Another advantage of having an array of independent lens-mirror combination is that you could adjust the temperature of the absorber according to your requirement. If your requirement is low, you can take a few mirrors off the focus.

Why is this tagged as "Announcement"?

The problem with this is that we need to come up with cost estimates.  It can easily be the case that it is not so cheap after all.  Plus, if you are going to use a lens you really can't beat a fresnel lens for small amount of material usage.

Here's an idea: (stir) welding, metal spinning and aluminum sheets.  A 4 by 8 sheet of Al is only 30 bucks.  That's less than bathroom mirror and it could turn out to be easier to make.

.lookup stir welding on wikipedia. Wire welding might be just as suitable though. Basically you weld the sheets together, carefully including leveling the bead to the same thickness as the surrounding sheet.  

Then put it in a giant metal spinning machine, which can be constructed at low cost from mostly wood.  Spin the sheet to a parabola and then it can either be used right away, or it you want it disassemblable, you can then cut it into segments.

Another idea might be to use cone-segments of reflective aluminum to approximate a parabola.  An alloy like what is used in aluminum cans could be nice, it is quite durable and springy but also quite cheap (about 4 cents per can I think).  A strip of aluminum can be made into a cone-segment easily.

The segment shape I am thinking of is to segment the parabola along the axis of rotational symmetry.  The aluminum would be much like rings, then.  If you think about it, actually it may be possible to produce a parabola-cone segment- approximation from a single (big) sheet of aluminum welded together like above (the advantage being that the parabolic dish provides most of it's own structural strength if is a single piece.)  The circular line, if you will, where one cone segment meets another is the same length for each of the cone segments.

Trying this with a piece of paper confirms that the bending resistance of the material is the only impediment to making a cone-segment parabola approximation out of a single sheet (with a sort of wedge removed from the sheet of course, much like making a cone out of paper by rolling it up if you see what I mean).   The material can of course be sort of creased at the connections between the cone segments to overcome this, i.e. use a little bit of bending to get the paper to become permanently bent at the angle desired in that region.

A machine which could take the large piece of aluminum sheet and bend it into the approximation of a parabola shape could be pretty easy to make.  A bunch of rollers basically.  Cut out the wedge and put the sheet in the machine.  I wonder if large rolls of aluminum sheet are available? That could save some of the welding needed to produce the initial sheet.

I think you may be able use a spiral shaped parabolic approximation too.  That might have some manufacturing advantages but I doubt it.

Of course the aluminum would have to be polished but that is easy to do by machine in a semi automatic way. 

One potential problem I see with these lens designs is how they scale. To capture a larger cross-section of sunlight, it grows as a cubic function (x^3), whereas the parabolic systems grow proportional to the surface area they're capturing (only x^2). When trying to boil water by concentrating sunlight, a critical minimum amount must be collected for it to work, otherwise heat will escape from the focal point before the temperature can increase fast enough.

In order to determine the viability of this, it must first be determined what volume of water is necessary to create a enough of a concentration, and what the tolerance is away from a perfect sphere or cylinder.

I would be curious what it would take to melt down glass and make our own reflectors.

It would also be interesting if we could make these into smaller modules by somehow insulating the collection point. For example, what if the collector was in an evacuated glass chamber, with all surfaces silvered except for the portion receiving the reflected/focused light?

I have been thinking about the collector in evacuated glass chamber idea. How bout using a stirling engine. Paint the surface that receives the sunlight black and then enclose it in an evacuated glass chamber. That may not be possible in the regular applications of stirling engines, since the source of heat is fire. But here we are only using sunlight. If this design increases the output of the stirling engine, mirrors would not really be necessary. We can use just one big water lens with 2 meter diameter.
Or two lenses if it is really necessary, as explained in my other idea--Two lens design for a cheap solar thermal collector:-

http://openfarmtech.org/forum/discussion/196/a-two-lens-design-for-a-cheap-solar-thermal-collector

I don't believe efficiency is a concern since the fuel is free. Efficiency is only a problem if the materials used are very expensive. A water lens should be very cheap and easy to manufacture.

I like the Auroville spherical solar dish design. The dish remains static and can be set in the ground or hillside and the collector moves to stay with the sun. The focus is a line in line with the sun rays so a tubular collector in evacuated glass would get seriously hot, and the bit of the dish that gets the wind is small. A fluid would be needed to transfer to the sterling motor. Concentration of energy is more at one end than the other, so move the fluid that way.