Re: ATM 'Frozen' liquid mirrors

[Suzanne Sarlette/Gerald Pearson <suegerry@mut1.muscanet.com>]

CHEN@uit.gsfc.nasa.gov wrote:
> 
>        Finally, a note on air bearings. For spincasting the
> speeds are very slow (a few rpm) and the load is fairly heavy
> (50-100 lbs), so I suspect that the setup would be different from
> the type of air bearings use in NMR machines.  Also some
> provision would be needed to center a fair size load (16 inch+)
> very accurately.  You idea of making a custom airbearing is
> tempting, Gerry.  I don't have access to a machine shop, which is
> fortunate since my skills in that area are not that great. 

The centering issue is the beauty of using a conical air bearing.  Just
as grinding 2 surfaces together is guaranteed to produce 2 mating
spherical surfaces, one conical surface sliding over a mating conical
surface is absolutely constrained to rotate about one unique axis.  

Slow rotational speeds are no problem at all for an air bearing.  As the
rotational speed becomes progressively slower, the bearing friction
approaches _ZERO_!  An additional advantage of air bearings is that they
show absolutely no wear, when operated properly (i.e., the moving
surface never touches the fixed surface, but always floats on the thin
film of air.)

Supporting heavy loads should not be a problem.  Just inject a high
enough air flow between the 2 surfaces, and those suckers will
separate!  Atmospheric pressure is about 15 pounds/square_inch, or
something like 1 Kg/square_cm.  I was thinking in terms of a truncated
cone with apex angle maybe something like 120 degrees, with maximum
diameter maybe a foot (30 cm) or more, and the missing apex of the cone
(a hole) somewhere between 1 and 3 inches in diameter.  Even if the
average pressure in the film of air between the surfaces were as small
as 1/10 atmospheric pressure, the bearing could lift a load greater than
150 pounds!  With higher but still very modest pressures, it should be
trivially easy to generate lifting forces in excess of half a ton, if
necessary.  I would think that the only _real_ problem would be to get
access to a good lathe that was big enough to machine a sufficiently
large bearing surface.  I think that most small engine lathes might
limit you to a maximum diameter of something like 6 or 8 inches.  (But
even that might be big enough for some projects.)

I've had one additional thought about the mating surfaces.  This is what
I would do if I had a good lathe with a large bed, and if I worked up
enough enthusiasm and craziness to actually try this thing.  :-) 
Machine one surface _very_ carefully and accurately, paint/coat it with
mold release, and cover it with with a very thick coat of
epoxy/fiberglass, built up gradually in several layers.  Epoxy loaded
with a filler such as silica powder or very tiny glass spheres might
provide some useful additional dimensional stability.  Another
possibility might be to load the epoxy with a metal powder -- perhaps
aluminum.  You machine one surface, and cast the other.  As long as the
epoxy/glass composite shrinks/expands evenly, the apex angle of the cone
remains the same and the surfaces _still_ mate accurately.

Hey, brainstorming this thing and reading the other posts on this topic
is the most "ATM" fun I've had in a long time!

BTW, FWIW, I bought some "Z-poxy" brand "finishing" epoxy from a hobby
shop.  It's designed for fiberglassing model airplanes and such-like. 
This stuff is great!  When mixed, it's a rather "thin" fluid and flows
and wets stuff extremely well for maybe 15 minutes, gradually becoming
thicker over perhaps 45 minutes.  It hardens moderately slowly, and
becomes a rather flexible solid after 1 to 1.5 hours.  The next day, the
stuff is as hard and rigid as a rock!  It breaks rather than bends. 
I've cast some stuff out of an epoxy/aluminum mixture (1 part epoxy
resin, 1 part hardener, and 2 parts fine aluminum powder).  The mix is
thicker than the pure epoxy, but still flows well.  After it hardens,
it's quite strong and rigid, yet very easy to drill & tap.  It should be
very easy to machine this epoxy/metal composite.  My brother-in-law uses
Z-poxy for band instrument repair.  He recently needed to fill in some
small holes & voids in a couple of black wooden clarinets.  He mixed
some Z-poxy with a little powdered black artists' pigment ("ivory black"
= finely powdered "calcined animal bones"), filled in the spaces, and
waited for it to harden overnight.  He said it worked great, forming a
very hard, black solid.  He particularly likes that fact that the liquid
is much thinner than the usual epoxy glues, since the stuff easily flows
into tiny crevices and fills them up.

        -- Gerry