Subject: no more "nocompute" offsets

From: Russet McMillan

Submitted: Fri, 2 May 2003 01:24:08 -0600 (MDT)

Message number: 668 (previous: 667, next: 669 up: Index)

Hello folks,

This message is primarily of interest to GRIM users, but also to any
observers who plan to use observing scripts or low-level offset

After the shutdown for emergency servicing of the azimuth drives last
December, I noticed that "uncomputed" offsets frequently caused
unpleasant-looking behavior of the telescope motors, including
accelerations and velocities higher than we like to see for the sake
of drive longevity.  At that time, Craig took steps to make certain
that all offset commands sent from Remark would be "computed," and I
altered the most popular GRIM dither scripts to make sure that all of
their offsets were also computed.  More recently I discovered that
there are additional advantages to using computed over uncomputed
offsets, so I have gone through all dither scripts in the visitor1
home directory tree to eliminate commands that would result in
uncomputed offsets.

With this message, I am warning users to expect a change in the
behavior of their dither scripts, and I am also requesting that users
not restore the "nocompute" request to their scripts.  If any users
keep GRIM dither scripts somewhere other than tycho:/home/visitor1/ or
its subdirectories, I ask that you go through your script and
eliminate the word "nocompute" wherever it appears.  A more complete
explanation follows to motivate you to comply with these requests.

A computed offset is essentially a re-slew of the telescope, including
new calculations of refraction and aberration, and conforming to
strict limits on acceleration, maximum velocity, and deceleration.  An
uncomputed offset is more what we expect from the term "offset"; it
simply moves the drives a specified amount without extensive
recalculation, and using a more generous set of restrictions on

For users of dither scripts, the advantage of using uncomputed offsets
is that they are considerably faster, taking only 1-2 seconds instead
of the up-to-ten seconds required for a computed offset.

In December, I believed that the only advantage to using computed
offsets in dither scripts would be the long-term effect of decreasing
wear on the drives.  I have since discovered that there are other
short-term science advantages as well.  Computed offsets are more
accurate in terms of positioning stars, and they also have a known
time of completion.  Uncomputed offsets return a message of completion
immediately, and in some instances the next exposure may begin while
the drives are still settling from the sudden offset.  After some
recent experience where I paid very close attention to timing, I
believe this is one explanation for the doubled images some users have
reported while using GRIM.  So I expect that using computed offsets
will result in a much lower incidence of unusable exposures for each

Because computed offsets require more time, GRIM users might want to
re-order their scripts to take more exposures at a single position and
minimize the number of offsets made.  However, I believe that the
benefits (both short-term and long-term) of computed offsets will
outweigh the small loss of time.

Please feel free to contact me if you have questions about your dither
scripts or about offsets in general.


APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 668 in the apo35-general archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to