Notes on Disk Set 19

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

  I'm running through the data in Disk Set 19, and thought I'd
share a few discoveries with others.

  First, the V-band follow run on disk 19-2 (and, I am guessing,
the I-band follow run on disk 19-3) contain target images exclusively.
That is, all the images on that disks show pictures of the night sky;
there are no dark images.  Some of the FITS headers _say_ that they
are of type "dark", but they really aren't.

  Second, the V-band camera had some ice crystals early during the
night, but they seem to have evaporated before the end of the night.
The crystals are prominent in images of the night sky taken early
in the "flat" run (on disk 19-1), but they disappear around 
image hvra2157588.fits.  The crystals don't appear on V-band images
in the "follow" run.  Thus, one can use only the last 8 or so images
from the "flat" run to make a night-sky median flatfield for the
V-band images in the "follow" run.  I tried it, and the results
weren't very good -- just too few images to combine.  Therefore,
I'm reducing the V-band images on disk 19-2 without a flatfielding
step.  That _will_ cause the magnitudes of, say, stars in the center
to differ systematically from stars near the edges.  Too bad.
However, if all one wants to do is look for variations of a single
star during the run, it's not deadly.  After all, each star stays
in nearly the same spot on the chip ...  It will make the final
photometric calibration of the data a pain -- probably easiest to 
break up the field into chunks of around 400x400 pixels and try
to calibrate each chunk separately.

  I've run a program to find stars in the images, setting a faint
limit at 5-sigma.  I find an average of about 3000 stars in 
each image.  I measured each star's position, using Tycho-2 as a
reference, and also measured each star's INSTRUMENTAL magnitude
through an aperture of radius 4 pixels.  The FWHM was around 3.25
pixels.  You can find star lists with (RA, Dec) positions, and
with INSTRUMENTAL magnitudes, for the V-band images:

        http://spiff.rit.edu/tass/ds19/

Each file contains 8 columns of ASCII text.  You can find a full
description of them at

        http://spiff.rit.edu/tass/pipeline/pipeline.html#astrom

but, in brief, the columns are

  1.the star ID number 
  2.the RA position (decimal degrees) 
  3.the Dec position (decimal degrees) 
  4.the local sky value (ADU) 
  5.uncertainty in sky value (ADU) 
  6.instrumental magnitude in first aperture 
  7.uncertainty in instrumental magnitude in first aperture 
  8.quality flag 


  A logical thing to do is to import all these data into some
kind of database, and then match up stars by their positions, 
and THEN do the good stuff: look for variations in each star's
magnitude over the course of the run.

  Warning: since these are instrumental magnitudes, they can vary
up and down from one frame to the next; but such frame-to-frame
changes are easy to uncover, because (almost) all stars will
share the same change.  

  I'll continue with this reduction process over the next few
days; my plan is to reduce the I-band frames, find stars in them,
match up stars in each pair of V,I frames, and then try to do
some photometric calibration.  The lack of flatfielding will
make this a pain, as I mentioned above.

                                            Michael Richmond