To: atm{at}shore.net
From: Jim Burrows 
Reply-To: Jim Burrows 


Here's a selection of ideas:

> > I think the Transverse Error Analysis is well worth keeping because it is
> > so instructive in giving a real handle on how different kinds of  errors ,
> > throw light more or less out of the region of formation of the Airy Disc,
> > due to their slopes and relative width.

Transverse errors are useless - the light doesn't go where this geometric
optics technique says it does.  There's two major problems with it:  1)
diffraction gets in there and screws things up, 2) the errors aren't
weighted by the mirror area that's producing the error, i.e., a big error
in a central zone may be nothing compared to a small error in an edge zone.

> > Nah, as above, forget the waves.  A very handy thing, following our guru,
> > Carl Zambuto, is to be able to see the deviations from parabolas with RoCs
> > different from the optical best-fit which might be better figuring goals
> > (usually shorter ROCs giving high centers).
>
>AH-HAH, so that's why the option to adjust ROC!  Perhaps its time for
>someone to post a treatsy on why and how to use this option.  Until this
>post I didn't "Get It", so I'm probably not alone! (heck, I
probably still
>don't fully "Get it", and why is this the first time I'm
hearing about this?
>:)

I'm just relaying what Carl said.  Edges are tricky and important, but you
can whack away at central zones with less worry about overshoot.  So if you
can find a parabola relative to which the current surface has a high
center, that might be a good figuring target, and generally amateur mirrors
don't have tight RoC requirements.

>In the mean time, I'd like to ask for some HELP!  Mike Peck came up with a
>truly GREAT idea, that being an accurately rendered star test simulation!
>Just WOW, wouldn't that be handy! :)
>
>The problem, "I'm no where near smart enough to figure this out!!"
>
>So I'm calling on the Mathematically inclined Code warrior Guru's here to
>come up with a sub routine that will make this option a reality.

This is related to Mark Holm's comment:

>Recently I translated a 2D-FFT routine from C to VB5.  I haven't tested it or
>made any use of it yet, but, assuming it works, it will be the core of
>what you
>need to do the star test simulation (I think.)  I was thinking of doing
>something similar.  Don't know if I have time.

Fortunately, the Foucault test assumes symmetry, so we don't have to use
the 2D FFT which takes a LONG time to run, even on present day 2 GHz
machines.  A 1D FFT will do the job and, like its name, it is FAST, so the
cute idea of a simulated star test image with a real-time slider to go from
inside to outside focus would be possible.

This is related to the following:

>The next problem to overcome is choice of language.  Currently I'm writing
>in VB6, which will probably cause the person capable of creating such a
>routine to cringe!  So, I suggest you write this routine in you choice of
>language, and once working I'll translate your code over to VB (Which I'm
>getting pretty good at, especially with C or C+)  In the mean time, you tell
>me what you need for input, I'll create a plain text file with this data.
>You can then load this data on your side (in what ever language), do the
>math, and create the array.

What I can do is generate a .dll file from my (heavily used) Delphi FFT
module which would be callable from C, VB6 , or whatever.  I
haven't done this for a couple of years, so it might take a little time. 
Also provide some FFT consultation.  Finally,

At 13:46 2003-02-19 -0500, Michael Lindner wrote:

>I'll go one better. I'd like to challenge the math-heads out there to publish
>this algorithm and others in open-source code. The reason I started
>maintaining TEX is that it was the only foucault analysis program that had
>source available, so I could port it to Linux. I see nothing so secret about
>mirror testing that we should all hide our algorithms. Nothing like a few
>hundred eyes to check for bugs.

I always tack source code onto my program zip packages (cf. Sixtests,
Diffract), even with an easy-to-read Word version.  My language is Delphi
which, unfortunately, seems to be falling into the fate of other really
good languages like Algol and Ada...


         -- Jim Burrows
         -- mailto://burrjaw{at}earthlink.net
         -- http://home.earthlink.net/~burrjaw
         -- Seattle N47.4723 W122.3662 (WGS84)

--- BBBS/NT v4.00 MP