Re: ATM "Frozen" liquid mirrors

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

CHEN@uit.gsfc.nasa.gov wrote:
> 
>     So here's the challenge.  Does anyone know how to make a
> turntable that is vibration free, cheaply?  The liquid mercury
> mirror telescopes use air bearings which are too expensive ($8-
> 15K), at least for me.  Since epoxy is lighter than mercury and
> more viscous (damps out vibrations better), it should be possible
> to use a less than perfectly quiet apparatus.  A high grade ball
> bearing motor, perhaps?  Come to think of it, something off of an
> old jet engine that is not smashed up might be a good candidate.
> Hmmmm.  Any suggestions?

I might be missing something, but I really don't think that it should be
all that terribly difficult to _make_ a suitable air bearing.

I never made or designed an air bearing, but I use small ones all the
time at work.  Air bearings are used to support liquid samples in 5 mm
dia precision-ground glass tubes spinning at 10 to 25 revolutions/sec,
and -- in the solids probe of one of our NMR spectrometers -- the
powdered solid sample is packed tightly in a 7 mm dia zirconia rotor
(ca. 1 mm thick walls) which we spin at speeds up to 5500
revolutions/second (Yes!  The centrifugal force is over a hundred
thousand times the force of gravity at the outer edge!)

FWIW, these are my thoughts on the issue:

I think that all you need are 2 mating precision-machined conical
surfaces with the axes of the cones being vertical.  The surfaces are
held apart by a very thin film of air, which air is injected through
several tiny holes (0.5 mm dia??) drilled in a circle through _one_ of
the surfaces.  IMHO, the required precision should be achievable with
careful work on a good lathe which is big enough to make the desired
size bearing.  It _is_ necessary for the apex angles of the conical
surfaces to match exactly.  A rounded cutting tool and a very slow feed
should produce sufficiently smooth surfaces.  Material is not critical,
so use some metal which machines very well -- perhaps aluminum or
brass.  

Alternatively, it might work better to use a thrust bearing (2 flat
surfaces separated by a film of air) to support the weight, and a
cylindrical bearing to keep the axis of rotation centered.  It should be
much easier to machine a _good_ thrust bearing than a _good_ conical
bearing, but then it might turn out to be _more_ difficult to machine a
good cylindrical bearing.  

You do not need a very high pressure or flow rate.  1 atmosphere gauge
pressure should be adequate, and 2 atmospheres (30 psig) should give you
a good safety margin.  Flow rate just needs to keep the surfaces forced
apart by a few thousanths of an inch, or less, so I would guess that a
few SCFM (standard cubic feet per minute) should be more than adequate.  

You want air flow to be extremely smooth.  That's easy, with the proper
precautions.  
        1.      Keep the air clean.  Use 2 air filters in series, the final one
capable of removing sub-micron particles.  Make sure you also remove any
oil mist.  You can kill 2 birds with one stone if you use a final filter
from Motor Guard Corp. in California.  A roll of toilet tissue works
well as the filter element, but even better is the custom rolls of paper
sold by Motor Guard for the purpose.
        2.      Keep the flow rate smooth and even.  Use a good pressure regulator
to meter the air from your source into a small ballast tank.  If there's
much pressure variation at your source, then use a 2-stage regulator, or
2 regulators in series.  The first regulator steps the pressure down
part way, and provides a fairly constant feed pressure to the 2nd
regulator.  Use a needle valve to regulate air flow from that tank to
the bearing holes.  A reasonable place to put the final filter might be
between the ballast tank and the needle valve.  Make sure the walls of
the tubing from the needle valve to the air bearing are not "stretchy",
or a resonance oscillation could be set up in the bearing.  I don't know
who makes the plastic tubing we use in the labs -- The stuff seems to be
something along the lines of nylon or polypropylene.

Filters, pressure regulator(s), valves, tubing, and fittings should cost
no more than a few hundred dollars.

        -- Gerry