Re: [ATM] Flat Field Cass with doublets subaperture corrector

[Richard <cnc@cncservo.co.uk>]

Hi Richard,

Saturday, February 10, 2007, 7:00:05 PM, you wrote:

RFLRS> At 39%, the corrector position (and
RFLRS> diameter) the primary RoC, the primary conic all moved roughly back
RFLRS> to the design I presented.

RFLRS> There must be something happening with these numbers.  I strongly
RFLRS> doubt that I happened on the exact original solution.

Well.. it's quite possible you did if you used the optimiser (and I'm
sure you did). If you read the manual section on the global optimiser
it talks about "falling into local optima troughs" and you could have
done just that. The purpose of the global optimiser is to have a good
rummage round and make sure there isn't some other local minima
nearby that may be better. I think of it like that analogy they use
for explaining curved space with an elastic sheet with a heavy ball
sitting in a well. Imagine the optimtimisation space as that elastic
sheet with many heavy balls, some heavier than others. The global
optimiser will go and weigh all the balls, (finding their dead centres)
and keep the best ten as it goes, so you can always back track. It
could be ideal for this problem.

-- 

Best regards,
  Richard in the UK

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