Re: ATM Foucault test accuracy

["Nils Olof Carlin" <nilsolof.carlin@telia.com>]

Carron wrote:

> I'm still wondering why mirrors tend to be undercorrected on star tests 
> (really I mean observing) compared to the expected results from the Foucault 
> test. I've run a couple dozen checks on mirrors that measure good, and had 
> them confirmed by a tester with better eyes than mine. Take them out and 
> check extrafocal images and they are undercorrected.

I read this to mean you find mirrors less corrected by startests than by Foucault? Others, I have seen, have not noticed this. They may be less sensitive to subtleties, or may perform either test differently. So, presuming the star test is what counts, the question to ask is what is the trouble with your Foucault testing (and of others who have found the same as you)?

Come to think of it, how do you do the Foucault? There are many variations, and if you disregard the wave optical aspects, you *may* get systematic errors that *may* be significant (AFAIK, Mike P does not bring up those aspects). One conceivable mechanism is using an equal area center zone, but simulations suggest that this error would rather be of the opposite sign...

> 
> I am not talking here about temperature anomalies, but measurements with 
> well-adjusted mirrors that have had time to reach equilibrium. On low FR and 
> large diameter mirrors, I've taken to intentionally overcorrecting by fudging 
> the target parameters (not the measurement values) in ADMIR (Yup!! I paid for 
> it and I intends to get my money's worth ;-). On a 22" f/3.7 we shortened the 
> target ROC by 9% !!! The extrafocal images look as if they had been computer 
> generated.

Have you tried one of the newer generation programs (like Dave Rowe's Figure or Jim Burrows' Sixtests)? 
I'd like a chance to try your Foucault data for a mirror with good star-testing correction, to see If I could get a clue (perhaps not terribly likely, but why not try?).

Nils Olof