New Data Sets

[Tom Droege <tdroege2@earthlink.net>]

Greetings to the data munchers among you.

This is by way of introduction of some new data sets that I will make 
available.  There are problems with the data reduction.  I have identified 
several critical ones.  I hope that some of you will ask for data sets (in 
the usual way - send your name and address) so that you can work on these 
problems.

I introduce two here.  There are more coming:

1)  The star matching program fails.  This almost always happens for the I 
image so it is not due to a bad position in the fits header.  Sometimes 
more frames will pass and generate .ast files if one fusses with the RA and 
Decl. position.  But it does not help very much.  A data set will be made 
available with good and bad runs from the set below.  Ask for Data Set 21 
if you want to work on this problem.  I am betting that this has something 
to do with the V focus being sharper than the I focus.

2)  Michael's photom-0.2 sometimes blows up.  When it does so, It produces 
*very large* offsets in the mag values.  +/- 10 mags or so.  Photom-0.2 is 
supposed to remove small differences over the 2.3 hour run due to changing 
sky conditions and the like.  It does some good, but sometimes introduces 
very large errors.  I suspect that it is also introducing smaller errors 
for other frames.  Note here the V correction blows up more often.  Again a 
sharpness problem?  As for Data Set 22 if you want to work on this problem.

I will make up data sets for you to attack these problems from the list 
below.  I will just pick out an appropriate set.  About a dozen CDs for 
each problem will be about right.

I have discussed this with Michael, and we will work out some sort of code 
management as we go along.  It will depend on how many of you take up the 
challenge.  I hope several for each problem.

There is at least one more problem coming.  If anyone wants to did deeply 
into the full pipeline and look for something subtle, then write me.

Here are the files I have been using and the results:

File		        Date		.clt found	photom-0.2
Hxra2246774.fits 	011117 Run 1	56		OK
Hxra2251710.fits 	011208 Run 3	53		OK	
Hxra2252683.fits 	011209 Run 3	26		Bad I
Hxra2262691.fits	011219 Run 3	17		OK
Hxra2263679.fits 	011220 Run 3	51		Bad V
Hxra2285602.fits	020111 Run 2 	26		OK
Hxra2286587.fits 	020112 Run 2	48		OK
Hxra2295569.fits 	020121 Run 1	55		OK	
Hxra2309558.fits 	020204 Run 1	56		Bad V
Hxra2310540.fits 	020205 Run 1	25		OK
Hxra2311530.fits	020206 Run 1	48		Bad V
Hxra2312524.fits 	020207 Run 1	56		OK

Hxra means the two files Hvra and Hira...

Date is the day the run started central time.  Always the evening before 
the run in UT.

Run is the sequential run for the evening.

.ast found means the number of files for which the astrmetric program 
produced a result.  This is not always due to bad RA or Decl. 
information.  The v frame almost always finds an .ast solution, while the I 
frame often fails.  The result is less than 56 collated files.

The photom -0.2 column shows the files where the photom fit blew up.  When 
this happens, there are very large mag offsets.  The values for on or two 
frames are 10 to 20 magnitudes off.  I suspect that other frames are also 
greatly in error.