[ATM] ATM: Thoughts about mirror cell design from a stubbornRTV-head

[sayre at foothill.net (Bruce Sayre)]

ATM cell designs -- mine included -- are stuck in time between Thomas
Grubb's whiffletree invention first used in the Armagh Observatory's
15-inch in 1835
(http://star.arm.ac.uk/history/instruments/Grubb-15inch.html) and
William Lassell's invention of astatic floatation levers for a 9-inch
telescope in 1842 (http://astron.berkeley.edu/~jrg/MIDDLE/). 

What's been missing in the discussion about "S-bending" and RTV is any
mention of flotation cells. A whiffletree by itself doesn't make a
flotation cell, even though many ATMs probably think so. Nor does it
really matter whether the mirror is glued to a whiffletree or not.  What
provides flotation is a system of astatic supports, employed in various
forms by professional cell designers for decades. So what's the
principle of astatic support?

The ideal cell holds the mirror so gravity won't distort its surface. To
avoid this bending, an astatic cell "floats" the mirror against the
force of gravity as the telescope changes its angle. To prevent any
bending, however slight, the mirror's weight must be carried by reaction
forces at each support point, which change with orientation. It's as if
the mirror was in a zero-gravity environment, or as if it was  buoyed
by forces felt if the mirror floated in a liquid of its own density, or
as if air jets just held the mirror off each lateral and axial support
point. A mirror then maintains its figure even if the telescope
structure bends or flexes. This is true whether the mirror back is fixed
to the cell or not: bending could occur in either case.
(http://medusa.as.arizona.edu/lbtwww/tech/ua9502/ua1.htm#principles)

Astatic cells were all passive until the advent of large, thin meniscus
mirrors. The era of modern reflectors, from the G. W. Ritchey's early
reflectors through the 200" Hale telescope
(http://www.weertman.com/bruce/porter/fullsize/mirror_support_cover.jpg)
used astatic levers, usually attached to the back of the mirror. These
were all passive with fixed points of support, usually with holes cast
into the back of the mirror. The big telescopes made today employ active
astatic actuators in their cells to counteract not only gravity, but
deformation caused by wind or accidental telescope acceleration.

Telescopes using honeycomb sandwich mirrors made by the Steward
Observatory Mirror Laboratory (SOML) at the University of Arizona
include the Phillips Lab 3.5m telescope, the MMT 6.5 conversion, the
Magellan 6.5m telescopes and the 8.4m Large Binocular Telescope. All of
these professional grade instruments use mirrors glued with RTV at their
back plates near rib junctions to push-pull astatic actuators that
provide both lateral and axial active support. The SOML experts have
written that these mirrors "could operate in principle with a passive
support system, but...[the active support system] has the advantage of
allowing lateral forces to be applied near the back plate rather than at
the neutral surface of the mirror, because the overturning moment and
deflections [read S-bending] can be compensated by the active axial
supports." For that reason, these mirrors don't need edge supports.
(SPIE Proceedings Volume 2199, Advanced Technology Optical Telescopes V,
p. 251-253).

So what's this got to do with ATM cells? How could they work at all if
none of them (except for Fr?d?ric G?a's magnificent 32" Dobsonian,
http://astrosurf.com/altaz/index_e.htm) use astatic supports? My guess
is that even today's "thin" mirrors have so much inherent stiffness that
any distortion introduced by bending or overbending, even with a perfect
mirror, is inconsequential because the wavefront error is dominated by
seeing on those scales even in the best conditions. Bubble wrap,
carpeting, whiffletrees: all seem to work. 

So why worry? JMI has been delivering telescopes with glued whiffletrees
for a long time, which by all accounts are anything but commercial
rip-offs (http://www.jimsmobile.com/images/ngt18_mcell.jpg). I also
believe some ATM cells are built without really eliminating lateral slop
in the whiffletree, which could itself introduce significant bending. If
you really think your mirror and seeing are so near perfection, perhaps
you should redesign your whiffletree to include astatic levers. (While
you're at it, you can toss those slings and use RTV!) ;-)

Perhaps we should be a little more clear on the concepts of good cell
design before we leap to isolated, insular conclusions.  

Build bigger binoculars!    

Bruce Sayre
P. O. Box 544
Applegate, CA 95703 USA
mailto:sayre@foothill.net
http://www.foothill.net/~sayre