The Cerebral Forum
Solar Projects => Heliostat Projects => Topic started by: Gabriel on October 08, 2012, 11:56:00 AM
Well, the Sun Tracking / Heliostat Program is mostly finished and fully functional, and the Sun Tracking / Heliostat Electronics are also pretty much done, so the next thing on the agenda is to come up with some machine designs.
There is, of course, this heliostat design http://www.cerebralmeltdown.com/heliostatprojects/Heliostat%20plans%20V2/default. (http://www.cerebralmeltdown.com/heliostatprojects/Heliostat%20plans%20V2/default.)
Since you only really need a saw and a drill to build the thing, it is accessible to most everyone who has basic construction skills.
There are some downsides though such as:
the lead screws would freeze in the winter
the lead screws tended to get bent easily which resulted in wobbly movement
the machine is very slow
the range of motion for the azimuth motion is restricted
Of course, many of these downsides could be overcome just by making some relatively minor changes to the design, which is something that I hope we will discuss here.
Another thing that I hope to discuss is an entirely new design.
What I'm planning on building for myself (and also possibly some other people since I keep getting asked if I sell them) is basically a long life, accurate movement, and minimal set up time heliostat.
This one will be based on wormgears, and I am going to try to make it capable of 360 degree motion along both the altitude and azimuth directions. I have attached a few screenshots to show what I am planning.
I wouldn't be surprised if it will end up taking me upwards of six months before this thing is really finished the way I want it to be. There are a lot of little things that I'm going to have to figure out. I am also going to have to invest in some small machine shop tools, so it will take some time for me to save up for them and also to figure out how to operate them.
Of course, the two designs I've mentioned so far certainly aren't the only ones.
There are a few others on this forum have shared.
If anybody has any input, I would be glad to hear it!
You've probably already seen this - http://stevenbrace.co.uk/2012/06/diy-arduino-motorised-time-lapse-head/ (http://stevenbrace.co.uk/2012/06/diy-arduino-motorised-time-lapse-head/)
but I was just drooling and thought I'd share...
Also, here's a neat article from a 1960 edition of Popular Science detailing a simple wayto make wooden worm gears. When I get some spare time, I hope to try out this method using a cheap HDPE cutting board from the dollar store instead of hardwood.
I actually hadn't seen either of those links before. Very cool!
On a side note, I'm currently trying to put back together the heliostats I made using lead screws as linear actuators, but with some modifications. I really want to hurry up and get something set up outside before the winter.
Yeah, I think wormgears would be great, but they do cost more (and I imaging a little tricky to make) than the lead screws you've used in previous designs. It's nice to use stuff off the shelf since it makes the design more accessible. I'm not quite sure how you're going to get the degree of rotation that you want with lead screws...looking forward to seeing your new design.
I've been brainstorming/daydreaming on how to modify common household items that use gear reductions for a heliostat design. So far I've come up with a salad spinner, which would be suitable since you could fit all the boards in the spinner itself to protect everything. The other idea was using an eggbeater, which is nice because the one I have is pretty skookum.
Ahhhh, brainstorming..... :P
I'm not quite sure how you're going to get the degree of rotation that you want with lead screws...looking forward to seeing your new design.
I won't actually be able to get any better of a range of motion for the leadscrew based design. At my latitude, and for the target angles I want to use, it isn't as crucial to have a large range of motion, but I still want to see if I can figure out something to increase the azimuth range though.
I personally think that Jim's design is the most simple and straightforward way to do it on a big scale ( 1-2m mirror)
I've had a few disappointing/frustrating days trying to figure out how to throw together a worm gear system with off the shelf parts. No luck here. I was thinking that maybe I could use something similar to this for the worm
How to make the corresponding wheel is the problem for me...
Anyone with any other creative solutions out there? My current design isn't working - it uses gears for the az, but I'm finding that when the driver board powers down, the stepper slips back down the tooth that it had partially climbed. *sigh* and it looked so pretty....:(
Here is an assortment of pictures of a linear actuator based heliostat that I put together. I am thoroughly convinced that a linear actuator based machine isn't really the best way to go since they are more complicated and have a rather limited range of motion, but I wanted to have at least one machine around to help with testing the Sun Tracking / Heliostat Program's linear actuator settings. I'm glad I did too because I did find one bug.
I will update the fixed version of the program soon (already done), but in the meantime I figure that I can at least show what I have been up to.
Overall, the machine does seem to track well,
but it is painfully slow (Nevermind, I tried it out at a higher speed, and it looks like the stepper motors can move the machine at a much higher speed without losing steps), which seems to be typical with the linear actuator based machines.
Here is a picture of the front of the heliostat. I haven't even bothered to clean the mirror yet since this is just a quick test.
Here is the back of the heliostat. I don't know if you can see it very easily, but instead of using door hinges for the alt rotation, I instead used galvanized lag screws as a pivot. I figured that they would hold up a lot better in the weather and are probably cheaper too.
It is a lot easier to align a machine to the correct azimuth direction if you are able to rotate the entire thing to point at the sun in sun tracking mode and then lock it into place to keep it there. Forget about that nonsense I have suggested in the past about using a compass for alignment, this is waaay easier. I have decreed it, all heliostats and sun trackers shall hence forth be designed with this ability built in. :)
It's probably hard to see this in the below picture, but the 2x4 sitting on top of the 4x4 post is screwed down so that it can't move. The 2x4 on top of it is clamped to the pipe, and the pipe is able to turn inside of a hole drilled into the 2x4 on the post. So the entire machine, leadscrews and all, rotates around the azimuth pivot point.
Here is a view from the side. Perhaps you can see the lagscrew better here.
Here is a view looking down on the leadscrews. None of the electronics are really protected from the rain since the machine is just temporarily set up for now.
If you want to get a closer look at how everything has been put together, here is the link to the Sketchup model. It's not exact, but it is close. I believe there are some errors, in particular where the mirror interferes with the 2x4s that are on top of the post.
Anyway, enough of that machine and on to bigger and better (well at least better) things. Like I said, worm gear based heliostats are a lot easier to understand. When using linear actuators, there are at least 6 more variables that have to be taken into consideration, and if you make a mistake on any one of them, your machine isn't going to work and you'll inevitable spend hours of your life trying to figure out why.
I don't have a lot to show yet as I've spent most of my effort experimenting with the design in Sketchup. It is still basically a variation of the design of the machine that I posted at the top of this thread. You can download the model at the below link if you want to see what I have so far.
Here is one picture of a mock model. This will never actually be completed as I just wanted to get a rough idea of the scale and the machine's rigidity.
The worm gears actually weren't very difficult to make. I think the method I used might have been similar to the one Jim used for his machine. http://cerebralmeltdown.com/forum/index.php?topic=293.0 (http://cerebralmeltdown.com/forum/index.php?topic=293.0)
For the worm part of the worm gear, I used 7/16-14 threaded rod. For the actual worm gear, I cut a circle out of some scrap plastic with a hole saw and then mounted it in a jig that it was able to rotate on. I then cut a groove around the perimeter of the circle by drilling "holes" with the drill press. Basically, I would drill a "hole", rotate the gear, drill another "hole", etc until I went all the way around. I then put a 7/16-14 tap from a tap and die set on the drill press set on its lowest RPM setting and used it to cut the threads in the worm gear.
It seems to work fine, although I haven't checked it for accuracy.
I don't think that it would be too hard to do it with just a normal power drill if you don't have a drill press, but you would probably have to make some kind of jig to make it work.
I made a worm gear in the past that worked pretty well. It's at this link. http://www.cerebralmeltdown.com/cncstuff/page3/wormgear/default.htm (http://www.cerebralmeltdown.com/cncstuff/page3/wormgear/default.htm)
I will add that cutting the threads with a tap was much easier than the method used there.
That's what I have so far. More to come.
It's very impressive how you can make this stuff. I wouldn't even have a clue where to start. It makes interesting reading and things are looking good.
I notice that the wormdrive arrangement you have in your final image is VERY similar to the arrangement in the Oyster unit I am using for my solar tracker. Probably jumping the gun here but are you aware that with a worm drive arrangement the horizontal and vertical movements are not totally independent? Specifically, as you rotate your cog around the worm drive shaft it will rotate altering the vertical pitch.
Worm Drive 8: Wheel rotating around worm locus (http://www.youtube.com/watch?v=RcsllgLLm70#)
In the Oyster's case this was pronounced enough to make a significant difference to the angle (and mess up the function of the limit switch). We had to add code to correct for this, it's all in my thread.
Yeah I know I'm going to have to compensate for that at some point. I don't think it will be too hard to do it though.
Interesting that the Oyster is similar in design. I guess there are no new ideas are there. :)
Just a quick question - I noticed the threaded rod you're using is quite larger than the 1/4" stuff that I'm using. Are you still using the method you described here:
or do you have a new method. I'd like to use thicker rod,just to beef things up, but I'm using helical couplers that have a 6mm openings on either end, so I'm stuck using 1/4" rod for now. It seems like I may have to invest in a tap and die set, which would be something new for me just so I can take a stab at making a worm gear and trying out your coupler solution. Also found another diy coupler similar to yours here:
No I actually just drilled a hole in the end of the leadscrew that was the same diameter of the stepper motor shaft. That's one reason I switched to a larger diameter leadscrew because it would be able to fit the shaft inside. Honestly though, it was really hard to get the hole straight enough and centered even on the drill press. It's time consuming enough that I don't think that I will try to do it again. Maybe if I had a lathe it wouldn't be so bad.
I think the method used at the link you gave would be better.
If you are only cutting plastic, then you can probably get by with the cheapest tap and die set you can find. However, if you think that you'd use it for metal at some point, definitely shop around for something better. My own cheap tap and die set has caused me so much aggravation.
I love to see clever uses for everyday objects - these guys use a car jack hack for their altitude adjustments! It looks like they're also using a gps and compass as Bob is currently doing to orient the tracker. Cool stuff! It's funny how people who have absolutely no contact with us are working on something so similar....I wonder how many others there are out there....
kudos to these guys!
Yep, it does sound similar. I've also wondered what would be the best way to have a solar tracker seek out the optimum direction to point on cloudy days. I've been thinking about experimenting with webcams pointed at the sky to see if I can somehow determine how much solar radiation my location is receiving at any given time and then logging it for future reference. I wonder if I could also use the same camera to guide the solar panel to point in the best direction during cloudy days. That's certainly not a weekend project though. :)
I'm happy to see that those small stepper motors are able to turn the leadscrew on that car jack. I ordered a couple of those (still waiting for them to arrive), but was uncertain if they would be powerful enough to do much. I also bought a couple of those EasyDriver Boards to try out. I'm planning on finally setting up two electronics systems so that I can leave one running continuously while they other is used for testing.
That's really interesting Paul L.
I came across Jeremy Blum's tutorials earlier and found them very useful but I had no idea he was working on anything like this!
Their design is very similar to mine actually. I notice that they are using an accelerometer to determine the angle of the panel, something I had considere. However, my machine actually uses the hall sensors in the motors and knowledge of the gearing system to determine the panel's angle.
Love their use of a carjack. It's brilliant in its simplicity.
It's been quiet on the forums lately, so I figured that I would do a quick post of a new heliostat design that I have been considering.
If you look closely in the attached pictures, you'll notice that this design moves differently than most others that you see.
Using this Wikipedia link as reference, you'll see that instead of having "pitch" and "yaw" like normal (where pitch = altitude and yaw = azimuth) this design has pitch and "roll."
This changes the math in the program some, but I think I already have it figured out. I actually didn't realize this when the idea first popped into my head, and it wasn't until I started experimenting with it in CAD that I realized that the Sun Tracking / Heliostat Program wouldn't work without modification.
Anyway, some of the reasons that I like this design are because it seems to be relatively easy to make, the structure for the mirrors is built into it the machine, and you can focus them if needed, and it seems like it would be sturdy.
That is one issue that I have with the first design I posted. It seems like it would be hard to build a strong one with basic materials, tools, and skills.
I've already attempted to make the gear boxes for this machine, but after a lot of effort still didn't turn out with something that I liked. It's obviously going to take some practice to make these things. :)
Happy New Years Everyone!!
I'm still thinking about heliostat design too. I've been interested in tensegritys for awhile and have often thought that it would be great to combine both heliostats and tensegritys. I was delighted to find someone else, Kevin Varner, had the same idea in 1982. I came across his patent here
Since I couldn't make heads of tails of it, other than the fact that it is lightweight and cheap, I tried to contact him. After a little e-stalking, couldn't find his email, but I did find his brothers', who I asked to forward my email. Hopefully he'll be in contact soon, and maybe could offer some insights on how to build a light, rigid support system....
I have always been interested in tensegrity type structures, but had never considered them for a heliostat design before. The patent link definitely looks confusing. I can't say that I really understand it myself, but I think that it uses the same axes of rotation as the heliostat design that I just posted.
If you're able to put this together, then kudos to you. :)
So I went over the tensegrity patent and thought I might throw a miniature version together just to see if I had the basic priciple down. I've uploaded a short video below. Eventually, if this were to work, I might consider flipping the whole thing on its edge; instant heliostat totem! It would take up less space in the yard too. The setup I have in the video looks complicated, but in actuality, was pretty easy to put together and took roughly 40 min.
I'm having trouble though with the azimuth tracking - I haven't put anything together yet because from what I can tell just by looking at drawings, the patent design just wouldn't work. The "Bail", which I've circled in the picture, moves along a wire, supposedly adjusting the mirrors azimuth. Am I crazy, or would this work? In any case, I'd rather use a different mode of getting the mirrors to move along their azimuth axis, but as to how, I'm not sure. If anyone has any ideas don't be shy about sharing. Otherwise, I'll be dustbining this setup.
I am one of the lucky few that has one of Gabriel's SunHarvester shields, along with three driver power boards, and have just finished soldering on all the screw terminals yesterday. I've hooked everything up and it's all working very nicely, though I've run out of wire to hook up my last pair of stepper motors. Nonetheless, I've got four motors running - enough for two heliostats, and I can't say enough about how I am impressed with the whole affair. Especially the joystick control and the ability to switch between target groups. And the button to set the time precisely is an excellent addition as well. Setup is fairly straightforward too (though I did have to go and watch his demonstration video again to figure out how to enable joystick control) . If anyone is thinking about ordering a SunHarvester Shield, do it.
Tensegrity Heliostat (http://www.youtube.com/watch?v=i3_t8xHzQIo#)
I'm impressed that you did that in 40 minutes. It definitely would have taken me a lot longer.
I'm having trouble though with the azimuth tracking - I haven't put anything together yet because from what I can tell just by looking at drawings, the patent design just wouldn't work.
That's actually what I was referring to in my previous post. It's not rotating along the azimuth, so if you try to use the current Arduino Sun Harvester Program, V097_e, it won't have any hope at all of tracking correctly.
I think I have the math for it figured out though, and it wouldn't be hard to try it out.
If anybody is interested in trying it,
Go to the "functions" tab and find the function named "FindHeliostatAltAndAz".
Inside the function, find this code
and replace it with this code.
machinealt = 90 + (-atan2(y,z)*(180/pi));//actually pitch
machineaz = asin(x)*(180/pi);//actually roll
That should, I think, make the program work with "pitch and roll" type heliostats.
Keep in mind that this is just a quick hack of the program, so a lot of things will be mislabeled since they shouldn't really be called "azimuth" or "altitude" any more.
The first heliostat I ever built was a "pitch and roll" type heliostat. I was absolutely clueless of that fact though until I just recently started experimenting with different heliostat designs.
I built both of these heliostats, and neither one worked. I double checked everything in what was then a PC based program, but I couldn't figure out what was wrong.
Then I built this heliostat, and it magically started working.
All of these years, and I finally realized what was wrong!
Now that I look at the design, it doesn't look half bad. The Pros are that the weight is equally balanced, so you wouldn't need very large stepper motors, and that you essentially have an "unlimited" range of motion. The Cons are that you have to figure out new dimensions for different sized mirrors and also that it would be a little trickier to set up correctly since you have to deal with two "posts" to hold it in place.
Now that those Sun Harvester Shields are in the hands of a little more than half a dozen people, hopefully we'll start seeing more and more designs popping up.
And now time for something completely different...
I was trying to think of ways to get my azimuth rotation for the previously mentioned tensegrity setup when I realized that I could use hydraulics. Then I though, well, why use gears at all for any setup? Hydraulics kinda make sense - I haven't done any comparison pricewise, but tubing might cost less than all the wiring we'd need to run power to each of the steppers on each of the machines. Additionally, the steppers and circuitry could be kept in inside, protected from the elements. Check out these videos to get an idea of where I'm going with this...
DIY Syringe Pump Fix (http://www.youtube.com/watch?v=Gq4RQHLfJA4#)
Engineering Project - Syringe Accurate Hydraulic Crane (http://www.youtube.com/watch?v=-ShET4if9aI#)
and finally, this beauty: http://iaus.com/StudioFiles/AniHydro/AniHydro.html (http://iaus.com/StudioFiles/AniHydro/AniHydro.html)
Cool stuff. I wonder what kind of precision one could get by using a hydraulic ram driven by a stepper motor connected to a positive displacement pump...
Could you equate the "pitch and roll" type heliostat to an equatorial telescope mount? If so, the pitch stepper motor (similar to declination) would only need to make a few steps each day, compensating for orbital motion.
Would a "pitch and roll" heliostat then be able to use less power? The pitch motor could be turned off for large amounts of time between steps, assuming the heliostat is self-locking. But the roll motor is probably making more steps, so it might even out.
It might even be possible to build a heliostat with only one stepper motor. If you went out every day to fix the pitch manually, or every few days, if you have a large target.
Hi dlaing, and welcome to the forums!
I'm not very familiar with equatorial type telescope mounts, so I'm not sure if they are the same or not.
The "single axis heliostat" is an idea that I've heard mentioned in various places around the net. I've always been skeptical of it though because despite all the talk I've never seen anybody build one.
I have, out of necessity really, written a simulator that graphs the angles required of these heliostats to reflect the light at the required target. At some point, I really need to clean up the code and upload it so that everyone can experiment with it, but that's for another time.
Anyway, after some experimentation, it looks like for this to work you would not only have to change the pitch manually but also you would have to use a different target altitude every day too to compensate for the sun's declination.
It seems that in order for the pitch to require essentially zero adjustment over the course of one day, you would also need to adjust the target's altitude according to this formula.
targetalt = 90 - latitude - sun's declination
The azimuth would also have to be zero, or at least close to it.
Considering that the sun's declination fluctuates by roughly 47 degrees over the course of the year, the target would have to be moved quite a bit every day, depending on how far it is from the heliostat of course.
So I guess you could technically make it work in certain circumstances. I guess it just doesn't work quite like most people expect though, which is probably why you don't ever really see them.
Strictly speaking though, the amount of power required to run a heliostat is extremely tiny compared to the amount of power you get back out of it. I'm currently experimenting with a kilowatt meter to get a rough idea of how much a heliostat might use. Hopefully I'll have something posted soon.
Thanks for running the numbers for this. I'm not surprised it doesn't work, like you said someone would likely have built one already.
The simulator would be very interesting to see. Is it a standalone program, or one written in matlab?
If we could quantify the power usage of the two heliostat types, then we might as well choose the more efficient geometry.
The program is written in MatLab. I think it is simple enough though that it would run in something like FreeMat or Octave which anybody can use. Mostly the script is just a mess right now, which is why I haven't gotten around to uploading it yet. I'll try and hurry up and clean it up though since it would be useful for everyone to have.
The geometry technically wouldn't change the efficiency. The mirror normal vector still points the same direction in both cases, it just travels a different "path" to get there. Although I suppose certain geometries are prone to shade parts of the mirror during certain times of the day just because of how their structures must be built.
I did a little more research and there appears to be a working single axis heliostat.
They appear to have the primary axis of rotation on the line between the heliostat and the target and the target azimuth is at 0. Does that mean the trick that the mirror cannot be in line with the primary axis of rotation? And it appears that the two axes aren't orthogonal. If you have time to run another simulation, I would be interested in the results.
I was refering to electrical efficiency, not solar efficiency, but I didn't say so explicitly. Sorry for the confusion.
Oh, this design actually makes a little more sense. I guess this is more in line with what you originally mentioned. From what I've read at least, you do have to use a target azimuth of zero, but I believe that you can use any target altitude that you want. You would have have to change the tilt of the mirror frequently though. In the document you gave, they referred instead to changing the shape of the parabola.
Incidentally, this type of machine is often called a coelostat.
The simulation I wrote wouldn't work with this since it is a slightly different design from the other heliostats I have experimented with. I'm sure that it would work though.
I did some testing with my Kilowatt Meter, and found that my heliostat circuit would use about 0.03KWH per day. Just to make sure that I would over estimate instead of underestimate, my example day was the longest of the year, summer solstice.
I'm sure that I could get the power usage significantly lower by using a faster speed and acceleration too. There are other variables that could be tweaked too. For example, changing the design of the machine so that it is well balanced and doesn't required much power from the stepper motors to move it would help. Adding a simple light sensor so that the machine only moves on sunny days would also help.
Anyway, 0.03KWH per day is 10.95KWH per year. Assuming a price of 10 cents per KWH, it would cost about $1.10 to run this heliostat for an entire year. This is like I said just a rough estimate, and I think I could get it down a lot lower than that even.
You guys seem to be suffering from the same problem I have: lack of a MakerBot or similar 3d printer.
Well it took four tries, but I finally managed to put together a halfway decent heliostat. It's still not 100 percent finished, but I took it out for a test run and it did pretty well. I need to add more mirrors and finish off some weatherproofing but here are some pictures of what I have so far.
Here's a picture with the top removed.
Here's a picture of the bottom part opened up.
I hope to make another one in the future after I find all of the bugs in this design. Building these things has required a serious learning curve as I didn't have much experience with welding and machining when I first started. You can obviously tell by comparing some of the earlier pictures in this thread with the ones here that I've made some big improvements in the way I make things. I still need a lot more practice at making precision parts, but I'm getting there.
Hopefully I will get faster at making these things too. I think if I make jigs to help align everything then I should be able to make them pretty quickly. For this one, there was a lot of extra time taken where I had to do things like grind parts down to make them fit. This was due in large part to the fact that I didn't have everything completely thought out before I started.
I plan on showing it off some more once I get everything finished. We'll see how it holds up long term, but hopefully it should work well enough for me to get some good use out of it.
Well, I have to admit that I've been anticipating this post with bated breath for awhile - you always seem like you're making leaps forward on your design, and this is no exception. Kudos! Let me know when you have them for sale to the general public! :) Fingers crossed that everything goes well for you - this build looks awesome!
Here are a few more pictures of my more recent progress.
Each time I try building one of these things I get a little bit closer to achieving a solid heliostat design. This one still needs some improvements, but it does at least work well enough to try it out in the sun. I will post more on it in the near future, but, for now, here are some pictures to get the ball rolling.
Each of the 6 1ft^2 mirrors is adjustable. This way, the light can be focused on a smaller target. I originally planned on adding 12 mirrors total, but I decided to keep it at 6 for now while I try things out.
For now, Iím only shining the light through a medium sized window.
This picture doesn't really show it, but the reflected light does a good job of lighting up the room. Standing in front of the light beam feels just like standing in front of the light from the sun.
Nice work Gabriel!
I'm wondering how you've made each mirror adjustable - are you using the same system in your first pics in this post with the adjustable blocks of wood with the threaded rod running through it? Set and tighten? I've wondered awhile about making adjustable mirrors and the only other system I've seen is the one used in the "solar death ray mirror - I can't find the link now, but they used several small mirrors with a dab of silicone to hold it in place, and then used screws on the four corners to set the angle of the mirror.
Anyway, good job!
Sorry about the very delayed response, but here is a picture that shows how the mirror is attached. The plastic piece is made out of 3/4" pvc. There are two parts to it. One mounts on the pipe for extra azimuth adjustment. The other has a rod that goes through both it and the first piece of plastic to give extra altitude adjustment. There is one of these for each of the mirrors.
The plastic is glued to the mirror using epoxy, which seems to hold really well.
This seems to work OK as a quick method of attaching mirrors and making them adjustable, but I think there is probably a better way to do it. You'll notice that the mirror in the picture is actually broken. I'm not sure exactly what happened, but the epoxy seems to have bowed the mirror outward as it dried. In this case it did it enough that the mirror broke. This bowing also has the effect of spreading out the reflected light after it hits the mirror. If you look at the pictures posted earlier, you'll notice that the light is spread out from one side of the window to the other. If the mirrors weren't bowed, I would have been able to focus all of the light into the window, which is what I wanted.
If you are interested, here is the Sketchup model of this heliostat. It probably isn't exact to the real thing, but it should at least give an idea of how everything is put together. I don't really recommend building it. I'm already working on a better design which I hope to start building soon.