priceless ! ! ---> http://www.redrok.com/
Post Edited (tommy) : 6/25/2007 4:13:45 AM GMT
-- and once it does - aim the mirror correctly, then rotate·it by time !
-- in the drawing above - (previous post) - if the motor is a stepper motor with 48 pulses required for each revolution, AND the worm drive's gear ratio is 50:1 THEN a full revolution of the sun-sensor (24 hours) would take 2,400 pulses to fulfill.
-- once aimed, 100 pulses per hour (one every 36 seconds) would keep it on track.
-- left/right only...
-- this example's gearing is too high - (and impractical) - redesign for 1000+ pulses per hour by making the lower pulleys ten times smaller, or using a stepper-motor with a gearbox, or both...
-- but you get the idea ? -- the Earth's rotation is constant speed and predictable.
·--- ps - that redrok site mentioned earlier·is priceless ! ! ---> http://www.redrok.com/
Post Edited (tommy) : 6/25/2007 7:31:45 AM GMT
--·your first try will be for function:· get it to work, damn the appearance.
-- your second try will be to get it to work, correctly.· Tinker, adjust.
-- your third try you can concentrate on looks, and things like getting the electronics to fit into a box, and making things waterproof, and kiddie-safe, etc...
·Here are some gears:
·
-- stepper motor / white, worm gear / large, (15") wooden, almost-a-pulley.
++++ motors: http://www.allelectronics.com/cgi-bin/category/400/Motors.html
++++ worm gears: https://sdp-si.com/eStore/·(no motors here - verry expensive)
-- I have only a small courtyard at the rear of my apartment.· (I also live on Long Island, which has unacceptable sunshine most of the year).· If I had a yard -·AND lived in the Rocky Mountains (Woodland Park, Colorado ?) - I would start with:
-- four used (used up), squarish, racing tires, as a base to build upon.· they can be bolted to, they're wind-proof, waterproof, free, and are easy to drag around the yard to position.· (three will do).
-- fasten a four by four piece of plywood on top for a base.
-- get a 3' 6" x 3' 6" piece of plywood to rotate on top <-- first real challenge
-- instead of a round, wooden pulley - fasten a (used, free) 26" bicycle-rim to the rotating piece:· Some are alloy, and therefore rust-proof.
-- when you get that far, we'll talk again.·
---- Do you have enough expertise with the SX or other computer to make a stepper-motor turn at your command ?
----·Can you make your own circuit boards - a real timesaver ? (it's not hard).
---- maybe you should start there.
-- also: I would mount a vertical sun-sensor separately, maybe even ON the moving mirror's base - that is a separate challenge (Oh boy !· Oh boy !)
there is more than one correct way to do almost everything.
·--- ps - that redrok site mentioned earlier·is priceless ! ! ---> http://www.redrok.com/
Post Edited (tommy) : 6/25/2007 7:26:54 AM GMT