From: Louis Boyd <boyd@pegasus.la.asu.edu
> >Not to throw cold water on your project but why would you want to
> >use worm gears to drive a telescope? By using disk and roller drives
> >you can simplify the machining and greatly reduce the periodic error
> >and backlash. Many modern small research telescopes use this
> >technology.
> >A single microstepped stepper motor on each axis can give excellent
> >tracking and slew rates with no gear shifting or clutches.
>
> >Lou Boyd
> >Fairborn Observatory
> What's a disk and roll3e drive? A brief description
would be enjoyed.
It's simply a round metal disc, typically one to
two times the diameter of the telescope mirror. There's one for
each axis. The telescope design can be equatorial or alt-azimuth.
The disk is usually made of a moderately hard stainless but it
can be aluminum. The roller is considerably smaller in diameter
and gives the first reduction. The rollers may also be used to
support the polar axis of an equatorial mount. The rollers are
driven either by a second disk and roller reduction or by a belt
drive. This is directly driven by a microstepped motor.
The key is that the pressure between the disk and roller must be high enough to prevent any slipping but low enough to prevent damage to the surfaces. The coefficient of friction of dry steel on steel is about
0.1 so with 100 lbs of loading it takes a 10 lb tangential force to cause any slippage. Our .5 meter telescopes run about that. This provides a safety factor in case something bumps the telescope. It will slip before any damage is done. The edge of the disk and the diameter of the roller must be large enough to keep the surface deformations low enough to not damage the metal. For a light weight 30" telescope (made entirely of steel) a 48" disk with a 1" wide face and a 1" diameter roller works fine. This
gives a 48:1 initial reduction. We use a belt drive
7.2:1 secondary reduction, then a x256 microstepped 400 step/revolution
stepper motor. For larger telescopes ( 1 meter & up) DC
servo motor are typically used with both velocity and position
encoders. On larger telescopes the disk diameter increases proportionately
with mirror diameter while the face width and roller diameter
increase at a faster rate to accommodate the weight which goes
up roughly with the cube of the mirror diameter.
The things critical on a disk and roller drive are
that the axis of the disk and roller bearings be parallel to each
other and the edge of the disk be cylindrical. No crown is used.
The roller must be mounted on low friction bearings. Normal
shielded ball bearings are sufficient. The material used for
the disk and the roller should have similar coefficients of expansion
or the drive rate will change slightly with temperature. The
roller should be harder than the disk for best wear characteristics.
While stainless materials are desirable, plain steel works OK
in a reasonably dry environment. The disk stays clean on the
rolling surfaces like a railroad track. With this type of drive
subarcsecond corrections may be made and promptly seen at the
focal plane without backlash. That doesn't mean the telescope
will point with arcsecond accuracy. There is still mount flexure.
These driver are repeatable in pointing to a few arcseconds and
thus flexure models may be made in software and pointing error
can be corrected.
For some examples check http://24.1.225.36/fairborn.html.
All but the first of these scopes use disk and roller drives.
They are simple, but are doing very high quality photometric
research.
Lou Boyd
Fairborn Observatory