CALSTIS CRJ Problem:

The Case of Dataset O4QP9G010

o4qp9g010_crj.gif

What's the cause of such a degradation ?

The reason for a so odd CRJ product is to be found in a mis-aligement
of the two exposures used to compute the product.

I computed the AVGs of RA, DEC and ROLL using the jitter file o4qp9g010_jit.fits
and this is the result:


Extension    RA          DEC            ROLL
---------   -----------  -------------  -------
O4QP9GLIJ    5:40:21.38  -69:04:39.871  275.114
O4QP9GLNJ    5:40:21.29  -69:04:40.285  275.114



The computed distance between the two exposures is: 0.64 arcsec (about 12 pixels!)

The surprising thing is that the pipeline could have discovered that
by simply looking at header RA_APER and DEC_APER keywords in the raw data file:




PHU:
RA_TARG         =       85.0894356129   (  5:40:21.465)
DEC_TARG        =      -69.0778232672   (-69:04:40.164)

Extension 1:
RA_APER         =       85.0894356129
DEC_APER        =      -69.0778232672

Extension 4:
RA_APER         =       85.0890577356
DEC_APER        =      -69.0779381543




It is clear that OPUS fills the content of the PHU by taking the values of
RA_APER and DEC_APER from the first exposure. (These are the coordinates listed in our obs. log database).  

It is as much clear that CALSTIS doesn't look at the RA_APER and DEC_APER values of the second exposure to decide whether or not perform the CRJ step.

Conclusions:


The stis pipeline should not blindly perform the crrej step, as it does now, by simply assuming that the two extensions are perfectly aligned. 
A check on the RA_APER and DEC_APER keywords should be introduced.

This could be done by setting the CRCORR switch to OMIT if the shift
is higher than a certain threshold.

This could be achieved with a quite simple change in the
ECF/CADC getref procedure.


Alberto


Stats on RA o4qp9g010_jit.fits[1]:
nrows = 143
mean = 85.08908453411148
stddev = 1.173638651439273E-6
median = 85.08908453396852
vmin = 85.08908162306706
vmax = 85.08908703772263

Stats on RA o4qp9g010_jit.fits[2]:
nrows = 96
mean = 85.08870382949186
stddev = 7.897594085233233E-7
median = 85.088703827906
vmin = 85.08870211775956
vmax = 85.08870596099145

Stats on DEC o4qp9g010_jit.fits[1]:
nrows = 143
mean = -69.07774204194124
stddev = 3.030645374780039E-7
median = -69.07774201295751
vmin = -69.07774288966662
vmax = -69.07774134501901

Stats on DEC o4qp9g010_jit.fits[2]:
nrows = 96
mean = -69.07785686515699
stddev = 2.816781070375085E-7
median = -69.07785685068006
vmin = -69.0778576641409
vmax = -69.07785623690479

-- 
Alberto Micol,Science Archive Software Specialist - Tel:+49 89 32006365
Space Telescope - European Coordinating Facility  - Fax:+49 89 32006480
European Space Agency           -     c/o European Southern Observatory 
Karl-Schwarzschild-Strasse 2,  D-85748 Garching bei Muenchen,   Germany
E-Mail: amicol@eso.org                     URL: http://archive.eso.org/