[ATM] Bubble Wrap mirror Cell.

[Mark Holm]

At the root of most overly simplistic mirror cell proposals is the
failure to realize just how small allowable mirror deformations are.  If
you are wanting fine images from a fine mirror, your allowable RMS
deformation is on the order of 5-7 nanometers with peak to peak being
perhaps 20-25 nanometers.  Now you will have to forgive me: my small
amount of machine shop knowledge is in inches, so I will have to
convert.  Good conventional machine shop tolerances are on the order of
1/1000 inch or more.  1/1000 inch is 25400 nanometers.  So, good
precision metal work is 1000 times too sloppy for a rigid machined cell
to have any chance of working.

Adding in a soft layer complicates the analysis rather badly, but keep
that factor of 1000 in mind.  It isn't easy to provide that kind of
isolation.

The best idea I have heard for a passive (not electro-mechanically
active), "rigid" cell comprises a carbon fiber composite back plate with
the mirror supported on it by Plop located silicone RTV blobs and the
carbon fiber plate itself on 3 point suspension calculated by a general
purpose FEA program (Plop might or might not be adequate).  The reason
for carbon fiber is that it can be made to have a coefficient of
expansion similar to the glass, so the plate won't create shear stress
by differential expansion.  The RTV aspect of the silicone allows it to
compensate for the gross irregularities between mirror and back plate.
How to support the mirror sufficiently stress free while the RTV cures,
is an unsolved problem.  Perhaps you could cough up $20 million and do
the job during a Soyuz jaunt.

Figuring out how stiff the carbon fiber back plate will need to be isn't
a trivial problem.

I think perhaps a hybrid cell might turn out more practical.  You could
make carbon fiber triangles and silicone those to the mirror, then joint
those to the underlying structure.  Hey! thats already been done, not
even requiring the carbon fiber.  And maybe you could attach
counterweights to all but three of the triangles, so that you could do
away with all those teetering layers.  Hey! That's been done too.


-- 
Mark Holm
mdholm at telerama.com