I'm uneasy about wire-trusses. You could get z trusses with rigid cross members on 2 of the 3 faces, the ones that go from either angle iron mirror edges toward the center collector tube. Since the focused reflected rays go along this side, all we have to worry about is letting the unreflected incoming parallel rays through. For this we could make a rigid member with near zero shade from flat pieces, such as two 1 inch wide strips facing parallel to each other periodically pillared by say half inch long 1 inch wide flat pieces. The orientation of these pieces should account for the latitude +/-23 degrees if the paraboloid is not tilted, using the most average angle.
For the side between the two angle irons across the mirror you could keep using wires to reinforce in tension, in an x truss way, or even if you skip trussing this side, the other two sides made in a rigid tension-compression z truss plus the two end triangles make quite a sturdy shape.

Where the collector tube casts a shade in the center you could make a z truss long flat piece edgewise, from which to anchor the focus fine tuning adjustment screws. You could also make a few holes here and there to let the rainwater out, if it happens to rain at noon when the mirrors are not tilted to either side.

Also instead of welding the collector tube in place, compression fittings might work well with water and oil, probably not so much with NaK temperatures.

I was thinking how we tend not to use trusses much. You can still see them in high voltage power line posts or some bridge designs, but, for instance windmills and airplanes don't use them much even if they are lightweight, at least not as much as they used to. Even bridges used to use them more.

Another thing. As you scale up the size of the paraboloid dish, say double it, to maintain the same focal accuracy you need to double the size of the collector tube, but that means its volume is going up as the square. Instead of using large tubes, you could use heatsink fins on smaller tubes, to increase the visible area. If the tube rotates with the paraboloid then only 2 fins on each side are enough, but if the tube is stationary and the mirror rotates around it, as the best setup is, then you probably need 3, 4, 5, or 6 fins, whatever is most cost effective.

I just thought of something - if the ribs are straight running from the center of the paraboloid toward the edge, it's easy to just bend the sheet material without much stretching required from the surrounding area, but if you want to imprint figure 8 or ununun, there is a lot of stretching requirement from the material and it may not be able to withstand it. The truss may handle the flexural modulus along the tube length, though it's true that easy flexing in that direction makes the mirror panels easier to pop out from the slots by wind and not as sturdy. Still thinking about this. For one, coke cans undergo a whole lot of stretching.

One way to add ribs to the backside of the mirror is welding strips of metal edgewise, but a better way may be corrugating via a knife-edge or a very astute v-shape to create the same effect, from the mirror material itself. This way you could have a good flexural modulus that holds the paraboloid shape, and the mirror surfaces would be level, with just a very small area missing where the rib indentation happens. The ribs should not be straight running toward the edge, but curved, and one option is indenting figure 8s into the panel but without the lines meeting at the waist, but more like ununun, in an curved 8 shape way.

For the truss a good solution would be having equilateral triangles formed by the collector tubing one vertex, and two other similar tubings running along the edge of the mirrors for the other 2 vertices. The larger the base of a truss the more sturdy it is and this is the largest base you can economically form in the given scenario. For the diagonal segments of the truss, because you don't want to shade too much mirror surface, you would only reinforce in tension with thin stainless wires, which means a z shape truss surface is not enough and it has to be x shaped, kind of like the very first airplanes had wire reinforced trusses for light weight.

For small scale installations you could use angle iron for the beams running along the edge of the mirrors, with slots to flex and snap in the corrugated mirror pieces, which would be flimsy enough even with the back-rib corrugations to flex enough to snap into place. If the center of the mirror is free floating then you could have and adjustment screw pushing or pulling on it to fine-tune the focal point, which you could check with a sheet of paper, knowing the image say 2 inches from the focus has to be 2 inches wide, 3 inches away 3 inch wide or something along these lines. For small scale installations such as backyards you could have the whole thing tilted 30 or 45 degrees, and the total length would be limited by how high you're willing to go on the north end, say 10 feet is manageable with a 10 foot ladder. You could have Home Depot sold cement blocks with threaded rebar sticking up out of them, and 3 legs for the whole installation would be stable even on lawn grass. Now for wind resistance you would want the mirror panels to be flimsy enough to where they pop out right before the wind tips over the contraption, lifting one of the cement blocks. After a big storm you'd find your mirror pieces laying on the ground somewhere.

The problem with small scale installations is that they are small scale, and for meaningful solar power you need to deal with very large areas, something that may not be meaningfully available in most urban backyards. For instance heating a home on average in the US in colder climates like Chicago for a 2000 sq ft home you need a 100,000 Btu/hr furnace, which is equivalent to about 30 kW. Solar irradiation facing the sun nonangled is about 1 kW per square meter, so you would need probably 60 square meters of area to account for all the losses and inefficiencies, which is 6 m x 10 m or about 18 ft x 31 ft, which is a lot of tilted area facing the sun. And this heating would only substitute the gas furnace while the sun is out, not during the night, for which you would need geothermal so double the area again or even more, and the whole thing would not even work on cloudy days, where photovoltaic still outputs about 5% of the full sun power, but parabolic solar thermal would be near zero. Though 30kW of heat converted via a Stirling or steam engine might be able to provide 5-10 kW electric, which is more than decent, and you could send it to the grid, and the whole rig may or may not be cheaper than photovoltaic.

I meditated some more about this today and the dish way works, it would even work with dimples in a flat surface, think of Fresnel mirrors, so the dish way works because a dish can be made to always directly face the sun and receive perpendicular rays, but the parabolic trough does not work mathematically exactly when the solar rays are coming in at an angle, which is how most parabolic troughs operate. In an east west orientation in the morning and evening the angles are very acute. A south north orientation can follow the suns path as the day goes by, but the sun is higher or lower above the horizon as the seasons pass. The Tropic of Cancer and Tropic of Capricorn are at a latitude of +- 23 degrees, and the sun travels between them as the seasons pass. For the continental US El Paso is 31 degrees latitude, Duluth MN 47 and Chicago near the corn belt at 42. To face the sun perpendicularly for a north south oriented parabolic trough the north end would have to be lifted so the whole thing is at an angle with the horizontal exactly the latitude angle. This is not done, it would be awkward, plus there is the +-23 degree=46 total fluctuation as the seasons pass, so in Chicago at the height of the summer the angle would be 42-23=19 degrees, and in the winter solstice 42+23=65 degrees, to face the sun perpendicular. This is not done and the troughs sit horizontal, level with the ground and the sun rays come in at an angle, and they focus on the collector tube no matter what the angle is as long as the parabola part is vertical against the rays. However if you corrugate and angle some portions of the mirror 30 degrees, then that focal line is not along the collector tube, but along a line intersecting the collector tube at 30 degrees, and perpendicular to the plane of the angled parabola. Now as the sun moves up and down with the seasons the focus of the rays follows these focal lines offset by 30 degrees, and if the deviation is large the reflection be out of focus from the collector tube. There is some wiggle room, as the collector tube is not a pure infinitesimal line but it has width, and this allows rays to be out of focus in front of or behind the actual focal point, near the focal point the rays only diverge so much if the distance is not great. So the math is not exact for a corrugated mirror with paraboloid corrugations when the sun is not coming in perpendicular but shifts position with the seasons. But corrugation can still be worth it, especially if you do only a little bit of corrugation, say a quarter inch deep then have flat paraboloid pieces for say 6 inches, repeat, even if you totally lose the rays from the corrugated part, it's like wasting 5% of the incoming sunrays to gain 200% improvement in flexural modulus or corresponding decrease in material thickness and capital cost. There are of course other ways to gain strength too, corrugation itself consumes extra material, which could be used to reinforce by the way of ribs on the backside. My original idea was to use U shaped snap-in skeletal sheet holders that would guide and bend flat sheets of metal into a paraboloid shape, would be easy to assemble and disassemble in someone's backyard, sit on a 3 legged frame, and would be easy to transport. But the sheet thickness would still have to be significant and pretty sturdy because you can only place the holders so dense, and a flimsy piece of sheet would easily get bent by the wind, or pulled out from the snap in location. A corrugated mirror structure is much more densely reinforcing and would only need a central truss- spine from which the corrugated mirror would branch out sideways. You would still need the holders but not as often, they would only be needed to hold the collector tube at the right distance. By the way I think in a proper setup the collector tube never moves but is welded in place and fixed, and the mirrors adapt and dance around it as needed to focus the rays.

Today I saw a fluorescent tube fixture panel with 4 tube-slots, and it had individual reflector zones, making the mirrored surface look like it had 3 giant corrugations. This gave me the idea of corrugated parabolic trough mirrors that would have better flexural modulus allowing for thinner sheets of metal to be used, which would be important in a 100 mile x 100 mile area of solar that would generate the equivalent of US energy consumption. Aluminum is too weak and the ideal cheap material in my opinion would be aluminum clad carbon steel sheets, aluminum providing the corrosion resistance and mirror surface. You would stretch carbon steel coils into very thin sheets, then pass them through a molten aluminum bath under argon atmosphere, argon constantly leaking out at the inlet and exit ports from the bath, like argon constantly is blown during arc welding, then one side would be mirror polished, the sheet cut into proper sizes for a corrugation mold press and the paraboloid corrugation imprinted, then the whole thing would be anodized for increased wind blown sand resistance and reflectivity.

The corrugation would be like thus: Imagine cutting a parabolic trough into thin slices across its width, you would have say 1 inch wide U shaped paraboloid strips. Now when you tilt them say 30 degrees to the left and the other to the right, they would touch near the center, the bottom of the U and diverge at the edges, at the top of the U. Because of the focal point being in the tilted plane of the parabola, still at the same distance at 30 degrees offset, the parabolic trough tube would have to come in closer by the cosine of 30 degrees which is 1/2*sqrt(3)=0.866 the original distance. To fill the gap between the divergent edges near the top of the U, you would need triangular flat shapes, but with a parabola that has this same focal length so it focuses on the same tube. Same for the gaps between the bottoms of the Us.

While this would improve the flexural modulus toward the width of the parabolic trough, lengthwise, because of the accordion shape it may not be that strong and may need a rigid truss based backbone running along the length.

A similar idea could be applied to concentrator dishes. You would put dimples into the dish that would each have a closer focal length, but you would orient them so that they all focus on the same point. In between the circular dimples that look like the bottom of a coke can, or make the dish look like a golf ball with dimples, in between these dimples you would have undimpled flat spaces with the original focal point, but to avoid having 2 different focal points you could micro dimple these regions with the same curvature as the main dimples, just smaller surface diameter, and then so on, the spaces between the micro dimples with even more nano dimples, like a fractal. This way you would have a dish with much stronger flexural modulus allowing much thinner sheets to be used in its construction.

A google search comes up with corrugated tubes but not corrugated mirrors.

Profit seeking behavior that private enterprise is good at does not necessarily serve the public interest, contrary to Adam Smith, it serves the self interest at the expense of public interest. Case in point is Pharma Bro. Someone who serves public interest drops prices not increases them. In a low risk low profit margin environment such as solar is you may have problem finding private investors, unlike in the high risk high profit nuclear environment where the risk when it comes due the private enterprise can just walk away and the public bears the consequences. In the low risk solar environment the government might have to fund and run the whole thing, similarly to how the Chinese do it with their nonprofit high speed rail system, building bridges, and solar farms. Government in the US also funds lots of basic research that's nobody's self interest because it's low return, low profit, but in the long run it serves public interest. Government run enterprise can be a terrible idea as the Soviet Union collapsed primarily because of economics, it can be high bureaucracy, lots of red tape, and low economic efficiency. In the US there are other low profit private enterprises, such as some farming, but nobody wants to live in the desert. Ultimately it could be the army core of engineers building up and running solar power in deserts, the military can be relatively efficient compared to other type of government bureaucracy. I heard you need about 100 mile x 100 mile amount of solar to fully supply all of the US's energy needs, haven't done the calculation myself. But we've certainly got the land for it.

Crescent Dunes Solar Energy Project wiki
A billion dollars for 0.1 GW plant that has issues running was an economic failure. But once that capital was sunk might as well run it now. It's like no amount of capital is too much if the interest rate is low enough, but interest on a billion dollars is hard to keep up with via low profitability solar unless the interest rate is really really low.

The wiki breeder reactor page explains better how to extend nuclear fuel reserves. It does say that there is more nuclear fuel in the world than once thought, but if you do the calculations for a 20 terawatt world, out of which the USA is 3 terawatt, you'll see that the reserves don't last very long. All reactors do some kind of breeding, with most conventional light water reactors having a conversion ratio of 0.6, and heavy water 0.8. A reactor becomes a breeder when conversion ratio is higher than 1, when it produces more fissile material than it consumes, by breeding from fertile material. Fertile materials are Th232 and U238. The only significant fissile natural material is U235, everything starts with that, but once you get a breeder going, you don't need any more U235, instead the bred U233 and Pu239 can be used as the fissile materials. All breeders require reprocessing, which is currently against the law in the US, but other countries, like France, have no problem with it. Solar power has no proliferation issues, but it's lower profitability, nuclear is higher profit but higher risk, and when the risk comes due, like in Chernobyl or Fukushima, it's not very nice, and those were not even worst case scenarios, worst case is say Houthis or other terrorists get their hands on a nuke. WIth solar power it's a more peaceful world, if you can only make the economics work.

Oops, if a minimum critical mass is 80-100 ktons using U233, then a 600 kiloton bomb is 6x this much, meaning 9000-12000 years worth of fuel at 20 terawatts. All U235 burning scenarios seem to only be able to burn a roughly equivalent amount of U238, whether it's a 5% enriched reactor getting a 10% burnup, a 25% enriched fast reactor getting a 50% burnup, or a 300kton U235 core burning U238 at 600 kilotons and minimal fusion, it's all only roughly equivalent amount of U238. I'm assuming a natural uranium burning heavy water CANDU can only get 1.5% burnup at max, and greatly increased nuclear waste amount unless it's reprocessed. But if there is plenty of U233 then burning up equivalent amount of U238 is no problem.

One of the great things about thorium derived U233 is that it is like U235 enriched to 100%, and instead of just burning it by itself it can be used to burn the otherwise unburnable U238. A conventional reactor has about 5% enriched U235 from the 0.72% natural, and can achieve a burnup of about 5-10% max, and the remaining 90-95% spent nuclear fuel is mostly U238, depleted uranium, unusable without any further U235. Or U233. To burn U238 there are 2 ways - one fission by fast neutrons in a fast reactor where its cross section is not negligible, and the other is converted to Pu239 in a conventional moderated reactor, which is then fissile just like U233, U235, Pu239, Pu241, Pu243. Pu240 and 242 are not fissile but should be fissionable in a fast reactor too. Th232 is also fissionable in a fast reactor. The Europeans/French reprocess spent nuclear fuel for the Pu239 and reburn it in conventional reactors, but the Superphenix trying to be a fast reactor had control issues. The Russians seems to successfully operate 2 sodium cooled fast reactors and Terrapower in Wyoming also wants to be a sodium cooled solid core fast reactor. If they can figure out reaction control issues.

So basically you need a fast core, such as a liquid core chloride mix of, say, 30% U233 / 70% U238 boiling Zn70(Cl37)2, which is surrounded by graphite moderator walls and leaks slow neutrons into a molten fluoride blanket that contains ThF4, from which either Pa233 or U233 are extracted. ThCl4 never comes into play. Also extra U233 stock piles could be quickly burned in a Geocore not just burning U233 but using it to burn fusion Li6D lithium deuteride with beryllium reflectors, the question being how big a bomb can you contain, how large a chamber do you need/can you build. So instead of having 3-400 times the practical reserves, you could fully consume 100% of uranium reserves in a fast reactor, making it 4-500 times the practical reserves, and if you can make a bomb based geocore work, 100kilotons is a U233 core that could burn a 200 kiloton total fusion added bomb, or 300 or 600 or 1 megaton, the sky is the limit the only question is how big a chamber you can build. I don't think you can ever build a chamber big enough to contain a Tsar bomba, even considering compression at 160 bars 350C vapor-liquid equilibrium steam, or if such steam is fully shock absorbing, which has to be tested by experiment. If it's shock absorbing then it's only a matter of money of how big a chamber you can build, to accommodate say a 600 kiloton bomb size. Thorium and U233 would be key for this, as there is not that much reserves of U235 reserves in the world to make a significant impact in a 20 terawatt world, but the 3-400x practical reserves of thorium is a different ballgame, meaning if you have enough uranium to power the world for 5 years at 20terawatts fully on uranium, then you have 1500-2000 years worth of thorium which is 300-400x. Which can be boosted to 500x by also burning all the U238, or possibly 800x via 200 kton bombs or 1200x via 600 kiloton bombs, meaning 6000 years instead of 2000. You may also find that seawater extracted uranium is easier than terrestrial extracted thorium in the long run, in which case burning U238 via thorium derived U233 makes a bigger impact than just going from 400x practical reserves to 500x.

Terrible idea. We go to https://www.nndc.bnl.gov/endf/... and put in Target: F*, Cl*, Zn*, Th* Reaction: n,g, Quantity: sig and pick F19 one of them, I like ROSFOND-2010 for all, so F19, Cl35 Cl37, Zn70, Th232 and plot, we can unselect items and repaint to see which is which, and we can see that Cl35 is terrible in a moderated environment near the left of the graph, especially compared to Th232. F19 is best and Zn70 is second best, but chlorine isotopes would only have a place in a fast reactor. If we zoom in to 1 to 10 MeV on x axis for a fast reactor, we can see that Cl37 is slightly better than Cl35, Zn70 is close to both, and while none of them are as low as F19 the situation is not as bad as in the moderated region where Cl35 absorbance is much higher than Th232.

I just noticed that on the wiki thermochemical cycle page, that the Deacon process is reversed above 700C, giving H2O+Cl2-->2HCl+1/2O2, because of entropy, 2 moles converting to 2.5 moles, and one does not need to go through Co3O4 processes to achieve the same thing. But 700C is high temperature and the O2 and Cl2 are corrosive, and silver coatings cannot be used with Cl2 like they could be with O2 and SO3 but perhaps enameled steel or glass coated steel might work if the glass doesn't crack or melt. But then it becomes an issue of tackling HCl. One way would be similar to the copper-chlorine cycle, Cu+HCl-->CuCl+H2, CuCl--->Cu+CuCl2 low voltage high current density electrolysis, CuCl2--->CuCl+Cl2 at high temperature, but possibly higher than the copper chlorine cycle oxygen releasing temperature plus more complex. If the reverse Deacon reaction works with bromine, which it would but probably at higher temperature, the CuBr2-->CuBr+Br2 would be at even lower temperature, as CuI2 never even forms, only CuI is available, there is a tendency for monovalent halides going down the periodic table.