>I've seen a few posts that mention that the Foucault test is only good to
>about 1/10 wave. This is something that has been going around for
>years. What is the limitation based on? Are we talking about human
>limitation or something about the Foucault test itself?
It's a human limitation in judging "r". For instance, the holes in a mask might be a half inch wide. Now calculate what a 1/16" error in judgement would be. See what I mean? Also, judging the shadow edge is subjective, not entirely objective, too. A cure for this is the caustic test which uses a slit and a wire. When viewing the test, diffraction bands appear on both sides of the wire. Getting the bands to match symetrically will increase the sensitivity to approximately gamma/100. The problem with the caustic test lies in the fabrication of the tester itself. It must be made to an extrordinary degree of precision.
>I've been using a slightly modified version of the data reduction program
>that was published in S&T a few years back. My tester is a slitless
>design with a green LED light source. It is mounted on a milling machine
>(Lagun FTV3, weighs about 4000 lb.). With this setup I get results that
>are consistant to better than 1/10 wave by quite a bit.
>I see quite a few mirrors because this is where most people in the
>Seattle club test them. I don't use a zone mask. The pin stick lets me
>see the whole mirror, which I feel is important. One more thing is that
>most of the mirrors are f/7 or longer.
The pin stick is clearly better than the mask, but how well can you judge a 1/16" shadow from 8 feet away?
>So, can anybody with a good solid test setup and a little practice get
>results significantly better than 1/10 wave? Is the limitation based on
>experience or theory and calculation?
My personal opinion is that measuring a gamma/10 wavefront is a little optimistic. I use pretty much the same test, but my tolerance analysis shows me that anything better than gamma/6 cannot be confirmed by independant observation.
>I'm not a real good observer, but the mirrors that have tested
>exceptionally well also are reported to perform better than most. What
>are others experiences and opinions about this?
The mirrors I've made with the slitless test clearly outperform commercial ones. I can see the festoons on the bands of Jupiter with my 8" f/6. So, even if I cannot make a totally accurate measurement with the slitless Foucault test, the optics still turn out very well indeed. I feel comfortable saying that my wavefront is better than gamma/6 or 7, but I don't feel that I can say anything about how much better the wavefront actually is.
LensNut
-- Steve Strickland Lensnut@tpoint.net