Date: Thu, 10 Jun 1999 06:24:04 -0600
From: long@apo.nmsu.edu (Dan Long)
Subject: 2.5m obslog 990609 ( MJD 51340 )
                          WEDNESDAY JUNE 9, 1999
                     APACHE POINT OBSERVATORY SDSS 2.5m
                              OBSERVING  LOG
    Observering team:   Dan Long                (APO)
                        Jeffrey Hayes           (APO)
                        Steph Snedden           (APO)
                        Francisco Castander     (UC)
                        Aronne Merrelli         (CMU)
                        Connie Rockosi          (UC)
                        Jim Gunn                (PU)
   Tonight was used to take pointing model data.  The secondary was moved
today to a collimated position via the spider vanes.  This was to allow the
secondary to have a collimated position even with the transverse actuators
out of commission.  The bottom line is the best fit model still gave 11
arcsec sky rms.
   The engineering camera was mounted with the corrector.  Counterweight
settings 245.
   We noticed when setting the rotator fiducial, that the valid fiducial (
the one around 1 deg ) would have a different value if the -3 deg fiducial
had been crossed first, as opposed to the +5 deg fiducial.  The difference
was about 4 arcminutes.  It would read 1:03.... when the -3 deg was crossed
first, and 0:59... when the +5 deg was crossed first.  This seems like a
problem.
   Overall it was not a very good night.  We lost about an hour at the
start because the fibers had not been swapped from the guider to the EC,
and it took a while to determine that the reason we were not seeing
starlight wasn't because the pointing was that awful, but that we were
opening and closing the shutter on the wrong camera.
   Once this was remedied, we found stars to be off about an arcminute in
elevation.  A pointing data set of about 6 points was taken ( twice...
because I forgot to LOG the points the first time ) by hand sweeping in
elevation at roughly constant azimuth.  At this point things were looking
up.  I fit the points to the existing model only allowing IE, IA, and CA to
change, and the fit yielded 4 arcsec rms.
   It should be noted here that image quality is much improved over last
night.  Star images are relatively round near focus, and not giving strange
double peaked and peanut shaped images as one swept through focus.  In fact
I was seeing sub arcsecond fwhm from the pterr centroiding numbers on
several occasions while taking pointing data.  Best focus was around -1375.
   I then proceeded to start a larger model covering the whole sky.  Here
odd things were noticed.  Namely that occasionaly stars were not showing up
on the chip, and that typical corrections were tens of arcseconds, much
worse performance than the above number would indicate.
   Since I was using regular DOPTERR instead of the KLUDGE version for the
first time since DRIFT was implemented in the MCP, I thought maybe
something was not working right here.  I started another model using
DOPTERRKLUDGE.
   The 48 point model took until twilight to finish, and sadly suffered the
same large sky rms that the partial model taken with regular DOPTERR did.
This model yielded 11 arcsecond sky rms.  This is much too large for
spectroscopy setup not to be a real chore.  Since the only thing different
between this and the earlier 1.4 arcsec model, is that the secondary is now
supported differently and has its transverse actuators disabled, I would
suspect the secondary is moving enough in a random way that it can't be
modeled out.
   I filled the spectrograph transfer dewar after I closed at about 05:00.
                                WEATHER
   Weather tonight was clear.  Winds were light southwesterly. Temperatures
were about 18 C at sunset, falling to 13 C by dawn.  The dewpoint
depression was about 25 C.  Dust counts were around 1200.
                                  Dan