[ATM] automated testing helps cure low edge zone

[jonah at eecs.berkeley.edu (Jeff Anderson-Lee)]

Last night it was raining here on the left-coast so instead of star 
gazing at the new moon star party, Cary and I spent some (eight!) hours 
into the night working on figuring my 8 inch mirror.

After making a new lap we started a series of 16 test/figure/test 
cycles, with occasional pressing.  In the beginning, the mirror had a 
major low edge zone and was very over corrected (about -3.5 best-fit 
conic) with about 1.3 waves of P-V error.  Using initial sessions of 
15-20 minutes of figuring, and working towards 3-5 minute sessions near 
the end, we focused on the worst zones and turned the mirror around so 
that by the end it was slightly under corrected (-0.9 best fit conic) 
and about 1/4 wave front (1/8th wave P-V) fit to a parabola.

And of course we are not done yet!  Even with what is now a better than 
average optic, with a few more sessions we expect to push it to even 
higher standards.  Were it not for a dying battery on the tester's LED 
and the need for sleep, we might have pressed on and finished it right 
then and there.

The exciting thing was seeing the changes happen!  By using Cary's 
Robo-Foucault setup, we could test the mirror with five to seven zones 
in anywhere from 3 to 10 minutes for each round of testing!  The amusing 
thing was to realize that part way through we were already better than 
most commercial optics.  Also, the inital outer zone had looked hopeless 
to me, but Cary insisted that it was caused by polishing (not grinding) 
and could be cured by polishing rather than returning to grinding as 
most people might propose.

I must say, I'm eager to build an automated tester of my own.  It makes 
*such* a difference on knowing how the mirror is progressing and where 
to focus the efforts.

Many thanks to James Lerch for his Robo-Foucault software and Cary 
Chleborad for his implementation of the setup, his skilled 
understanding, and his time!

Jeff Anderson-Lee
Sacramento, CA