Telescope Reviews: Mirror support on a machine

Rob, sadly this isn't true, though it's a very small amount, the contact is far from as perfect as one could wish for.

You probably know well enough what Josephson (?) blocks are, and that they actually stick together through molecular force. Neither the mirror back nor the table are ever going to be that flat on a micro/macro level.

The whole point of the flotation is that, no matter how flat you have the table or the mirror back, they are only guaranteed to make physical contact in 3 points, unless one or the other deforms. And it isn't going to be the table.

So the mirror sags onto the table when you're talking about wavelengths of light (kind of like a pitch lap would if you laid it face down). And every time you pick it up and put it down again, it's going to take on a different shape, and when you polish that and then pick it up off the table, you don't have a good sphere (or whatever), you have the shape of the table (in reverse) + the shape polished onto the mirror. You need the flotation to fill the gaps evenly.

Note that I'm talking about polishing, not about grinding. I doubt that it makes much difference in grinding, but I use the flotation for both, so when I'm done with grinding I just hose the table off and it's ready to go for polishing, no adjustments.

There's always two sides of course - on my generator I put the blank straight on the flat table and the diamond cutter just chews away at it with what I suspect is hundreds of lbs of pressure (sometimes). But since that blank is cleated down like heck and can't go anywhere, and I want the same rough curve ground in, it comes out even all around (like a planer would), to within the tolerance needed (a few thou on the edge thickness).

Every step along the way from generating to figuring the tolerances you need to hold the parts to get smaller, that's about the size of it.

John would know more about the first part of it...

You don't need consistent (or any) rotation of the mirror on the table on a machine, if all of the following is true:

Table is really flat
Table is really trued
Workpiece is centered, or, if not restrained, centered on average
Workpiece is on good flotation

Some people do practice regular rotation of the substrate on the table, even swearing by it, but it's only useful if one of the things in the list isn't true. If you do it, I think it has to be consistent and indexed physically, and you have to give equal time to all the different positions over the course of any work session, and it can't be done just randomly unless you're consistently lucky.

But you don't need randomness there (from grinding thorough polishing out), you just need predictability.

If you're using spin grinding/polishing (fixed quill position) you need the tool to be held at the neutral spot (holds the ROC constant) and you need a faceting pattern on the polisher that doesn't imprint on the mirror.

If you're using strokes (SFAIK!) you need a pattern that doesn't repeat (rotation speed and stroke speed should be incommensurate, ie, not divide evenly into one another), and the average position of the stroke work needs to hold the ROC constant.

When you get to figuring an argument can be made for randomness. It will help to disguise machine artifacts, SFAIK. But I figure everything completely by hand, and I'm not trying to be random when doing it, though I'm not trying not to be random either - there are inevitable small variations. I don't think they are part of the essential process.

I just put that in the list, it's hard to do considering that you're always going to line the blank up in the generator somehow, to within a fractional mm, and that's all you need. Rough grinding centers the curve automatically...

Not often and only for small mirrors, anyway. The odds of getting a perfect sphere by simple chance are tiny in larger sizes. But I did once pull a 1/6th wave paraboloid off the polisher by "chance." Not quite good enough to use, though I had to give the polisher big props for making it automatically.

Huh? The table rotates on my generator with the blank locked down on it. I added an extension table to it once and didn't check it carefully enough - it turned out a few astigmatic generations before it was caught. Now it has a really flat 5/8" aluminum table running on 3 big casters, it's scary how accurate it is.

Hogging out by machine takes an even amount off all around, so if there was a lot of wedge in the blank when you start, there's still some left when you're done. It shouldn't cause it though, rotation or not.

Grinding a mirror with wedge, or generating a blank with wedge if you don't clean it up - both produce a shape that when polished by machine shows astigmatism, as the polisher hits the high zones harder. You can polish it by hand without any issues though.

It averages about 1/2 lb per square inch added to the tool weight, but since the tools are tile/glass embedded in Hydrostone they aren't 100% grinding surfaces. So I'm guessing anywhere from 1/2-1 lb per square inch of grinding surface, like 25 lbs for an 8.5" tool, or 50 lbs for a 12" tool (as for a 20" mirror). These are both grinding and polishing out loads. Figuring is about half that much, 10 lbs on a 7" polisher, for example. Or less, depending.

If you want the explanation that will take a lot longer.

It's just what works best for me.

Best,
Mark