The Rossen Mini-Heliostat FAQ


This FAQ (Frequently Asked Questions) was written to respond to those people who have asked questions about my current project, the design and eventual implementation of a cheap heliostat system for electricity production. It is not yet an exhaustive list of answers, so feel free to ask more questions.


Click on the thumbnail image to see a larger version.

rectangular field, side view rectangular field,  eye-level view rectangular field, top view
LEFT: A schematic of my original vision of what a mini-heliostat might look like.

PNG 15k

CENTER: The back side of my prototype heliostat showing some of the gear system.

JPEG 132k

RIGHT: The front side of the heliostat prototype reflecting the sky.

JPEG 101k


1. What is a heliostat?

2. What are heliostats made of?

3. What is a Mini-Heliostat?

4. What is the advantage of a Mini-Heliostat?

5. How much will it cost?

6. Have you made one yet?

7. Why didn't you calculate a manufacturing profit?

8. Why hasn't this been done before?

9. Can I get involved?

10. What did the Australians say?

11. What microcontroller are you going to use?

12. What type of heat engine or receiver are you going to use?

13. How big is your work shop?

14. What is the status of the project?


1. What is a heliostat?

A guided mirror. In solar energy applications a heliostat is a mirror that reflects the sun's rays to a fixed target and is adjusted to follow the sun so that the target is always illuminated. The adjusting mechanism can be anything, including human slaves, spring-wound clockwork, or computer-controlled motors.

2. What are heliostats made of?

Glass, steel, concrete, and money. Until now, most heliostats have been arrays of glass mirror facets about 1 to 2 meters wide, mounted on a rigid steel frame, and carefully bent to focus on a receiving tower. Typically, there are between 50 and 100 square metres of mirror surface on a frame. The frame is mounted on a steel pillar which is fixed in a concrete footing. There are two motors in the heliostat, one for azimuthal movement and one for elevation movement. Everything is controlled by a central computer that makes the calculations for adjusting many heliostats.

Heliostats are designed to last for 20 years and withstand winds up to 120 km/h. They cost about US$300 per square meter of reflective surface. A new design, based on a drum-shaped flexed-membrane mirror of 50 square metres area, promises to cost US$80 per square meter, if made in quantity (between 5,000 and 50,000). Given the present low demand for solar energy, it is unlikely that any company will invest in the money necessary to start producing this new generation of heliostat.

3. What is a Mini-Heliostat?

Basically, a small heliostat with a mirror 30 cm x 30 cm. Six to 16 of these mirrors are mounted on arrays, ganged to 2 shared motors. The motors are controlled by a mini-controller shared with 8 other arrays. For a 15k PNG picture of what a simple mirror unit might look like, click here.

4. What is the advantage of a Mini-Heliostat?

It is easier to make. By making the mirrors small, one reduces the wind loading on the structure, thus allowing fabrication with weaker materials, i.e. plastic. When one uses plastic instead of glass, steel, and concrete, the costs of development, tooling, production, installation, maintenance, and modifications plummet.

5. How much will it cost?

Between Sfr 12.- and 25.- per square meter of mirror surface. The current (Mar. 1998) US-Swiss exchange rate is about 1.48 francs per dollar. This cost takes into account production and installation costs, but no profit for the manufacturer. Why no profit? See question number 7.

6. Have you made one yet?

Yes. Version 0.1 of the Rossen Mini-Heliostat rolled off the production line on September 26, 1997. The mechanics work (sort of), but for the moment there is no computer control. Computer control will be active around mid-October 1997. The main purposes of 0.1 were to test the ease of manufacturing from semi-finished plastic and to get hands-on idea of what the acceptable tolerances are. The secondary purposes were to show-off and have some fun.

UPDATE November 2, 1997: Priorities changed and computer control is still not ready. There was an accident with the heliostat and it is now no more that a couple of photographs and some bits of plastic. It was left outside on a table during a wind storm. When it slid off the table its heavy glass mirror smashed to bits. Autopsy showed that only the glued joints, the glass mirror, and the overly-delicate elevation ring were in any way damaged, i.e. the solid plastic pieces were unharmed. This is an argument for either reinforcing the joints and/or molding the heliostat in a couple of pieces.

For an idea of what it looked like before the disaster, here is a 132k JPEG showing backside of heliostat prototype. Note that not all of the gear train was in place when this photo was taken. Here is a 101k JPEG showing the front of the prototype and the sky reflected by the mirror. The white stuff in the middle of the mirror is a small cloud, not bird crap.

7. Why didn't you calculate a manufacturing profit?

Because selling heliostats is not my primary goal. First of all, I don't believe anyone (except for maybe high-temperature furnace companies) would be immediately interested in this system. In other words, there is no immediate market large enough to keep me and my company from starving if I try to sell heliostats. Secondly, I wish to exploit the system myself and sell solar electricity without having to trouble myself with the hassle of salesmen, marketing gurus, and other salaried employees. Partners maybe (see question 9), but nobody who is unwilling to risk their time and money with mine.

Finally, the system is probably unpatentable. As far as I know, heliostat patents ran out about 10 years ago, and the only new idea in this system is that it is small. This is probably not sufficient for a new patent, even though it is a good idea that no-one seems to have done yet. This means that if I go into the business of selling these things, I am completely unprotected from the first knock-off company who sees that there is a market. Since the tooling cost for this system (or something similar) is under US$50,000 and the system is very simple, it is easy for anyone to copy. Therefore I will build for the purpose of exploitation, and if someone comes along with a cheaper system than mine, I will be happy to buy from them.

UPDATE June 1, 1998: That is not to say that I won't sell some heliostats to anybody who wants some. It is just that at the moment I simply don't have the time to go looking for customers. However, if you are interested in buying enough heliostats to make a 0.5 MW solar furnace for only Sfr50,000 (about US$33,000), reply to the address at the bottom of this FAQ.

8. Why hasn't this been done before?

Cheapness has never been the primary goal before. Specifically, it was never the primary goal of the aerospace companies who designed the first heliostats to manufacture a cheap, low-tech system by the millions and then exploit it to make electricity. They were being paid with government money to do research and it was more important to have a big, impressive, high-tech system to show to the politicians (who were giving them the money) when the politicians came to have their photos taken for the newspapers.

To be fair, I will admit that in the 1960's, when the first heliostats were designed, it made good sense to control large mirror surfaces with the minimum amount of electronics possible. Since then no-one has reconsidered the parameters of the problem (i.e. the reduced cost of computers and other electronics) and we are stuck with systems that are described by solar engineers as "proven technology". I prefer to use the word "obsolete".

I acknowledge that I am not the only person to think small, but I believe that I am the only one to consider the implications of using plastic construction. The other two people that I know about are Duane Johnson of Red Rock Energy and David Wells. They publish their results and opinions regularly on the home-brew solar concentrator mailing list. Both of these guys have built smallish heliostats (mirrors on the order of 2 square metres) for significantly less money than the US DOE spends, but the systems are made of glass and metal and probably will need a lot of start-up money to get out of the "garage-built" stage. Good luck, guys!

9. Can I get involved?

Why not? Make me an offer. The original concept of this project was as a one-man job and I have enough resources to finish this project alone (the design and prototype of the heliostats, at least), however there are a number of skills that I need to learn or improve to finally arrive at a successful implementation. It would speed things up tremendously if I had some competent partners. In the near term, I could use someone to put together the heliostat mini-controllers. I am thinking about using PIC-based microcontrollers, but anything suited for hobby robotics will do (as long as it is cheap). It would also be useful to know someone with experience in designing plastic parts.

I have very little doubt that I can get the mini-heliostats to work for the above-mentioned cost, but I am concerned about the receiver system. I think I will have need of a mechanical engineer who is good with his hands or a machinist who is good with his brains. David Wells has given me a reference to an Australian team who converted a standard diesel engine into a 25 kW steam engine to use with solar-generated steam (see question 10), and I am thinking about copying their technique if I can't find a better system for power conversion.

Naturally, I would prefer partners living in Switzerland, but it still might be possible to work over the Internet and by mail. They say that virtual corporations are the next Big Thing... At the very least, I am interested in any comments and suggestions that you might have for making this project a success.

10. What did the Australians say?

On January 27, 1998 I finally met with the team at the Australian National University who were responsible for the White Cliffs Solar Thermal Power station. For those who haven't heard of it, the project involved using 14 parabolic dishes, 5 metres in diameter, to produce steam for a converted diesel engine. The power station produced 25kW electric on sunny days for the small opal-mining town of White Cliffs, NSW. The power station was designed to be stand-alone and had batteries and a diesel back-up generator. The station was decommissioned in 1993 when the town was finally hooked up to the rest of the NSW grid and the equipment was sold off.

Using the experience gained from this project, the group has built what they call "the Big Dish", a 24 metre parabolic reflector that also uses steam in a similar converted diesel engine. It was this system that I saw when I visited their lab in Canberra. Dr. Ken Inall showed me around and gave me more details than I could understand on the workings of the system. (I wish I knew more about engines!) In one sentence, Inall regards the project as success since they proved their point, but he thinks the future is in high concentration photovoltaic systems.

For those who want the details of the of White Cliffs Project, it is possible to buy a 242-page report for AUS$50 ($25 for the report, $25 for shipping) from the Publications Branch of the Department of Energy of the New South Wales government in Sydney. The report contains photographs, schematics, blueprints, operating procedures, tables and tables of data, and all of the references you could possibly want. There is also a fair amount of dry humour in the report that makes it quite interesting (dare I say fun?) to read. I highly recommend it to anyone interested in setting up their own solar thermal system. The NSW DOE has given me permission to reproduce one chapter of the report describing the converted diesel-steam engine.

Ask for "White Cliffs Solar Thermal Power Station" (Design, operation, and results. 242 pages, 1991, AUS$25)

For mail orders, post to:

Department of Energy
PO Box 536
St. Leonards NSW 2065

For fax orders, call: +61 2 9901 8247

For phone orders, call: +61 2 9901 8269

You can pay with a cheque or major credit card.

UPDATE July 13, 1998: I recently discovered that the NSW DOE website has a publication list on it with an ordering form and a short description of the White Cliffs report. Check it out.

UPDATE February 3, 2008: The above info is no longer valid. According to Buddy Bramwell (thanks for the info!), the report is out of print but available in electronic format. You can contact Mal Williams of the NSW Department of Water and Energy (mal.williams AT and/or visit their website at

UPDATE May 9, 2011: After having received dozens of requests over the years for a full version of this report, the NSW government has finally scanned it and published it online! And just in case they decide to reorganise again, here is my local copy.

11. What microcontroller are you going to use?

After posting a question on comp.robotics.misc about factors affecting DC motor lifetime , I got several messages from people wanting to know what microcontroller was going to be used to control the motors of my heliostats.

The answer is that I have no definitive plans, yet. My initial desire was to use one of the PIC series, but I am also considering the idea of cast-off x86 motherboards since it is possible to get them for next to nothing. I still haven't decided if I want to spend the time configuring a bunch of motherboards and supplying them all with 4 levels of current, or if I should buy PICs.

PICs are nice due to their low power requirements and big public domain software repertoire for robotics, but I find that they are a bit expensive considering how little power and memory they have. At least they use less current...

People have proposed Z80s to me but they need more support chips than PICs.

The final configuration will probably be something like sets of 8 PIC motor controllers controlled by one common x86 motherboard for doing all of the floating point calculations. Each PIC motor controller will control 8 arrays of two motors each, therefore each motor controller needs to be capable of controlling 16 DC motors in forward and reverse directions. I am contemplating shared current-controlling circuitry in order to run the motors at high speed if there is an emergency requiring rapid defocusing of the arrays.

UPDATE May 15, 2000: Last year I bought one of the first StrongARM-based uCsimm Linux micro-controller modules. My plan was to use this low-powered computer as a master controller for groups of 16 or 32 heliostat arrays, but I still haven't gotten around to assembling it.

12. What kind of heat engine or receiver are you going to use?

Oof! This is a big problem because there is nothing on the market (that I know of) that would be suitable for generating electricity and I don't have the resources or knowledge to design and build something from scratch.

I'd love to use one of those experimental Stirling engines that get wonderful efficiencies and run forever, but they don't exist except as one-of-a-kind works of art or vapourware. The steam engine that I saw in Australia was great, but again, it is not on the market. I've heard a rumour of a company in Melbourne that is supposed to be commercialising a high concentration photovoltaic receiver for parabolic dishes this year (1998) but I haven't heard anything more about it. Chances are that they will have their own concentrator system that they'll be trying to push, so they my not sell any receivers to me. (Like BP Solar and their "turn-key" linear concentration system.)

Luckily, I have a plan B while I wait for someone to offer reasonably priced receivers (or I get the money to built them myself). I plan to sell a few hundred square meters of heliostats to anyone interested in doing experiments using solar energy. Let them figure out the best receiver to use! Since the price I'll be offering will be extremely reasonable, it should be relatively easy to convince people who are more interested in high-temperature processes than mechanics, to buy a ready-made, configurable, trouble-free system. This gets me back to the problem of clients, marketing, and all that jazz, but the kind of people who might be my clients are usually smart enough to be sold on tech specs and price alone. I hope.

13. How big is your workshop?

Just some space in the hallway of our apartment, about 4 by 2 metres. I've been able to get by on very simple equipment until now. Just an old 486 running Linux, some hand tools, a tiny work bench on wheels, and a set of miniature electric tools for plastic work. I've recently bought a cheap flatbed scanner, which accounts for all of the new photos that you see on this website. Although it would be nice to have a fully equipped machine shop to make whatever I want, the lack of tools forces me to concentrate on design, which is probably a good thing.

14. What is the status of the project?

Moribund. I have kind of given up on the project except as an occasional hobby ever since I got into the computer consulting business since it is relatively easier to make money. Having a baby daughter eats up a certain amount of time too. Every now and then I get an email or a phone call from someone who wants to buy some heliostats and I have to tell them that there was never anything more substantial than a pre-alpha-level plastic prototype. Maybe one of these days when I am bored of computers, I will get back to this project.

(2002-04-12) Recently I have become interested in redoing and expanding the heliostat simulator as a project in Forth. One of the things that stopped me from continuing with the project was the lack of a viable, inexpensive CAD system for GNU/Linux. That CAD system still does not exist (maybe VariCAD would be sufficient), but there was still much simulation and cost-calculation work that had to be done before doing plastic mold design.


Erik Rossen
Chemin de la Crétaux 9
CH-1196 Gland
Tel: +41 78 617 72 83

Revision dates of FAQ


Erik Rossen <>
OpenPGP key: 2935D0B9
Tel: +41 78 617 72 83
Home URL:

Copyright © 2000 until the heat-death of the Universe (thanks, Mickey!), by Erik Rossen
Last modified: 2016-02-07T12:43:24+0100