[Date Prev][Date Next][Thread Prev][Thread Next][Search]
[HEASARC Mailing List Archives]
ROSAT Status Report #158: FTOOLS 4.0 UPDATE #1: ROSAT PACKAGE UPDATE
Date: Wed, 19 Nov 1997 13:18:09 -0400 (EDT)
Subj: ROSAT FTOOL patch released
FTOOLS 4.0 UPDATE #1: ROSAT PACKAGE UPDATE
WHERE TO GET THE UPDATE:
The FTOOLS update can downloaded via anonymous FTP from legacy.gsfc.nasa.gov
or via the WWW from the URL:
The tar file contains the files README.UPDATE1 and INSTALL.UPDATE1 as
well as the actual patch.
DESCRIPTION OF UPDATE:
*** If you use the script PCPICOR to correct your PSPC data, then you are
using PCECOR and you need this patch ***
Progress has been made on understanding and correcting residual variations
in the temporal and spatial gain of the PSPCs. A subroutine has been added
to the ftool PCECOR to correct for the temporal- and energy-dependent gain
variation discussed in Prieto, Hasinger, and Snowden (1996, A&AS, 120,
187). PCECOR allows users to correct the PI assignment of PSPC events for
the variation in the linearity of the PSPC response with absolute gain.
This tool became necessary because it was observed that the effective gain
in the 0.25 - 1.0 keV energy range was lower than expected after
normalizing to the 1.5 keV calibration-monitor line, and this effect became
worse with lower absolute gain (for example, a 0.75 keV X-ray might be
detected as a 0.72 keV X-ray when the PSPC was operating at low gain).
However, a bug in the PCECOR ftool has resulted in too small a correction
being applied to the PI channel. This bug has now been fixed and the
alternative correction algorithm, based on the Prieto paper, has also
been incorporated into the tool.
In the new version, users can choose from one of two options for the PI
1) Using calibration data from the PANTER facility in Neuried. The data
were collected several years after the launch of ROSAT, using the
engineering-model PSPC. Data from calibration lines (covering the PSPC
energy range) allowed a check of response linearity over an extended energy
2) Using in-flight observations of calibration sources (such as N132D)
to determine the PI correction as a function of time. See Prieto,
Hasinger and Snowden (1996) for details.
Method 1 was the only option available in all previous versions of PCECOR.
However, a comparison shows that method 2 produces significantly better
results for data taken during the later epochs of PSPC data (1994). In
combination with tasks to fix the SASS spatial gain correction bug the fits
to data from constant-spectrum sources observed at several times throughout
the mission show considerably greater self-consistency with the new version
of PCECOR, using either method 1 or 2. This is particularly true for the
LMC SNR N132D whose spectral characteristics were the most sensitive to
the residual gain variations. We provide both correction algorithms so
that users may have all options available to them. A paper discussing
these data-corrections will be made available as soon as it is accepted
The time dependant calibration correction is taken from the AlK calibration
file closest to the users observation. Previous versions of the task used
8 calibration files:
The new version of pctcor, v1.1.0, utilizes 13 files in place of the original 8.
The new files allowed a more accurate time-dependant calibration
correction, since they encompassed AlK data covering a larger fraction of
the PSPC lifetime.
The perl script PCPICOR calls PCECOR as the last part of the PSPC
data-correction procedure. A new version of PCPICOR is required, as the
user interface has been updated to allow the selection of data-correction
algorithm for PCECOR.
PCFILT / HRIFILT
A problem was also found with the dec alpha versions of PCFILT and HRIFILT.
There was a bug which caused those tools to have a segmentation fault under
under some circumstances.
A bug has been fixed which caused RBNRPSF to set the wrong area-weighting,
under some circumstances (this bug did not affect the radial profile data,
but ultimately resulted in a mismatch between the psf model normalization
and the data further down the data-analysis pipeline, in the output from
The perl script srcdetect had a typographical bug.
This routine now allows the user to compare the psf model and data
without the use of a pha file (which allows a counts weighted model to be
generated). In the absence of a pha file, a mean psf model is calculated
assuming uniform weighting over the PSPC energy range. The code also
allows the user to bypass the wmap, even if a pha file is used. This
allows a very fast run of the program, although not as accurate as
a full consideration of the wmap weighting.