Re: [ATM] 3 point cell for 12 inch ?

[Jeff Anderson-Lee <jonah@eecs.berkeley.edu>]

A 3 point support at 0,44r will have 21nm P-V and 3,9nm RMS error for 
this mirror.  That's better than 1/60th wave on the wave front.  The 
refocus change in focal length is 0.022mm, which about the same as the 
focus tolerance of an f/4.8 mirror and likely less than the flexure in 
the overall OTA.  Furthermore, you should not suffer a 2 to 4-fold loss 
in performance due to construction errors because the cell is so simple 
and the measurements are so large.  You have the added benefit of no 
stiction, because there are no pivots.  It's well worth a try.

Jeff

Treur, Harro wrote:

>Hello,
>
>I am rebuilding the mirrorbox for my 12,5" Pyrex f4.8 with thickness
>2"(51mm).
>I got the advise that I should make 3 point cell.
>But this type of cell I do not consider a Flotationcell, because there is no
>rotational movement involved (like in a 6-point). Is
> 
>And is a 3 point with 0,44r still a good choice?? (I am not experienced, but
>0,44r seems a bit small to me...)
>
>I would like this to be confirmed before I start cutting aluminum and steel.
>
>Harro Treur
>Netherlands
>
>
>
>-----Oorspronkelijk bericht-----
>Van: Mark Holm [mailto:mdholm@telerama.com] 
>Verzonden: dinsdag 21 december 2004 4:31
>Aan: atm@atmlist.net
>Onderwerp: Re: [ATM] PLOP, new developments in mirror support, starte of,
>the art edge support, etc
>
>
>  
>
>>Even if they did make my
>>1/240th wave cell a 1/120th wave cell would I or
>>anyone else be able to tell the difference?
>>
>>    
>>
>One of the things ATM's have come to realize, at least some of us, in 
>the past 5 - 10 years is that there are ways to estimate, fairly 
>reliably, how much mirror (or other optic) error can be tolerated before 
>one might notice the difference.  The beauty is that it really doesn't 
>require terribly complicated math.  The RMS wavefront error can be fed 
>through fairly simple equations to get an estimate of the Strehl ratio 
>and the Strehl ratio can be, and has been, related to observational 
>experience.
>
>Everybody who has gone to the trouble of testing it out on a well 
>characterized mirror says you can tell the difference between a Strehl 
>of 0.8 and 0.9+  From 0.9 to 0.95+, is probably getting near the limit 
>of detectability and above 0.95, probably nobody can reliably tell the 
>difference.
>
>This plus Jeff Anderson Lee's results on likely effects of construction 
>tolerances on cell performance are what lead me to recommend a maximum 
>of 2.5 nm for RMS deformation in Plop.  2.5 nm all by itself is way 
>below detectability, but Jeff's work says that factors of 2 to 4 
>worsening are possible.  10 nm RMS gives a Strehl of not quite 0.94, so 
>I consider it at the edge of detectability.  The 2.5 nm recommendation 
>should keep cell induced deformation below detectability in most cases, 
>even with reasonable construction tolerances thrown in.  (Of course with 
>the caveat that Plop does not model lateral forces on the mirror.)
>
>My RMS to Strehl calculator, using equations I cribbed from Jim Burrows, 
>is at  http://pong.telerama.com/~mdholm/atm/cells/rmscalc.html
>
>  
>

_______________________________________________
ATM mailing list http://www.atmlist.net/