Summary of Analysis Method and Scripts
Very basic summary
All necessary files are retrieved from the HEASARC ASCA
database. Makefilter and attitude files are retrieved along with all
unscreened and screened files. The makefilter files are used to create
new screened event files.
The nominal target position is used to extract source and background
events from a GIS2 image and the source detection significance is
determined.
If the source is not detected with a significance of at least 5 sigma,
then an upper limit is determined and no further processing is
conducted.
If a source is detected, then the source position is refined and
source and background events files are extracted in full, soft and
hard energy bands. These events files are used to create various
images, spectra and light curves. The spectra are fitted with both an
absorbed and unabsorbed power-law model in the full, soft and hard
bands.
A description of each of the main processing tasks is given below.
Each task is a shell script which utilises various FTOOLS tasks. The
tasks are listed in the same order in which they are invoked in the
pipeline. The description of each task is intended only to be a
summary. More details can be found in the scripts themselves and
comments therein. The gory details can be found in the log file which
is provided for each sequence.
Task-by-task summary in pipeline order
Uses script: do_cats.xco
Any attitude files with known errors are altered, and any unscreened
SIS events files that are know to cause processing problems (owing to
the observing mode in which the data were collected) are hidden. The
SIS effective event threshold is determined for each of the unscreened
files to determine if this changes during the observation, in which
case a lower energy boundary of 0.6 keV should be used instead of 0.5
keV. An observation catalogue is then created and the BRIGHT2 SIS mode
data are used if there are more BRIGHT2 than BRIGHT mode files. The
makefilter files are merged if required and the minimum allowed
bright-Earth angle is determined. The GIS bit assignments are
confirmed to be standard. The SIS screened files are then examined and
if the observation contains data with different SIS CCD clocking modes
then the mode having the longest exposure time is used. The SIS0 and
SIS1 are confirmed to be operating in the same clocking mode, and the
unscreened events files using other modes are hidden. Only data
from the primary SIS chips are used.
The FTOOLS task ascascreen is used to create a .xco script that is
then used by XSELECT to filter the unscreened data. Some sequences are
also screend on time to ensure that they do not contain data outside
the time bounds of their respective attitude files. The screened GIS3
file is examined to determine whether those data suffer from the loss
of digital resolution in the GIS3 spectra (the "bitfix" problem).
Uses scripts: t_zipc t_detimg
t_veron do_inst_sky_cent.xco
do_gis2_sky_bin2.xco
do_gis3_sky_bin2.xco do_gis2_sky_init.xco
do_cent_extr_spec.xco do_cent_extr.xco
The nominal RA and Dec of the target is obtained from the attitude
file. Some initial unsmoothed and smoothed detector images are created
using the script t_detimg
. Some sky images are created for each detector. The script t_veron is used to
create a combined GIS2+GIS3 image with an overlay of objects from the
VERON 2001 catalogue. In this image, the nominal position of the
target is indicated with an "N", and the VERON object closest to this
position is indicated with a star.
A source and background spectrum is extracted from the GIS2 data using
either the nominal target RA and Dec or a "custom region", and an
image is created to show this extraction region. This spectrum is
used to determine whether the object is detected with a significance
of at least 5 sigma. If the object is not detected, then an upper
limit is estimated and the processing halts.
If the object is detected, then the centroid of the object is found by
finding the positions of maximum intensity in smoothed sky images.
This is converted to detector coordinates for each detector. These
centroids are used to construct source and background regions. For the
SIS instruments, the default source region is a circle with a radius
of 45 rebinned (by the default factor of 4) pixels, which corresponds
to 4.8'. The SIS default background region comprises most of the
remainder of the primary chip. For the GIS instruments, the default
source region is a circle with a radius of 27 unbinned pixels, which
corresponds to 6.6'. The GIS default background region comprises the
area that is both inside the GIS field-of-view and within a
source-centred annulus having inner and outer radii of 35 pixels
(8.6') and 55 pixels (13.5'), respectively. If custom regions are
available, then these are used instead. Custom regions are primarily
used to excise serendipitous objects from the target source and
background regions, and are determined by-eye.
The script t_detimg is then called again to create detector images
showing the extraction regions for each detector. The script t_veron
is also called again. It creates a new image on which the centroid of
the target is indicated with a "C", and the VERON object closest to
this position is indicated with a star.
Uses scripts: do_prod_sis.xco
do_prod_gis.xco
The PI columns in the SIS0 and SIS1 events files are updated using the
FTOOLS task sispi. Events files are then extracted for each detector
for source and background regions for various energy bands. The GIS
energy bands are defined as follows: 0.75-10 keV (full), 0.75-10 keV
(hard). The soft band is not used for GIS data. The SIS energy bands
are defined as follows: 0.5-10 keV (full), 0.5-2 keV (soft), 2-10 keV
(hard), 0.6-10 keV (full2), 0.6-2 keV (soft2). In the SIS effective event
threshold was change during the observation then the full2 and
soft2 bands should be used instead of the full and soft bands,
respectively. A sky image for each detector is created, and both
source and background spectra are extracted.
Uses script: t_group
Spectra are grouped, first according to a standard grouping scheme
then also to ensure that there are no fewer than 20 counts per bin.
The latter task is performed by the script t_group . Ancillary response files are created
for all instruments, and new restribution matrix files are created for
SIS0 and SIS1.
Exposure maps are created for each instrument. The source spectrum is
used as well as three simulated power-law spectra with photon-indices
of 1.4, 1.7 and 2.0. These exposure maps are not applied in the
processing pipeline but are included in the products.
Uses scripts: t_lcvs_nbint do_html_flc.xco
Uses windows: win_full1.wi win_full2.wi
win_10per.wi
A variety of combined SIS0+SIS1 and GIS2+GIS3 light curves are created
for both source and background regions using various energy bands.
Time resolutions of 16 s, 256 s and 5760 s (the ASCA orbital period)
are used. Light curves having a time resolution of either 16 s or 256
s are required to be fully exposed, and the bins in the 5760 s
resolution light curves have a fractional exposure of at least
0.1.
The source and background spectra for each instrument are examined to
determine the value of BACKSCAL. This value is used to determine the
ratio in between the source and background region areas. The number of
source and background counts is then determined for all energy bands
and the background-subtracted number of counts and counting rate is
calculated.
Uses scripts: t_fits_pl_fullf.xcm
t_fits_pl_softf.xcm
t_fits_pl_hardf.xcm
t_fits_pl_nhfullf.xcm t_fits_pl_nhsoftf.xcm
t_fits_pl_nhhardf.xcm ezyfit_v4.tcl
A series of unabsorbed and absorbed power-law fits are conducted using
XSPEC. The spectra from all 4 instruments are fitted simultaneously.
One of three energy ranges is used for each fit: full (0.6-10
keV), soft (0.6-10 keV) or hard (2-10 keV). The energy range 5-7.5 keV
is excluded from fits to avoid possible Fe features in the
spectrum.
Each fit is performed via one of six XSPEC scripts:
unabsorbed/full (t_fits_pl_fullf.xcm) , unabsorbed/soft
(t_fits_pl_softf.xcm) , unabsorbed/hard
(t_fits_pl_hardf.xcm) , absorbed/full
(t_fits_pl_nhfullf.xcm) , absorbed/soft
(t_fits_pl_nhsoftf.xcm) , absorbed/hard
(t_fits_pl_nhhardf.xcm) . The tcl script ezyfit_v4.tcl is
called by the XSPEC scripts. An initial fit is conducted and only if
the reduced-chi-squared of this fit no larger than 15 is the complete
fitting process carried out. An XSPEC script is provided with the
Tartarus products so that users can easily being spectral fitting at
the stage of the initial fit.
As well as the best-fitting parameters,
the flux of the best-fitting model (calibrated to SIS0) is also
given. The flux is measured over the following ranges: full (0.5-10
keV), soft (0.5-2 keV) and hard (2-10 keV).
|
|
