This is actually a response to Gregors post http://forum.opensourceecology.org/discussion/257/this-solar-fire-steam-stuff-is-going-ahead-as-is-then

But that thread seems sidetracked, so I've started a new one in the GVCS development section, as the crux of the question is of course cost-effectiveness.

Hello everyone, I'm Eerik Wissenz and I develop Solar Fire, www.solarfire.org.
Since joining OSE, I have also been working heavily on the IT front, helping move OSE to one of the root servers I rent. Next to this I have only had time to create the basic documentation, and not so much time to follow all the forums. I was more concerned about the forum simply continuing to exist.

Also, I can be contacted directly at wissenz (at) gmail.com

Gregor wrote:

"400 bucks per peak kW sold.  I searched the wiki and couldn't find
anything on the actual materials cost - even assuming the best case
scenario of that being the power actually transferred to the steam,
it's not enough to get anywhere near $1 per peak watt as hoped.  The
whitecliffs engine alone got about 20% efficiency.  That would be
almost $2 per  peak W of electrical output for the collector alone,
right there - and that ignores, well, everything."

400 USD ber peak Kw is offered by Tinytech, so it is not a DIY price. Also, Tinytech insisted on the most conservative peak-watt estimate, as they prefer to sell something that over-performs than under-performs (they were also skeptical of the definition of peak watts as maximum production, and prefer to cite a figure the client can expect over the day, based on our calculations in Rajkot where pollution blocks about 20% of the sunlight). So they insisted on marking it at about 12 peak kilowatts. However, the machine has 32 m^2 of mirror, loses about 90% to optical losses (inclination away from the sun's rays), and with good mirrors reflection should be around 90%, which results in peak 25 Kw hitting the boiler. Of that 80% can be absorbed with a steal surface, so 20 peak Kw should be the steam production.

Tinytech is able to supply that machine (as in willing to take the contract in the here and now) at 5000 USD, with sufficient on quantity (over 10). However, instead of market the SFP32 Tinytech is so convinced by the technology that we immediately designed a 70-90 m^2 model.

So a more realistic peak rating would be 250 USD per kilowatt.

This statement can be verified by the Solar Fire Simulator: http://www.solarfire.org/Simulator-Lesson-One

Now, this is for the existing machine a company is actually willing to make, it does not factor in any DIY reductions nor any cost improvements over the SFP32 which was built entirely in steel. About half the cost for Tinytech is labour, and the rows of the machine can be replaced with wood or bamboo reducing weight (and thus strength of the supporting structure) and cost. We must keep in mind that the SFP32 was the first generation Solar Fire concentrator of that size.

Another fundamental point is that from a solar thermal system it's possible to extract high quality thermal energy, directly, or relatively high quality thermal energy from the waste heat of the steam engine. For PV the panels must remain cool so it is only possible to extract low grade heat.

Thermal energy accounts for 80-85 percent of energy used domestically and in industry, so when associated with these applications, electricity can be produced and large amounts of thermal energy. It is with these multi-uses where the machine becomes much cheaper than PV.

For instance, to make charcoal, a solar concentrator can easily do that at about 80 % as there is no mechanical loss. For PV it would be necessary to make the electricity (~15%) then convert it to heat, which would be far more expensive and take up more surface area.

I discuss the thermal problem in http://www.solarfire.org/The-Thermal-Problem