Date: Sat, 12 Jun 1999 06:40:03 -0600
From: long@apo.nmsu.edu (Dan Long)
Subject: 2.5m obslog 990612 ( MJD 51342 )
                           FRIDAY JUNE 11, 1999
                     APACHE POINT OBSERVATORY SDSS 2.5m
                              OBSERVING  LOG
    Observering team:   Dan Long                (APO)
                        Jeffrey Hayes           (APO)
                        Ellyne Kinney           (APO)
                        Steph Snedden           (APO)
                        Connie Rockosi          (UC)
                        Jim Gunn                (PU)
   Tonight's plan was to get a pointing model after the secondary support
system work during the day, then put the imager on and take data.  The
pointing model had 16 arcsec rms  which after some investigation seems
linked to long azimuth moves.
   The telescope was set up with the engineering camera in the evening
twilight.  Balance with no corrector was at 180.  It took a couple tries to
get the rotator fiducial to register, even after crossing the adjacent
fiducials.
   After prelimary setup, we were able to start a model by about 22:30.
This ran until around 00:30.  It was a 6x6 grid 36 point model.  The good
news was that even after the secondary work, stars showed up on the chip
most of the time using the old model.  The bad news is that the fit to the
model was a rather dismal 16 arcseconds.
   Once, near the end of the model, the azimuth axis ran away slewing to an
fk5 star.  An axis init was issued which halted motion.  I then re-issued
the track command and all was back to normal.
   When the residuals for the model were found to be so large, we decided
to see what sweeping up and down in elevation at roughly constant azimuth
would look like.  A sweep at 120 deg azimuth, between 25 and 80 deg
elevation yielded a fit of 1.2 arcseconds.  A second sweep at 0 deg azimuth
yielded a similar 1.9 arcseconds ( when the first star in the sweep was
omitted ).  Put them both together however and it yielded a hideous 26
arcseconds.
   This all would seem to indicate that the secondary is not the source of
the problem, since elevation moves are quite repeatable.  It does seem to
point to the azimuth axis.  When moving a long way in azimuth, large
pointing errors occur.  In fact when moving about 270 deg in azimuth after
the first pointing model data set, the fk5 star appeared nearly 6
arcminutes from the center of the field.  A reset of the azimuth fiducial
showed a 5.5 arcminute change in position.  A check of the other axes
fiducials showed shifts on the order of an arcsecond.
   I kludged ( and I do mean kludged ) together a pointing model which I
output.  I took both elevation stripes I did ( 120 and 0 azimuth ) and
subtracted the first calib offset for the second stripe from the Des Mt -
Act Phys Azimuth term for each star in the second stripe.  This was about
.03 deg in azimuth.  The resulting model got a fit of around 6 arcseconds
instead of 26.  I'm not really sure how much of this is valid, but on a
test of a couple stars, the elevation errors were really small ( a couple
arc seconds ) and the azimuth errors were quite large ( a couple arcminutes
).  So hopefully the model manages to characterize the elevation, leaving
azimuth as the only question mark.  If so, it will make hunting down stars
when doing spectroscopy a bit easier.
   Of course, it would be much better to find and fix what's wrong with
azimuth, but until then, this model might get us by.
   After this, we collimated the primary.  It didn't take much from the all
zeros position to get a reasonable collimation.  The secondary was set at
the following for the process.
        setmir: -1100 -227 -90 0 0
  The primary tilts which removed coma follow
        setmir: 0 0 -8 0 0
   The tilt and translation equivalent is then
        setmir: 0  0  6.3  341.6  0
        galil: 4600  -2300   350  -12900  4400  4150
        mig: ?*  .5560  .5555  .4135  1.9030  .4030
                * mig #1 was still not being read properly by the tpm
   These new numbers were put into default.dat
   The telescope was closed at 05:15.  The EC was left on, and the
cartridge with the corrector on it was placed in the Lindy cart shed.  I
also noticed the corrector clips sitting on the tool box in the enclosure,
if anyone is looking for them.
                                WEATHER
   Weather tonight was mostly clear, but there was quite a light show on
the eastern horizon.  Winds were light southwesterly until 03:00 then
turned easterly and it got much more humid and the seeing got much worse.
Temperatures were about 20 C at sunset, falling to 6 C by dawn.  The
dewpoint depression was 25 C most of the night, dropping to 6 C by Dawn.
Dust counts were 1000 to
1500.
                                  Dan