[Date Prev][Date Next][Thread Prev][Thread Next][Search]
[HEASARC Mailing List Archives]
ROSAT Status Report #155: PROBLEM with Flipped-Y coordinate
Date: Wed, 8 Oct 1997 11:31:12 -0400 (EDT)
Subject: ROSAT Status Report #155: PROBLEM with Flipped-Y coordinate in some ROSAT events files
ROSAT Status Report
Oct 8 1997
PROBLEM with Flipped-Y coordinate in some ROSAT events files.
A problem has been detected with some ROSAT events files in the data
archive. The problem is that the TCDLT2 keyword is negative in some
The affected files are the German Rev1 sequences, specifically, the events
files which have been converted to US fits format. The German Rev 1
datasets are easily identified as they are those with directory names like
wp600185, and the converted events files have names like wp600185.fits
(i.e. no n00, ao1 etc trailing the sequence name, and no _bas in the
filename, those are all later versions of ROSAT data, and unaffected).
The list of German Rev1 sequences can be found on
along with the value of the TCDLT2 keyword in each case, a negative
value indicating that the sequence was affected. About 10% of German Rev1
sequences have a correct value for TCDLT2. Currently 302 observations
appear to be affected.
The TCDLT2 keyword specifies how DEC changes across the field, from the
reference pixel. The flip in sign means that if you make an image from the
events file with software which uses this keyword, like xselect, then the
resulting image has the Y-direction inverted, i.e. the DEC
increases in the wrong direction. This Y-flip obviously results in an
incorrect position derivation for sources in the image.
The problem has been traced back to an early version of a conversion
program, which reformatted German _events.tfits files into w*_seq.fits
files. As this is isolated to the events conversion program, there is no
problem with other data files in the affected sequences. Most notably the
"standard" SASS images in the same dataset (the _im1, _im2 etc files) are
OK. The WGACAT database is also unaffected, and source positions in that
database are correct, even for "affected" sequences.
We are currently working on fixing the keyword in affected datasets. We
anticipate that all affected files will be corrected by the end of the week
but users may wish to check files they have retrieved in the past.
If you have any questions please email email@example.com