DIYbanter - View Single Post

(Donald Nichols) writes:

I don't know for micrometers, but I have seen a Gertner (sp?)
travelling microscope (a microscope mounted on a carriage for a
lathe-bed style optical bench) with a long horizontal travel with a
micrometer leadscrew. But the nut is captive end-to-end, but free to
rotate, except for a radial rod which engages a groove in a bronze bar.
That bar has fine setscrews to adjust the angle relative to the axis of
the leadscrew so it introduces very small corrections into the pitch of
the leadscrew. All in all, a very precise bit of metalworking -- and
*very* expensive. :-)

Do you know if this was intended to compensate for errors in
manufacturing, or to allow temperature compensation, or something else?

Reminds me of another compensating mechanism, this one in softwa
Astronomical telescopes need to move to stay aligned with a star as the
earth rotates under the telescope. This is usually done by aligning one
of the axes of rotation of the scope mount with the earth's rotation
axis, and then using a motor to rotate that axis continuously at a rate
of one rotation per sideral day (23 hours 56 minutes). The timing
source could be a synchronous motor driven from line AC, or a
synchronous or stepper motor driven by an oscillator.

Just the motor drive is good enough for visual astronomy, but not for
photography where you need to keep the location of stars from drifting
on the image plane during exposures that may last many minutes. The
usual solution to this is having an auxiliary high-power eyepiece that
grabs a bit of light from outside the camera frame. You align a star in
the field of view with the crosshairs in the eyepiece, and then monitor
its location during the exposure. As it drifts away from centre, you
use buttons or switches on a guiding controller to bring the star back
to the centre of the field. Since this eyepiece is operating at a much
higher magnification than the camera, if the wandering motions are kept
small in the eyepiece they will be invisible to the camera.

The corrective motions may be provided by a small motor whose motion is
added to the main tracking motor by a differential (the old way), or
by temporarily increasing or decreasing the frequency of the oscillator
driving the tracking motor (the newer way).

There are several sources of error in all this. One of them is
inaccurately cut gears, which cause the telescope movement to
periodically speed up and slow down. Even if the average speed is
exactly correct, local speed variations cause star image movement. Some
of this error is due to periodic error in worm or plain gear cutting, and
repeats every revolution of the worm or every tooth of the plain gear.

Someone figured out that if you start out an exposure with manual
guiding, and you accurately keep the guide star centred using the hand
controller, then the history of manual corrections you make is actually
a measurement of the error in the drive train. There's a feedback loop
where the instantaneous motor speed is increased and decreased under
human command in order to keep the telescope motion constant, as viewed
by the human operator. If you record the pattern of these corrections
over several cycles of the drive system (e.g. several rotations of the
worm gear), you can estimate the corrections that are necessary *every*
cycle to compensate for mechanical tolerances. At this point, the CPU
inside the telescope controller can "play back" this sequence of corrections
over and over again, without further human help.

Dave