Rochester Mark IV news

[Stupendous Man <richmond@a188-l009.rit.edu>]

  Last night, I ran the Rochester Mark IV for about three hours,
following a field as it moved across the sky.  I acquired about
38 pairs of images, each 60 seconds long.  This is still the
"commissioning" phase, but it's a good step forward.  I'm learning
quite a lot.

  Air flow: I noticed that images taken earlier this summer 
showed "measles", which I believe are caused by water condensing
on the silicon.  It's a big problem during the hot, humid summer
months, as Tom has noted.  Anyway, I wondered why the dry air
system wasn't doing its job.  I took apart the pump and discovered
that the two connections to the air tubing were _not_ "intake"
and "output", but instead 2 separate outputs!  That is, the pump
was pushing air into both ends of the loop which ran though the
camera heads!  Yow.  No wonder the moisture wasn't getting sucked
up by the dessicant: there must have been very little flow.

  So, during the afternoon yesterday, I hacked up a method 
to send the air back into one of the two pump units, so that it does 
make a closed loop.  I don't think I sealed it very well, but 
at least the air should flow.  Last night, the images show a much
more mild case of the measles than before.  Rah!  I am leaving
the air pump on overnight and through the day today, and hope
to remove all the moisture by tonight.

  FITS header: As I try to reduce the data today, I have found 
four or five different errors in information written to the FITS
header.  I need to fix the "tait" program so that it writes the
data in a format that the pipeline can read.

  Time: The clock on the computer controlling the telescope was fast
by 19 minutes.  Oops.  Must fix that.

  Pointing: I know the Dec of the camera home position, but not yet
the relationship between degrees and steps of the Dec motor.  I know
roughly the Hour Angle of the home position, but not yet the relationship
between (time leaving the RA drive running at high speed) and motion
in RA.  I need to write code which allows the user to point the
telescope to a given (RA, Dec) position.

  Control: The "taitcl" clien program does let me control the
camera remotely.  Good.  However, as currently written, the client
sends a command to the camera (good) and then immediately terminates,
without waiting for a reply that the command has finished executing
(bad).  As a result, I can't easily send a string of commands.
Need to fix the interprocess communication so that one can opt
to wait for one command to finish before sending another.

  Despite the many errors and mishaps, it's exciting!

                                                     Michael Richmond