Low Tech HELIOSTAT. How Do I Attach the Suntracker?




About: I am a stone mason. My hobby is making new solar cooking and gardening stuff. I have used solar heat to cook soil for a couple of years. In mother earth news in January, i read that their compost expert does...

I recently made a solar accumulator which is a limited heliostat, I guess. It can only send the light to somewhere on the axis of the equatorial mount.
The new heliostat is different. It should be able to send the light just about anywhere you want and keep it on that spot as long as the sun shines.
This feature could be massively useful for solar cooking (multable heliostats shinging light on one area) water heating or solar lighting, (shining the light into dark corners or northern windows).
I want low tech ones that poor people who do not have computers can make.
I am pretty sure this works now after making a little model. I am using gimbals on the real thing having seen a problem on the model. Thanks to all who helped me choose gimbals for where it joins to the equatorial mount!
The theory follows on the next couple of pages.

Step 1: Basic Optics

When sunlight hits a mirror, it bounces off at the same angle as it hit the mirror.
The plane of the mirror is exactly at right angles to a line bisecting the bouncing lightrays. I think we should focus on the bisecting line!
If we allow the mirror to swivel on its centre point, and tie the corners of the mirror to somewhere on that line and keep it taut, the mirror will point at the correct angle to send the light to the target!

Step 2: Like a KITE

If we get 2 similar elastic bands and tie them together, and tie ones end to the sun pointer and the other ones end to the target pointer, where the bands are tied to each other will define a point on the bisecting line.
Furthermore, the kite mirror will now be pointed in the right direction as long as the sunpointer follows the sun.

Step 3: Put It on Equatorial Mount!

If the bands do not work, you could try a spring loaded spool of thread with 2! threads coming off the spring in the middle. One thread in place of one band and one in place of the other. This will keep the "kitemirror" taut.
Here is what it might look like on equatorial mount.
It might work better if the fixed target pointer comes up through the middle but it makes for a more difficult to build device.
The sun pointer moves during the day, the bands move as it does so, and the centre and the mirror is moved as the bands move too.
Perhaps a device with 2 pointers was used by The Archimedes Heat Ray to aim it and keep it on target?
Archimedes was a whiz with parabolics too so his heat ray mirrors might have been concentrating mirrors.

Step 4: Gimbal the Start of a Practical Heliostat

I have been hunting for ways to turn the idea into reality. Many thanks to people who answered my forum question. The universal joint answer led me to a gimbal as my solution!
While searching for a metal ring to make a gimbal, I made one from wood.
I stuck mylar to a rectangular piece of plywood for the outside and used stiff wire and screws to attach the gimbal section. I hope that it is close to equatorial mount.
Since taking the photos, I have added a target pointer.
I have not yet decided how to attach the sunpointer.

Step 5: How to Set Up and Adjust Equatorial Mount?

If the equatorial mount is pointed exactly north (in the northern hemisphere) and in line with the axis of the earth, then all is rosy.
How can we get it to this state? It is hard to find exact north! I suggest the first thing to do is get the heliostat set up pointing up at exactly at your latitude no of degrees. Point it approximately north.
Over the course of the day, as it turns the reflection will move a little. You can use this movement to readjust the mount until it is perfect. I have not yet figured the movement out because I have not yet made one! Perhaps this work can start tomorrow!



    • Woodworking Contest

      Woodworking Contest
    • Colors of the Rainbow Contest

      Colors of the Rainbow Contest
    • Party Challenge

      Party Challenge

    63 Discussions

    Nice effort! For anyone looking for a turn-key solution, see the H1 heliostat over at http://www.lightmanufacturingsystems.com/heliostats . It uses a metalized PET mirror, which is unusual.


    9 years ago on Introduction

    I have been working on this for  about 7 months now.  An equatorial mount does work at equinox as you say with a 48 hour rotation or any day with a 24 hour rotation if the target is in line with the polar axis (principle of coleostat). At redrock heliostat website, 3 mechanical heliostats are shown: Focalt, Gamby, and Silberman.  The last two suffer from gimbal lock when the sun and target are closer than 30 degrees, there is a good article on this in Wikipedia.  I have not been able to understand the Focalt mechanism yet. It seem as if the tape, rubber band and corkscrew methods might also have mirror control problems at low sun, target angles.

    2 replies
    Daniel Murphy's latest work can be seen on this video.   (I have other easier projects where I can contribute more!)
    I think if you pause the video from time to time it is easier to understand what he means. Brian

    Good to hear from you!  I abandoned it for a while.  I think last time I emailed with you we noticed that a reference picture of a Gamby had the equatorial axis  pointing the wrong way.  Correct? 
    But really glad that you have made progress.
    When we know the problems beforehand, we can make arrangements to move the target or heliostat  to get round  the problem during year!
    I have made progress on tracking (turning the equatorial mount)!  I have a fuller sized image at http://solardesign.ning.com/photo/liquid-piston-tracker and I may do an animation too.


    I have been looking into a heliostat on equatorial mount with a 48 hour rotation. I have not been able to conduct experiments due to clouds. Even if the 48 hour rotation does not work perfectly, it may be part of the solution. If it is "nearly right" then all that is needed is a 24 hr rotation for the sun pointer to make fine adjustments to the angles of the mirror. This means no huge swings on the gimbal (perhaps a small ball joint is all that is needed). Like the ball joint in a tractor 3 point linkage. What do you think?


    10 years ago on Introduction

    Hello, I had a similar idea a time ago, and I never got to develop it totally. Seemed in its objective, but nothing with the rubber band. The equatorial mount by itself is sufficient. If the sun takes 24 hours a turn, we only needed a mechanism that gives a turn in 48 hours. for example, the hour axis of a clock, geared down four times. In this way, if at sunrise dawn? we put the mirror to 45 degrees on this axis, so that the reflected ray upwards vertically, at noon the mirror will have to be horizontal, and to the dusk again to 45 degrees in the opposite direction. This approach must be duplicated symmetrically to the equatorial axis, so that the following day the process can be repeated. But the sun is not easy to convince, and it is guided by its own temperament. It would be necessary to make almost daily adjustments to the approach, because there are no mechanic ways to follow the sun in the sky. Pardon my "automatic translator" English.

    13 replies

    Sorry. The 48-hour rotation idea won't work, except in one case. That case is if the date is one of the equinoxes and the target direction is on the celestial equator. In other situations, it may seem to work at sunrise and sunset, but not at other times of day.

    For example, suppose you are on the equator at an equinox, so the sun rises due east, rises straight up the eastern sky, passes overhead at noon, then goes vertically down the western sky to sunset. And suppose you want to reflect light due north, horizontally. At sunrise, the mirror must point horizontally north-east, so light from the sun on the eastern horizon will be reflected to the north. And at sunset the mirror must be aimed north-west. So, you might say,the mirror turns 90 degrees in 12 hours, so that's 48 hours per revolution, with the axis of rotation vertical.

    But that would mean that at noon the mirror is pointing due north, aimed horizontally, and that won't work at all! In order to reflect light from the overhead sun toward the northern horizon, the mirror would have to be aimed upward at 45 degrees. Your 48-hour rotation won't do that.

    We have discussed some heliostat designs that work. A mirror that rotates once every 48 hours won't. Sorry!

    And it *is* possible to follow the sun mechanically. It doesn't have any "temperament". Its motions are highly predictable.


    I don't agree.

    If the assembly is equatorial, my 48 hours rotation idea functions EVER, be equinox or solstice or intermediate. Your assumption of an equatorial subject is yours, not mine. Please rethink that thing.

    About the "equation of time", that is to say the movements of the sun in the sky, the mechanical complexities of a device that keep in mind them, they are larger than what a person can confront. Or, if you prefer it, the relation cost / benefit is too high so that be worth while.

    david williamsredrok

    Reply 10 years ago on Introduction

    Yes. I know of two designs for mechanical heliostats that allow for the Equation of Time. One uses a cam-and-lever device, with a cam that rotate once a year and is shaped so its radius is basically a graph of the Equation of Time. The other makes use of the fact that, to a fairly good approximation, the Equation of Time is the sum of two sine waves, one with a period of six months and the other with a period of a year. Two rotating cranks, one turning once a year and the other twice, can be linked so that a motion like the sum of two sine waves is produced. As far as I can see from the diagram on your website, the Hultberg machine works basically as the second of the above designs. There are probably other ways to do it, too. But, nowadays, that kind of technology is kind of archaic. Mass-producing the machines might be fairly cheap, but hand-making them one at a time would be very tedious. It's so much easier to use a computer! David

    david williamsredrok

    Reply 10 years ago on Introduction

    Yes. I did skim through the patent. I didn't study it in detail, but as far as I can see, the design *should be* capable of great accuracy. The limitation I can see is that the gear ratios must, obviously, be ratios of integer numbers of teeth on cog-wheels, which means that they are unlikely to be exactly right to match the astronomical parameters: the number of days in a year, the tilt ("obliquity") of the earth's axis, the eccentricity of its orbit, the dates of perihelion and the vernal equinox, and so on. So the machine may be good, possibly *very* good, but not perfect.

    But I am puzzled as to WHY Hultberg went to all the trouble and expense of inventing this machine and patenting it. The patent reads like something from the 19th Century, but actually it was filed in the 1980s. The cog-wheel technology shown in the diagrams was already being replaced by computers. (Actually, in one of the claims, computers are mentioned. It looks like he realized at the last minute that this new technology was about to supersede his invention, so he tried to claim it too.)

    I doubt very much that any of these machines have been made except for Hultberg's own prototype.

    Fun, though.



    Reply 10 years ago on Introduction

    Thanks, Duane. I need the apparatus for a solar kitchen, therefore I don't need precision. My discussion with David was conceptual, and I interrupted it for lack of time. We could not come to an agreement.


    Every inventor eventually realizes that every simple idea has already been thought of, probably many times. If your 48-hour rotating mirror worked, we would all be using it for heliostats, and wouldn't be discussing the more complicated machines that are described on this page. Unfortunately, it doesn't work, except in the very limited case I described previously. You don't have to believe me. Try it! Just take a clock mechanism, add a 4:1 gear to slow down the 12-hour rotation of the hour hand to the 48-hour rotation that you want, attach a mirror, and see if you can use it as a heliostat. Be sure to try it with the target direction well away from the equator, e.g. to the north if you are in the northern hemisphere. You won't have to spend much money, if any, which is good because you'll be disappointed. Sorry! As for the Equation of Time, sure, it is a fairly complicated function, but it would be easy to follow mechanically. Just use a cam-and-lever mechanism. Have a shaft that rotates once a year, and put a cam (a non-circular wheel) on it. The shape of the cam should be such that a lever resting against it moves according to the Equation of Time. Then use the motion of the lever to adjust the heliostat clock so that the machine follows the Equation of Time properly. This sort of thing was commonly done back in the time when intricate mechanisms were widely used. Nowadays, we tend to use computers and electronics instead, but there's no reason why the mechanical device should not still be used. However, in the old days mechanical heliostats worked by clockwork which had to be wound up every day or two. So some sort of manual maintenance had to be done frequently, which meant that the adjustments for the Equation of Time and the variation of solar declination could be conveniently done by hand. I have never seen them done automatically in a mechanical heliostat, even though it would be perfectly possible. Anyway... Try building your 48-hour thing. Let us know what happens!


    Thanks for the response, David.

    When you say "...Just take a clock mechanism,..." etc, lacks to add "put the array over an equatorial mount". If you omit this step, the device does not function. You think the following thing: during a single day, and especially considering only the diurnal hours, the displacement of the sun in the sky is practically circular, be in winter or in summer. Therefore, an equatorial mount is the unique thing that does you need to trace it. You can be sure that it will not be set apart of its path more than some seconds of arch.

    An equatorial mount or assembly is one whose axis of turn is parallel to the terrestrial axis, that is to say that is oriented in north-south direction, and inclined with regard to the horizontal one angle equal to the latitude of the place.

    I live in Argentina, south of South America. Then, I speak Spanish. Therefore my English is not good, pardon. I use http://freetranslation.com/ and then I apply corrections according to the little thing that know of the language.

    Thanks also for the cam-and-lever mechanism idea. I did not think this, I only thinked combined movements, this is much simpler. But one must say that even this method is not the last word as for an exact position, because the displacements of the terrestrial axis are not contemplated in the equation of the time, due to they are erratic and unpredictable. Obviously, the differences are despicable for the almost totality of the applications.

    I did something seemed to what you suggest, three or four years ago, and functioned here in the backyard of my house, approx. 32º of south latitude, in full autumn. At least, functioned between the hours 11 and 13 approximately. As soon as it have time I will try to refloat the apparatus and I will it to work several days. It does not contemplate the annual displacement of the sun, that is to say that in this epoch I would have that to adjust the north-south inclination of the mirror almost daily.


    Thanks for your e-mails. Okay, I'll visit here to correspond with you.

    I managed to see the 1843 machine to which Brian referred. For some reason, the link would not work for me yesterday. Actually, the image shows a collection of machines made by Silbermann, who was a famous maker of these things at that time.However, I cannot see any reference to a mirror turning once every 48 hours. The machines have an ingenious mechanical device for bisecting the angle between the directions of the sun and the target. The basic mechanism, defining the direction of the sun, rotates once every 24 hours.

    I suspect that the person who wrote the "explanation" for Brian knows even less about the thing than Brian himself does.

    I am beginning to wonder if the 48-hour machine is something like Leonardo da Vinci's helicopters - imagined and maybe drawn, but never made and tested.

    Back to our discussion. For simplicity, imagine that you are at the earth's South Pole. The earth's rotation axis is vertical, with the Celestial South Pole directly over your head. Suppose the date is December 22, the summer solstice, so the sun is moving horizontally 23.5 degrees above the horizon. Suppose that the target at which you want to reflect light is in a horizontal direction, as seen from your mirror.

    At some time of day, the sun will appear to pass directly above the target. The bisector of the angle between the directions of the sun and the target will be pointing 11.75 degrees above the horizon, or 78.25 degrees from the vertical axis of the earth.

    12 hours later, the sun will be 23.5 degrees above the horizon in the opposite direction. The target is still where it was originally. So in what direction does the bisector now point? Almost over your head, that's where! it will be only 11.75 degrees from the vertical axis of the earth, at the same azimuth as the target.

    So the angle between the earth's axis and the direction of the bisector - along which the mirror must point - varies between 78.25 degrees and 11.75 degrees during the day. So it is absolutely certain that the mirror can *not* be fixed to a polar axis, rotating only about it.

    The same situation occurs, but is less simple to describe, for any other location on the earth. It is not peculiar to the poles.

    As I have said so many times, please DO THE EXPERIMENT. HAZ LA PRUEBA. Forget about everything that I and others have said. Learn from actual observation. That's science.

    Best wishes.




    Where you say " ...it will be only 11.75 degrees from the vertical axis of the earth",
    you must say "...it will be only 11.75 degrees south of the intersection of the meridian of the place with the heavenly equator"

    That is, in my latitude, "it will be approx. 21.25 degrees north of the zenith." You cannot see it?

    If I could send you a drawing, I believe that you would understand it. But you have said me that cannot receive it.



    A drawing would be misleading because it is two-dimensional. We MUST think in three dimensions. Use a globe instead.

    Why do planes flying between Europe and the west coast of North America go way to the north over the arctic? Because that's the shortest route. Look at a globe (NOT a map) and you will see this is true. The mid-point of the journey between Vancouver, on the west coast of Canada, and London, England, is close to the Arctic Circle, over northern Canada, even though both cities are close to the same latitude of 50 degrees north. This shortest path is called a "Great Circle".

    If the earth were transparent, an observer at its centre would see the direction of the mid-point of this Great Circle route as the bisector of the angle between the directions of Vancouver and London.

    The same is true of the situation I hypothesized above. In three dimensions, the angle between the target, on the horizon, and the sun, 23.5 degrees above the opposite horizon, is in a *vertical* plane, which also includes the axis of the earth. The bisector of the angle is only 11.75 degrees from the axis and the celestial South Pole.

    It is essential in this problem to think in three dimensions and to consider Great Circles, rather than straight lines, as the shortest distances between points.