[ATM] If it's in a supremely stiff cell, that does not twist,why would a thin mirror sag?

[jonah at eecs.berkeley.edu (Jeff Anderson-Lee)]

artbianconi@blast.net wrote:

>In any case, there also is no demonstrated need for the a 1/10 wave precision 
>on either surface The mirror will "float" in a thin (0.001 to 0.005") film of 
>Teflon that is soft enough to conform to slight surface irregularities and 
>slippery enough to float and achieve equilibrium.
>  
>
If you don't polish both surfaces, then they will have irregularites on 
the order of several wavelengths to several 0.001".  These 
irregularities will govern where the initial points of contact are 
between the layers of the sandwhich.  If you make the backplane 
sufficiently stiff (your plan), then any difference will be made up by a 
combination of sagging of the glass and compression of the teflon.

If you do an analysis, I think you may find that a 0.001" to 0.005" inch 
layer of teflon is neither as compressible nor as uniformly thick as you 
might think.  That is, it while it will (mostly) conform to the surface 
irregularies, it will not significantly mask them and may introduce 
irregularities of its own.  [To avoid this, you would likely need to 
introduce a significant thickness of a highly compressible material such 
as carpet or foam.]

Thus, the glass WILL SAG between the intial points of contact, until 
other contact points reach the teflon and continue to sag until 
sufficient back-pressure establishes a new support point.  The process 
will continue until there are enough points with sufficient force to 
balance the weight of the mirror.

Now, the same sagging happens with a conventional cell, except that care 
is taken to (a) carefully place the contact points and (b) through 
balancing of leverage, match the forces on the points to within 
tolerable limits.  Using FEA (e.g. gui_plop) we can predict the sagging 
of the mirror for given arrangement of contact point placement and 
relative forces and be assured that with careful contruction we can 
bound the extent of the sagging to within reasonable tolerances (i.e. 
RMS wavefront error on the mirror significantly less than 1/10th 
wavelength).

With your design however, there no predicting where contact points will 
be as it is left to random chance of surface irregularites, and you have 
not yet convinced me that there is sufficient compressibility/uniformity 
in your thin layer of teflon to make up the difference before 
significant distortion sets in.

>I enjoy these exchanges a great deal. I am also appreciative of the feedback 
>even though I am not always in agreement with some of the conclusions drawn.
>
>What concerns me is the amount of time I have to spend explaning or justifying 
>my design. 
>  
>
You asked for our thoughts and we are giving them to you.  Differences 
of opinion are par for the course.  On the other hand, progress depends 
on trying a new way, so feel free to try it and let us know how it worked!

>Before I commit any more time to the design, I will supply them with the FEA 
>results and the models that were used in the analysis. I do not expect any 
>serious flaws in the design to be uncovered but I am prepared for anything.
>
At least you are trying to analyse the design first! As for flaws, check 
your assumptions regarding the compressibility/uniformity of the teflon 
and about how well the surfaces match under zero gravitational loading.

Jeff Anderson-Lee
Sacaramento, Ca