ATM Re: Icarus (was "Would someone direct me to Michael Pankratz' design?")

[Michael Pankratz <papilio1@juno.com>]

>Interesting design. Very nice work. Have you done a tolerance analysis 
>on
>this?
>
>Mike Peck


Hi Mike,

Thanks!

Yes, I've done tolerancing on every element.  It turns out that this
design is quite a bit more forgiving than I would have expected -- still
tight, but certainly doable.  For example, it takes the following
deviations from nominal to reduce the Strehl ratio to 0.8:  de-center of
the triplet corrector of 0.25mm:  primary-to-triplet de-spacing of 1mm
(which recommends carbon-fiber tubing to avoid changes during thermal
cycling).  These are only about half as tight as a Cass with the same f/x
primary.  Plus, another thing in our favor is that the spherical primary,
unlike a paraboloid, has no pre-defined optical axis, so tilt and
decenter are for all practical purposes the same thing.  This means that,
as long as the primary is reasonably well aligned following assembly in
the field, collimation can be conveniently accomplished from the top end,
while looking through the eyepiece, by adjusting the centration of the
upper ring (triplet/diagonal/focuser as a unit), via a spring-loaded
micrometer arrangement.  Any small asphericity of the primary remaining
after fabrication can be "tuned-out" by adjusting the lens spacings,
allowing exceptional spherical correction of the final image -- perhaps
even allowing field adjustment to compensate for thermal changes in the
figure of the primary.  I'll check out this possibility once I've
actually built the thing.

The tilt of the diagonal, once set square to the optical axis, should
require no adjustment, assuming a reasonably stable mounting.  The final
doublet will require provision for axial adjustment, but again, once set
by examining star images should not need re-adjustment in the field. 
This doublet (just think of it as a ParaCorr or barlow) should be
physically attached to the base of the focuser (not the drawtube), and
provision should be made for a small x-y adjustment of the entire focuser
unit perpendicular to the optical axis.

All lens radii (the air-to-glass ones at least) should be accurate to
within 0.1%, but this is fairly easy using Burch's optical bench
technique in ATM3.  Those steep internal curves will be difficult to test
by interference against each other, which is the reason I eventually
opted for mating surfaces rather that the coma-correcting air-spaces of
the Jones-Bird -- by keeping the internal surfaces equal and oil-spacing
them, the figure requirements are 5 to 10 times looser (depending on the
index) than they'd be with air-spaced surfaces.  Also, with all the
degrees of freedom in this design, reasonably small amounts of wedge can
be compensated by appropriate adjustments elsewhere -- unlike triplets
used as objectives, where wedge and centration are critical.

All in all, I think, a fairly friendly design.

Cheers,

Michael Pankratz
Tesuque, NM
papilio1@ juno.com
********************************
Check out my latest optical design at: 
http://members.tripod.com/mipan/IcarusHome.htm