update to TN 89

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

  I have added a section to TN 89 which analyzes the 
Mark IV 7x7 grid data from DS24 in a new way: completely
unflattened.  You may recall that I have been creating
night-sky flats, using them to flatten the data, then
trying to correct for the large-scale gradients introduced
by the night-sky flats.  Well, it turns out that skipping
the whole flatfielding step yields results which are slightly
better, on average.

  See

     http://a188-l009.rit.edu/tass/technotes/tn0089.html

down at the end.

  One could use light-box flats on the images and see how
well they work.  I believe I have a set of light-box images
taken by Tom, included as part of DS20.  

  Tom wrote recently:

> I keep looking at the ASAS[1,2,3] and HAT[4] and similar papers.  ASAS is 
> the most useful and I keep looking at the mag vs mag sigma plots.  These 
> seem to be about a factor of two better than the data that I have been able 
> to analyze.

  I looked at the most recent ASAS paper, 

      http://xxx.lanl.gov/abs/astro-ph/0210283

In Figure 3 of this paper, we see some plots of scatter in the 
ASAS measurements.  In Figure 3c, the difference between ASAS measurements
of the same star (when it appears near the edges of the frame)
are shown; the scatter is 0.038 mag for the brightest unsaturated
stars, which are around V=8.

  If you look at the new section of TN 89, which considers stars
moving all the way across the frame in Mark IV images, you will 
see that I measure a scatter for the brightest stars
(also around V=8) of approximately

                         V                    I  
------------------------------------------------------
 raw (unflattened)     0.022                0.017

 ensemble photometry   0.015                0.015
------------------------------------------------------

  Rah!

  So, if we reduce our data properly (i.e. not using night-sky
flats), we can do as well as they can.

                                             Michael Richmond