[ATM] Thoughts about mirror cell design from a stubborn RTV-head

[sayre at foothill.net (Bruce Sayre)]

Mel wrote on 2/17/04:

<<...yes, but you can see rather tiny distortions in the star test...>>

For all cells, or just those not well-executed, or just those of a
certain type? 

<<...There is a very good article in an old Sky and Telescope about a
Brazilian 20 inch...>>

It's so old (early 1960) it's not available from S&T as a PDF download.
Library, here I come.

<<The drawback is the expense and extra weight of all that
counterweighting...>>

Agreed -- and they're probably overkill. 

I think a well-designed and built cell with a whiffletree glued to the
mirror need not detectably degrade a mirror's inherent capabilities,
even that rarely-achieved, 1/20 wavefront error optic  made with a
passion beyond reason because "it is human nature to strive for
unattainable perfection." But I'm making these assumptions: 

1) the mirror's thinnest ratio is 1 to 15
2) the three bottom whiffletree points are not also used for collimation

3) the whiffletree is compressed, or as thin as possible
4) the whiffletree rests on a rigid, flex-free metal structure
5) the whiffletree's flexible pivots are incapable of lateral movement
6) support locations are determined with Lewis' PLOP
7) the silicone pads are 1.5mm to 2mm thick

Successful mirror cells made with slings for lateral support probably
observe the first six constraints. But slings probably have to be
well-executed too -- something RTV-heads don't have to know anything
about!

Bruce Sayre
P. O. Box 544
Applegate, CA 95703 USA
mailto:sayre@foothill.net
http://www.foothill.net/~sayre