ATM Tri-space Apo

[Roger Ceragioli <rogerc@dakotacom.net>]

On the advice of Mike Spooner, I've been studying a subaperture color
corrector that Roland Christen designed about 15 years ago, and which he
called the "Tri-space Corrector."  It consists of a triplet lens placed
about 1/3 of the distance from the focus of an ordinary coma-corrected
achromatic doublet.  The triplet contains KzFSN2 glass as the abnormal
dispersion element, and can turn an ordinary achromat into a very well
color corrected apochromat.  I think that this scheme must be more or less
what Teleview is now producing for some of its larger refractors.

After playing around with various possibilities for 3 or 4 hours on ZEMAX,
I seem to have confirmed that Roland's choice of glasses for the corrector
is certainly excellent, if not the best (I couldn't find anything else that
even worked).  The glasses in question are Corning's B58-53 (not sure if
this is still available, but couldn't find anything else to replace it in
the Schott or Ohara armory); KzFSN2 (N4 won't work); and BaK1.  The
advantage of this design is that you can achieve impressive performance for
much less cost than with full-sized apochromatic objectives, since the
KZFSN2 need only be 1/3 the diameter of the full objective, bringing the
cost of the glass way down.

The disadvantage of the system is that the subaperture corrector must be
positioned very accurately in the beam, since tilt and decentration of if
will result in lateral color and coma.  These two problems are in fact the
residual errors off-axis in the case of a well-centered system.  With an 8"
f/15 objective, at .25 degree off-axis, the blur spot is about 2 or 2.5
Airy disks in diameter and is smeared out as a comatic spectrum.  In
practice, I suspect, you would see nothing in the telescope, since your
magnification would be too low.  But problems would crop up in the center
of the field, if your corrector were not positioned well.  In a properly
designed and centered system, the central blur is much smaller than the
Airy disk, and with a slight aspherizing of the objective, you can get C,
e, and F all simultaneously to show wavefront errors of less than 1/10th
wave.  The performance of a full-sized triplet with KzFSN-4 is not nearly
this good.

The spacing of the corrector in the tube and thickness of its elements seem
to matter little (within reason).  How difficult it would be to make and
cement the elements is hard for me to say at present.  But none of the
curves is unrealistically fast, and all three elements can be cemented
together, protecting the unstable Kz glass from water.

Seems doable.

Roger Ceragioli