[ATM] Floppy Disk, was: New? Idea for mirror support

[mdholm at telerama.com (mdholm@telerama.com)]

> Mirrors seem to absolutely defy quantitative testing. I mean, something
> as 
> manifestly simple as a chunk of glass placed upon a support system that
> 
> endeavors to preserve its surface figure should have, by now, produced a
> consensus 
> among designers and/or users of same. And yet, it has not.
> 

Well, actually there is some consensus, but a lot of atm's don't know it, or
aren't willing to accept it.

The first consensus point is that when you are trying to control deformations to
the level of a small number of nanometers, everyday wisdom doesn't cut it. 
Nanometers are really tiny!!!  These days, very good machinists working with the
more stable and readily worked materials and the highest quality tools and
machines can hold tolerances of about 0.0001 inch.  That is 0.00254 mm which is
2540 nanometers.  If we want to hold our mirror surface to no more than 1/20
wave deformation (1/10 wave at the wavefront), that is 25 nanometers, or 100
times better than really good machinists do with optimal materials and equipment.

Furthermore, we need to hold those 25 nanometer or less tolerances over
distances up to (for atm's) 1 meter, and for the pro's, up to 10 meters.  Now
I'll bet even the best machinists working with the best tools and machines can't
hold 0.0001 inch over a yard!

It's not true that glass is like jello at these dimensions.  If glass did not
have rather high stiffness, we would be sunk.  The first stage in getting to the
tolerances we need starts with a quite stiff, highly uniform, non crystalline
material with quite low thermal expansion and decent environmental resistance
that can be worked to extremely high precision.  Not many materials fill the
bill.  Even glass has to be specially formulated, processed and selected in
order to be good enough for really fine work on a large scale.  But, at the
nanometer scale, even glass is not stiff enough to hold the required tolerances
under it's own weight for anything larger than about 6 inch diameter.  (Even 6
inch mirrors should be on a  well dimensioned 3 point support.)

Mark Holm
mdholm@telerama.com