ATM RE: Stellafane Report; Computer Assisted Digital Figure Analysis

["Sidor . Kurt" <KSidor@drc.com>]

Tony and the List,

Thanks for being one of the few who stayed at the Friday night talks till
midnight to see our presentation.  We also got to set up the working demo
under the tent for about 3 hours on Saturday, the few people who came over
looking a bit puzzled at our set-up seemed very excited and interested once
we got to explain it to them.  The test was developed by Rochelle Prescott
(retired Physicist/Scientist/Engineer, 82 yrs old) about 50 years ago and
has remained unused (as far as we know) since.  He developed this test for
evaluating the parabolic mirrors he was using in a Schleiren photographic
setup for his pioneering work on some of the worlds first supersonic wind
tunnels. The test is very similar in set-up and appearance to a standard
center of curvature type "Ronchi" test.  The mirror is illuminated by a
single point source located in the plane of the grating (moving source).
The biggest difference is the design of the grating itself, it is a linear
array pattern but instead of being a standard clear line followed by an
equally wide opaque line the periodic transmission profile instead of being
a square-wave is more sinusoidal.  With this type of aperture for each slit
"appodization" occurs when viewing the mirror through the grating and
higher-order diffraction effects are suppressed to the point where no
noticeable degradation of the illuminated mirror image occurs (this allows
accurate data to be taken right up to the mirrors edge).  The mirror and
grating are only illuminated by a single point source (in my working model a
superbright yellow LED and a fiber optic strand) instead of many multiple
slits as in a standard "Ronchi" test.  I captured video images of the fringe
pattern with a Snappy and a home video camera.  These images were then
processed using the shareware Scion Image and some QBASIC code that I wrote.
The "fringe" centers were found crossing the central diameter of the mirror
and based on Rochelle's mathematical derivation of the geometry involved the
mirrors slope can easily be determined and mapped.  He originally exposed
photographic negatives and spent hours using a toolmakers microscope to
accurately measure each fringe position and even more hours to crunch the
data by hand and get the final mirror profile plot.  My working model has
reduced this process to about 2-3 minutes with most of the delay simply
being multiple pieces of software and software interface.  At Stellafane we
presented the working model evaluating a professionally made quartz F # 3.25
, 4.5" dia parabola.  The test showed a 1/10 wave high (at the 50% zone) and
a 1/10 wave low central depression.  The video captured fringe image was
capable of showing 17 complete fringes yielding 34 data points across the
mirrors diameter (the software finds the dark and clear centers of the
fringes).  With this resolution the size and shape of the zones were made
quite obvious and the test data showed an easily discernable test resolution
of about 10-20 nanometers.

With some help from some great people we met at Stellafane who have
graciously offered their assistance in this project we hope to soon have a
web-page describing the work in full and providing refined software (free)
as well as providing a source for the gratings.  We are also rewriting the
math for the software to be able to analyze all conics not just parabolas to
the same degree of resolution.  We also need to do comparative testing
against other methods such as caustic, Foucault, and perhaps even
interferometers to gauge test accuracy.  Rochelle has previously been
published in The Optical Society of America and will probably be submitting
this work as well.

We hope that this test will offer something to the optical community (ATM's
especially) as a relatively inexpensive and easy alternative to
interferometers for testing concave surfaces to very high resolutions.

I will let the list know as soon as more info will be available.  I will try
to upload a few pictures of the work to a free image-hosting page soon and
send links to the list.

Thanks!

Kurt Sidor