Re: ATM Mirror Evaluation (was: Getting close - figuring 21"f/6)

[Aplanatic@aol.com]

There is a direct way to measure the edge quality of a mirror -- interference.

I have made, or have been directly involved in making, four Cassegrain secondary mirrors.  In each case the mirror was tested against a carefully polished and figured concave mate.  The largest of these secondaries is 14.5" in diameter and the smallest is 4.2".  Here are some observations and thoughts that I have collected over the years from doing these mirrors:

1) Interference testing is an excellent way to determine the figure of a mirror near its edge.  To do so one offsets the mirrors during testing so that the center of one mirror is near the edge of the other.  In this way the structure of the edge of the mirror becomes very -- sometimes painfully -- obvious.

2) The standard Foucault test can be used as a good judge of the edge quality.  If the edge is turned down badly enough to be a problem then one does not see a clear diffraction ring around the mirror when the KE is set to null the 90% zone.  As the TDE gets worse (as measured by interference) one sees a steady increase in the brightness of the edge opposite to the KE when trying to null this outer zone.  Admittedly, it takes experience with both interference and Foucault to judge this semi-quantitatively.

3)  During polishing I typically turn UP the edge by keeping the strokes short.  If I then try to figure the mirror into a parabola, the edge always turns down before I can complete the correction.  In fact, it always turns down in a region about 3 mm from the edge and has about a 1/2 fringe error at the edge.  That's about a half a wave on the wavefront.  If you can do better than this then you're more skillfull than I am.  

An interesting story:  A couple of years ago I made a 5" spherical secondary.  It was about f/4.5.  After polishing the concave Pyrex test plate to a sphere and figuring and testing the convex secondary, I decided to parabolize the test plate and turn it into a little scope.  To make a long story short, I turned the edge every time, as measured by interference.  After 4 or five iterations trying to bring the edge back and re-figuring to a parabola, I simply gave up.  The edge ended up in the condition described above, i.e., 1/2 wave error at the edge and TDE over the outer 3 mm.  Personal moral:  Always plan on masking the edge.  

4) It is simply not worth the time to try to get the edge perfect.  Start with the knowledge that you're going to mask off 5 mm all the way around.  Make a mask that's 10 mm smaller in diameter than the diameter of the mirror.  If you see a diffraction ring all the way around the mirror when nulling the 90% zone with this mask in place, then be very happy and don't try to get more out of it.  After coating add a nice clean baffle to the mirror of the appropriate diameter and enjoy the crisp star images.

5) The above should suggest an interesting strategy.  Make several masks of different diameters.  Look for Foucault diffraction rings with these masks when nulling the 90% zone.  When you find the largest one where the diffraction ring is still barely visible all the way around, you've found the optimum diameter of your mirror.

YMMV,

Dave Rowe