25 thoughts on “A Space Telescope”

  1. Problems problems problems. 2 mil Mylar is about 1000 times thicker than the required surface accuracy of an optical telescope mirror, so they Mylar has to be extruded and rolled with optical precision. Plastic has memory. When you fold it and then unfold it, it remembers the fold. So far, nobody has been able to get plastic films to work well for optical lenses even on a small scale in lab setups with elaborate control systems.

    A liquid mirror avoids those pitfalls, but I don’t think a very thin liquid film on top of a thin plastic film would work because a small deviation (dip) in the plastic would hold a thicker and heavier amount of liquid, which would further distort the plastic, which would fill up with more liquid, leaving other areas of the plastic dry. And then you get into the undamped oscillations of large fabrics, etc. But it would be a fun cubesat experiment.

    On an unrelated astronomical note, some think a neutron star is perfectly smooth to optical wavelengths, due to the immense gravitational forces. If they happen to be reflective at optical wavelengths (does visible light reflect off neutrons? Is there a sea of charged particles floating above the surface?) then they should be like perfect mirrored balls.

    If you could build a truly giant telescope that could clearly image a neutron star outside the galactic plane, you’d be seeing an image as seen from a mirrored ball outside that plane, which would allow you to reconstruct what our galaxy looks like when seen from that vantage point, all without leaving the comfort of our own solar system (which inconveniently would be filled with a giant telescope looking at a distant neutron star).

    1. You’d need a cold neutron star, not the kind we’ve found surrounded by accretion disks and spheres of hot gasses plasmas and jets. That could be a bit of a challenge as well. Maybe if you could find one as part of a binary system with an active star that isn’t leaking into it? So let’s see we need physics of motion, semi exotic materials and liquids and the rare cold neutron star. Yep pulling all this together on a cost benefit basis should put the JWST to shame. ^_^

      1. I suppose that a binary star would be the only way to detect a cold distant neutron star, because the companion would have a wobble that would show up as a spectral shift.

        So, you have a neutron star about 1000 light years outside the galactic plane, and about 20 km in diameter, and you want to image its surface in at least 4K resolution, which would mean each pixel sees about 5 meters of it. Dawes limit says we could do that with a primary mirror that’s just 410,000 miles in diameter. It’s the obvious follow-up to the JWST.

        Meanwhile, around the neutron star’s companion, some guy named G’dorf is snapping a photo of the Milky Way with his smart phone, and his wife says “You know, way down in that middle arm there’s a planet that is about to spend 180 quadrillion dollars to build a giant telescope to take the same picture you just took.”

        1. Since there are no cold (and black) “white” dwarfs — too little time in the 13.8 billion years since the big bang for any to have even begun to cool off — how can there be any cold neutron stars?

  2. It won’t work worth s damn with high altitude balloon, because “Since a balloon moves with the wind, the payload would feel no wind,” is utter nonsense. A balloon may drift with the wind, but there will be turbulence and Bob can’t show a single high altitude video that isn’t bouncing around like a chihuahua on crack.

  3. Bob’s making it way too hard. Very large aperture telescopes are an active field of development, and they are neither reflector nor refractor. They are diffractors, with objective optics consisting of thin metalized mylar or even thin metal foil, with either Fresnel zone (concentric circles) patterns or fractal zone patterns etched through the material. There is no optical material to deal with, and, in fact, these objectives are able to image from infrared through X-ray wavelengths. They don’t need to be precisely flat to produce high resolution, high-contrast images. The plastic or metal foil can be folded into a compact space, then deployed by any number of methods – my favorite is having a thin-walled plastic tube at the periphery of the objective, which is then inflated to stretch the objective to its final shape. If the telescope is to be one integrated structure, instead of two free-flyers, a truss of thin-walled inflatable tubes can also be used.

    1. The diffractive primary element is a _good_ thing, because if you put a lens array ahead of the sensor, you can use plenoptic deconvolution to reconstruct the ray traces, giving hyperspectral imaging while not wasting any photons. Of course, you have to trade subimage pixel count against main image resolution- if you have 100 spectral bands, the x-y resolution is reduced 10x. But at large scale, you can add one hell of a lot of sensors to capture billions of rays to make up the difference.

      The sub lens array even allows the gaps between sensors to be hidden by the sublenses, no gaps in the output. In the extreme, each sublens is slightly larger than the entire sensor behind it, and the diffractive telescope’s image plane would have millions of these. I’m thinking of an image plane with millions of tiny side-by-side smartphone cameras, capturing trillions of rays. Gigapixel hypercubes with 1nm spectra across a range from 300 to 1400 nm. A terabyte of date per image, glorious.

      On a smaller scale, the Vera Rubin telescope’s camera has an array of 189 16 megapixel sensors plus a Rube Goldberg filter swapping mechanism and three huge lenses for field flattening, the cost was rather literally astronomical. A plenoptic array would need only one (intentionally chromatic) lens, no filters, have no edge gaps, and waste no photons. The 45000 individual 3mm subcameras could be positioned along a curved focal plane without need for field flattening. The ray calculations can remove all the effects of coma, residual field curvature, and subcamera defects while producing full spectra for every pixel.

      Damn, I need to write up a white paper.

        1. It turns out somebody else thought of it years ago, it just makes sense. For astronomy, most objects are point sources and straightforward to deconvolve, too.


          Just for comparison, the huge CCDs for the Vera Rubin Telescope have very large pixels, 10 microns square. The pixels in a mini industrial CCD can be as small as 1 um, and the more the merrier- by shrinking the pixels in the subcameras, the deliberately defocused main image gets spread across tens of the subcameras and you can get excellent spectral and image resolution as the same time. It just requires 100x as many sensor elements and one hell of a lot of processing power. A side benefit is that cosmic ray flashes don’t matter- the real signal is inherently blurry at the 300 billion sensor elements, while the cosmic rays make small numbers of pixels light up sharply. This would be easily removed in the processing pipeline and the dual 15-second exposures would not be needed for CR rejection, potentially doubling throughput. The tiny pixels are also smaller targets, so while the same number of pixels would be trashed by cosmic rays, they would be a smaller fraction of the whole.

          And oh by the way, every image would have a hundred bands of spectral data. New telescopes in the future could have truly magic capabilities.

      1. White paper?

        Why don’t you talk to some NSF directors and send in a full proposal?

        1. Honestly I’m just a well-informed layman in optical design, and while I have a large folder of papers on plenoptics, I still can’t quite grasp them completely. Maybe I can write it up well enough to get someone who knows more than I do to fill in the details.

  4. Instead of using thrust to provide the axial force, why not place a relatively rigid grid below the reflector, and give them opposite charges? The exact shape could be adjusted by controlling the voltage on small sections of grid.

    1. One of the commenters on Zubrin’s article linked a paper on that very thing

      Stretched Membrane with Electrostatic Curvature (SMEC) Mirrors for Extremely Large Space Telescopes

  5. While we’re on the subject of big telescopes, a few months ago I wondered about building large radio telescopes in a different way.

    Instead of the conventional dish on an alti-azimuth mount on a pedestal base, which concentrates the forces at the central mechanism, build the whole structure as a big stove-pipe joint, just as you see on hard-shell space suits. Basically, it’s a space suit knee or elbow joint with one open side sitting on the ground, where it can rotate in azimuth, several rotating wedge sections, and a dish plugging the other end.

    All the loads are transferred to the outer skin, which is the only real structure it needs. The normal center of mass of a uniform hollow wedge is exactly one-half the radius, directly toward the biggest side. If you a fixed counterweight into the thin side of the wedge, handing down inside and beneath it, you can shift the wedge’s center of mass to the center of its lower face. That makes it’s center of mass location rotation insensitive, so torques and loads don’t vary with rotation angle.

    If you start this balancing at the top wedge, then it becomes just another fixed mass relative to the wedge beneath it, and that mass can be balanced out by the lower wedge’s counterweight, on down through the structure. And you can optimize each wedge for overall cost, with perhaps the top wedge being aluminum and the bottom one being thin cast concrete.

    But there’s more! That would create a load-bearing skin with a hollow interior, except near the inside edges where you’d need free clearance for the wedge-counterweights. That’s still a lot of empty space inside, so you build another one inside the first one, with thinner wedge segments that have short straight sections, so it duplicates the outer geometry. And inside that second one you add a third one, etc., all with common rotation axes.

    This nested pattern evolves into a solid structure, such as you might carve out of giant foam wedges, and allows the weight of the dish to be transferred continuously and evenly down to the base. Or perhaps you could make the whole thing out of giant airbags with internal straps, with your upper frequency limit determined by how wobbly and stretchy it is.

    You still have the feed horn supports to deal with, but the structure seems a lot easier and cheaper to build than going the conventional route, as long as you come up with some pretty nifty DIY edge bearings and don’t let government contractors design and build it on a cost-plus basis.

    Or, taking a page from Doug’s visible-spectrum idea, get rid of the dish and horn and replace them with an array of refractive or diffractive set of microwave lenses, or perhaps a large short-focus Fresnel zone plate antenna, which would focus inside the hollow stove-pipe support structure.

      1. One of the key phrases there is “chronically underfunded.” I just take that as a given design constraint and roll with it. It sounds like their project could have kept operating if it had been made with concrete instead of dirt, as it basically eroded away. It sounds like it really wasn’t useful in the long term because of its rather limited resolution, so other instruments superseded it.

        I think the stove-pipe method has a plenty of potential advantages. Supporting a dish by the outer edge is structurally stronger and more rigid than supporting it from the center. Wind loads should be much less of a problem because it’s more like a wall or roof than a sail.

        If you go with a solid dish instead of a mesh, any underlying support structures become indoor instead of outdoor, massively decreasing the ongoing cost of upkeep and repainting. You could use lower cost bulk materials like precast concrete instead of steel. The bearing loads are widely distributed instead of concentrated. If you go with a large hollow version, you can stick tennis courts or exercise machines inside and generate side revenue.

        You could also just use two short, thin wedges, allowing a large dish to steer five or ten or twenty degrees from the zenith, probably without costing much more than a truly fixed dish of similar size.

        I think the big question on feasibility is whether you can make the required bearings, which need to be large, potentially high load, and low friction. For the bottom bearing, could you use other tricks such as floating fluid bearings, an air bearing with a large surface area (basically a concrete pad rotating on top of another concrete pad), or having the bottom floating in a circular moat, or even have it built as a circular ship in a pool? Could you build the whole thing out of ice and build it in the Arctic?

        But it also might answer the question of how to best build your own backyard dish out of cement and plywood while passing it off as a storage shed to the local HOA. It could at least create some interesting student design studies for a civil or mechanical engineering class.

    1. I got an even better idea.

      Find a natural hollow in some mountains, somewhere, say roughly near the Earth’s equator to get more sky coverage, excavate a bowl and line it with metal mesh as a radio reflector and position the receiver above that bowl using strong, steel cables?

      Maybe we could get NSF to fund it?

      1. You could do that, but what I’d suggest is building a similar structure on a large concrete or steel pad. Such a dish naturally sweeps in right ascension due to the Earth’s rotation, but is quite limited in declination. So to fix that, we put the giant fixed dish on a wide swath of railroad tracks (like you’d normally only see in a major switching yard), and run those from the Panama Canal to northern Canada. That would allow the giant dish to scan much more of the sky over its operational lifetime. Now if the tracks were, say, all high speed rail, the dish could move much faster in declination…

        As an aside, I just finished watching General John Raymond on C-Span. The interview ran from 6:30 to 7:30 Eastern, and was quite informative. It’s not posted on C-Span’s website yet, though.

        1. I guess our respective snarky jokes got crossed someplace?

          The NSF funded such a thing at Arecibo, Puerto Rico, but they let it collapse in a clattering, smashed up heap when the cables wore out.

  6. Could such a mirror be used to concentrate sunlight? I can think of several handy applications for that…

    1. It wouldn’t need to be optically perfect in that case. Just enough to focus sunlight on a boiler.

Comments are closed.