Re: ATM An efficint algorithm for robotic Foucault testing

[Dale Eason <atmpob@yahoo.com>]

The requirement is that the knife cuts into the beam
at about the same point laterally.  One way to do that
is to move the knife laterally until the average of
all the zones match a set value. I set that value
arbitrarily to less than 1/2 full intensity other
values may be just a valid. I am still investigating.
I'm not sure how much that requirement must be met. 
But you have to cut into the beam somewhere and that
rules has given me consistant results.  I got that
rule from Mick Peck's description of his flip and diff
testing.

There is an interesting thing that happens when you
set the knife that way.  If the knife is positioned at
at the null of a zone when the average is reached,
both the left and right values of the nulled zone are
at that value as well.  I think if you analize the
math the Nils Olin Carlof has on his site you can
predict that this would be the case.

--- Hugues.Laroche@ses-astra.com wrote:
> 
> 
> Hi Dale,
> 
>  Hm, I am interested! Especially since I not intend
> to motorise
>  my tester, but wish to take frames of the mirror's
> apearence.
> 
>  >>>
>  The hard part is making all the images at the same
> average intensity.  I'm not sure how different
> average
> intensities affect the results.  I will experiment
> with that next.
> <<<
> 
> I am not sure I understood right: your analysis
> software has access to all
> pixel values over the mirror? Then why not normalise
> the values of each
> frame, say betwen 0 and 1? You see what I mean?
> 
> Also one thing you did not mention: with that you
> can analise the mirror
> not only along the horizontal diameter, but others
> as well. Then you
> may quantify astigmatism, and have to test other
> positions of the mirror
> to separate the flexure part from the eventual
> "intrinsic" part.
> 
> Clear skies
> 
> Hugues
> 
> 
> 


__________________________________
Do you Yahoo!?
The New Yahoo! Shopping - with improved product search
http://shopping.yahoo.com