In this section we give an example of how to find the hardness ratio for a source (61 Cyg) in the center of the field, and show how this can be used to constrain the spectrum. We assume that the user is familiar with PROS, FTOOLS and XSPEC.
xs> task hrifits=hrifits.cl fu> hrifits rh201950n00_obs.tab rh201950n00Here rh201950n00_obs.tab is the input file from qpspec, and rh201950n00 is the root of the output file (``_pha'' will be appended). Note: you must have run the FTOOLS package futils for this script to work. You must have the files ``fitscol'' and ``head'' (part of this distribuion.) The task hrifits will create an output file rh201950n00_pha.fits which can be read by XSPEC. The unshifted ground-based response matrix (hri_dtmat_15.fits) is hard-wired into the FITS header, as is the HRI Ancillary Response File (hri_arf.fits).
Fill in the parameter file xgain.par:
# good pars: 48.0 -40.0 0.0 rev0 # 48.0 -40.0 -0.4 rdf # EVENTS rh201950n00_bas.fits ASPECT rh201950n00_anc.fits ASPCOLS TIME,RA_SC,DEC_SC,ROAN_SC MAPS /data/andrea/HRI/jds/gain.fits MODE RDF BORESIGHT 48.0 -40.0 -0.4 TDELAY -0.65 XPOS 4168 YPOS 4056 GHIST 2.0 7.0 0.1 steps /proj/jcm/Science/hri/hv.dat
and run xgain. You will need to edit the parameters EVENTS, ASPECT, MAPS, XPOS, YPOS, steps. It may be necessary to edit the other parameters if you are not using RDF files, or are doing more detailed analysis (see Section 4. For 61 Cyg, rh201950n00_bas.fits and rh201950n00_anc.fits are the names of the event and aspect files. These are the FITS files obtained from the HEASARC archive. MAPS contains the location of the gain.fits file your local system, XPOS and YPOS are the coordinates of the source in sky pixels obtained from visual inspection of the image displayed in SAOimage. Finally, steps is the location of the hv.dat file on your local system.
The output of xgain is displayed to the terminal and written to a file xgain.fits.
This script shift_matrix.sh uses the output of xgain to construct two response matrices. The first is a matrix derived from the ground-based response matrix by shifting it by the ammount corresponding to the mean BE gain of the observation (default name ``new_mean_resp.fits''. The second is a matrix composed of several shifted response matrices, weighted by the time spent in each corresponding gain state (called ``new_resp.fits''.) If the source was not wobbled over too many gain states, the response matrix corresponding to the mean gain is fine. If the source was wobbled over a wide range of gain states the time-weighted response matrix should be used. Note that the spectral resolution can be degraded still further if the source exposure covers many gain states.