ATM Ball Bearing Light Source

[DROWESMI@aol.com]

When Foucault testing fast primaries for Cassegrains (focal ratio < 1.5) I
currently use an on-axis approach which requires a beam splitter.  The beam
splitter in my current tester is an unsilvered plane parallel piece of glass
advertised as being flat on both surfaces to 1/8 wave.  In use, this beam
splitter produces noticable astimatism during a test for such, which worries
me.  At first I thought it may be due to mounting stress, so I have remounted
it in various configurations and orientations with little noticable
difference.  The astigmatism is definately associated with the test set-up in
some manner because it does not rotate with the mirror being tested.  In fact,
the mirror is quite free of astigmatism which I have shown by other methods.
Now I am worried that the test set-up is giving slightly erroneous results
when used as a Foucault tester.  With the beam splitter in place, the measured
astigmatism using a high power eyepiece is about 1 to 1.5 mils when the
eyepiece is focused on the focal plane of the mirror, i.e., this is the
minimum achievable circle of confusion.  Comments on this are welcome.

So, how about using the following method of creating a light source that is
nearly on axis:  Mount a ball bearing with a diameter of about 0.5" as close
as possible to the knife edge (KE) without interfering with the light being
returned by the mirror.  The bearing is mounted so that it moves towards and
away from the mirror when the KE is moved longitudinally (moving light source)
but is stationary when the KE is moved in a transverse direction, this being
easy to arrange mechanically.  The bearing would be illuminated by a laser
which is placed close to the mirror being tested, perhaps mounted on the
mirror support box.  Given the following assumptions: f/1.5 mirror, parallel
rays from the laser, perfectly spherical bearing 0.5" in diameter, bearing
placed 0.25" off-axis; then I calculate that the maximum longitudinal error in
this test for a spherical mirror is less than 25 microns, which I deem to be
quite good enough.

My questions are the following:  Is there some flaw in this method that I have
overlooked?  Will the bearing be precise enough for this test, and how do I
measure it?  Will the illumination be sufficiently bright?  I will be able to
test this last item in a few days.

Thanks for taking a look,
Dave Rowe
Torrance, CA