[ATM] PLOP Request for 8"F7

[astroguy]

Some things to keep in mind when building and designing mirror support 
systems are:

1    The angle that any support plate needs to move is probably a degree or 
less assuming a relatively flat back for the glass.  Making a support plate 
with this degree of motion is relatively easy.

2    The support plates only move one time - - - when the mirror is placed 
on the support system.

Think simple and keep it simple.

Kreig


----- Original Message ----- 
From: "Mark Holm" <holmmarkd at gmail.com>
To: <jack.swaton at starryhost.com>; "ATM List" <atm at atmlist.net>
Sent: Sunday, January 31, 2010 3:45 PM
Subject: Re: [ATM] PLOP Request for 8"F7


> Hi Jack,
>
> I ran Plop, assuming plate glass and a 37mm secondary, with diameter of 
> 200mm, fl = 1400mm and thickness of 12mm.
>
> 3 points is getting near 10 nanometers RMS.  This is a bit high, but the 
> Z88 calculation brings it down to about 6nm.  Z88 is a new feature in Plop 
> that reanalyzes the design using a more sophisticated modeling method.  It 
> can't, in the current version, be used to optimize the design, but can be 
> used to refine the deformation calculation from a design optimized using 
> the previous method.  Another point to keep in mind is that 3 point cells 
> are fairly robust against construction tolerance errors, and the use of 
> real dimension pads, instead of theoretical points, also works to reduce 
> the real world deformation from the predicted value.  What I am saying is 
> that 3 points may be just fine for this mirror, especially if you are not 
> a high magnification freak.  If you don't jack this scope up to 400x, you 
> may never be affected by the very slight trefoil astigmatism that will be 
> the main cell induced error.  Trefoil astigmatism isn't so bad an error 
> anyhow, compared to the main errors that Plop helps us avoid.
>
> 6 points gives a residual deformation so low it might as well be zero.  9 
> points, using the standard options, doesn't do so well as six.  I know 
> that is counterintuitive, but it is well established as a real, and 
> expected result.  Still, the 9 point, standard option cell will work.  The 
> thing I have against it is it has more, and smaller, and harder to 
> measure, critical dimensions.  Also, the triangles have to pivot freely in 
> all directions, a spherical bearing situation.  With a 6 point cell, the 
> arms only have to float in one plane, an ordinary cylindrical bearing 
> situation.
>
> I couldn't run Z88 for the 4 point cell, but the basic RMS deformation is 
> about 4.7nm, getting small enough to be a real consideration, especially 
> because a Z88 run, if I could get it to go, would undoubtedly give a lower 
> estimate yet.  5 points gets you down to 2.3nm, below the limit I have 
> used for a long time (a conservative limit).  So, If you want to be able 
> to wow your atm friends with something a bit different, you could go with 
> a 4 or 5 point cell.
>
> I am very far from an expert in cell construction.  I have gained some 
> expertise in Plop operation and some of the related issues, but really 
> can't claim to be a great expert at that either.  If you doubt me, feel 
> free to post more questions to the ATM group.  You won't hurt my feelings, 
> and I won't come back with an offended tone.
>
> -- 
> Mark Holm
> markholm at verizon.net
> holmmarkd at gmail.com
>
> _______________________________________________
> ATM mailing list http://www.atmlist.net/
>
>
>
> -- 
> Internal Virus Database is out-of-date.
> Checked by AVG Free Edition.
> Version: 7.1.413 / Virus Database: 268.18.23/740 - Release Date: 3/30/2007
>
>