And there's the rub ... the frame. Here's a bit more detail as to why (just no evading the mechanical engineering theory and practice, which I've been trying to avoid) >>> Transmitted wavefront, no problem; bend a pellicle a bit and the transmitted wavefront is unchanged. But the reflected wavefront catastrophically degrades. Here's why: The simplest frame is a hoop. The pellicle is necessarily in tension (potential energy per unit area); this puts the hoop in compressive stress. The system seeks its lowest energy state. This causes the hoop to "want to" flop into a figure eight and fold in half (mechanism related to why hoses kink). Other non-linear contributors don't allow it to get that far, but the result is that the hoop twists enough to manifest a terribly astigmatic reflected wavefront. The math is called ~Knot Theory~. You don't notice this instability just looking at it... but an interferometer clearly shows it.
And how we solved that... needing pellicle beamsplitters from 6-inch to 36-inch diameters... which need superb transmitted and reflected wavefronts! >>>
Make the frame of solid Fused Silica or ULE. Grind, polish, and figure to twentieth wave surface PV. Then machine out the central part where the pellicle will go. Vacuum Metalize the surface where the pellicle will attach, so it's electrically conductive. Fab/Attach the pellicle (usually takes several tries to get a good one)... Coat it, electrically ground it (to prevent electrostatic puckering), integrate it and use it. So you see, the frame far far exceeds the cost or a solid flat mirror of the same size (which the frame started out as).
Conclusion: The frame is the toughest part... far exceeding the cost of an equivalent solid mirror. Tom
Aside: Most of my training was in optics, math, philosophy, and psychology/physiology. But mechanical engineering would have been my 4th or 5th choice, so I audited a lot of grad courses (gratis, as an alumnus) and picked up a lot, hanging out with the mechanical engineers and scientists. Although telescopes are optical systems... all of the components and assemblies are mechanical builds, subject to all the vagaries of trying to craft things that hold their geometry to millionths of an inch perfection... forever! Notice that building a good optical flat is much tougher than making a sphere. That's because the radius must be exactly infinity. With spheres, the radius tolerance is generally far far more forgiving. Tom
Bit more on where the astig in the reflected wavefront of a pellicle comes from. Examine a bag of rubber bands. Notice that they naturally relax into 1, 3, 5... odd-loop potential well energy states, because the net twist is zero for all of those. The compressive stress on the pellicle frame upsets the one-loop well. The strain response is to (try to) transition to state 3, by (necessarily) passing through unstable state 2... It never gets there, but becomes severely astigmatic, as a consequence. Related topics from physics: Hamiltonian, Lagrangian. Tom
thanks for the in-depth explanation. It's really interesting, particularly the hoop wanting to flop into a figure 8, which makes sense in retrospect (thinking of folding bike tires when I went on trips).
I used to think of structural parts of machines as being hard and not really considering them much in an overall design but now I tend to just imagine them as rubber and ask myself how they're going to flex because at the tolerances in consideration they basically are rubber.
That said, I don't really have a huge constraint on the hoop part for the pellicle and since it's a 100% mirror, I can have a solid backing part and the hoop can just be a raised element on a surface, which should provide a ton of stiffness. It seems like at some point the tendency to flop must be made insignificant with enough frame material. The structure could be a made from regular plate glass, say, 20mm thick with a 1mm high ridge on top (and a hole in the back to eliminate pressure making a pressurized chamber) which the pellicle stretches over. I guess that would be stiff enough?
So, take a piece of plate glass, cut out a circular body for the pellicle backing structure, say using this simple technique https://www.instruct...y-Sized-Hole-in. then, imagining the raised 'hoop' we want to create, remove glass from outside that diameter with a silicone carbide grinding cup of appropriate diameter (like https://www.google.c...nding&tbm=isch, which I think I've seen at the flea market for a few bucks), and remove glass from the surface inside the hoop/ridge with another small SiC grinding wheel fixed parallel to the surface while the glass rotates below and the two slowly come together and the SiC wheel moves in and out radially to do the entire hoop-enclosed surface area (maybe a grinding bit on a flex extension on a Dremel, or die grinder if the Dremel is too wimpy). Maybe do those steps submerged to avoid the glass dust danger. Then polish the top of the hoop/ridge with normal grinding/polishing techniques. Aesthetically I wouldn't care if the hoop/ridge was not circular but maybe it needs to be for some reason to stretch the pellicle on without making undulations? (seems like you could stretch a pellicle flat also over a rectangular aperture with proper edge tensioning...?) I suppose all points on the upper surface need to be planar but maybe for such a small area of glass as the hoop/ridge upper surface is that could be ground out rapidly? So, ok at this point I'm at a loss for ideas how to make the hoop/ridge upper surface really flat if that is a requirement. Kind of back at square one, to make a flat glass surface (for a mirror) except that instead of having to make a 50-60mm flat mirror, it's just a ring surface ~60mm in diameter and 1-2mm wide...so ~120mm^2 to polish vs. ~3000mm^2 (for 50x60mm surface).
Or maybe the backing+hoop/ridge could be machined out of a single chunk of cast iron (but that would not longer be cost effective unless I had the machine shop, which I don't...so, just mentioning it). Or maybe just take a long-ish (for the stiffness against figure 8 flop) section of thick-walled iron pipe which is 60mm in diameter and prep one end flat and smooth to stretch the pellicle over. Maybe that would be much easier overall.
Not sure how much thermal expansion matters if one is willing to wait a little longer for thermal equilibrium, but just for reference:
0.55 Fused Quartz
5.8 grey cast iron
9.0 Plate glass
... in x 10^-6/°C:
I guess making the hoop/ridge top planar and flat could be difficult.
Edited by ToSpaceOrBust, 01 August 2020 - 07:29 AM.