BAT spectra can be extracted for the full event-list, or over a specified time by using
the additional parameters tstart
and tstop
on the command line. These times must be in Swift MET (seconds since 2001-01-01). The time of the beginning/end of the dataset can be determined from the event-list by using fkeyprint
:
> fkeyprint sw00377487000bevshsp_uf.evt+0 tst # FILE: sw00377487000bevshsp_uf.evt # KEYNAME: TST # EXTENSION: 0 TSTART = 281296394. / Start time TSTOP = 281381107.0579 / Stop time
In the example below, we extract a spectrum for the T90 interval, determined from battblocks dur.gti file.
>batbinevt detmask=grb.mask tstart=281296751.488 tstop=281296755.584 >Input event file name: sw00377487000bevshsp_uf.evt >Output file name: grb.pha >Make light-curve or spectrum: pha >Histogram time bin size: 0 >Time binning algorithm: u >Energy bin list: CALDB:80
The standard BAT 80 channels have been used here (CALDB:80). Note that the spectrum produced does not need to go through the grppha
ftool; this last step has binned the file. Again, if you processed the BAT data using batgrbproduct as described on the index page, then the detector mask will be in the newly created auxil directory, under the name *b_qmap.fits.
A
systematic error vector must be applied to the BAT spectra before use to account for
residuals in the response matrix; there is a specific
task to do this, called batphasyserr
.
>batphasyserr >Input spectrum filename: grb.pha >Systematic error filename: CALDB
In order to ensure that the position of the burst in instrument coordinates is
known (since the spacecraft slews, this can change, but the knowledge is
required for the flux determination), the ray-tracing keywords must be
correct. After the spectrum has been produced, batupdatephakw
will perform this task, providing the sw[obsid]bevtr.fits ray-tracing file is
present; it should be in the [obsid]/bat/event directory. Alternatively, if you have run batgrbproduct
, it will be in the auxil directory, called sw[obsid]b_all.evaux
[If this file is missing from the archive (possibly the case for data before late
2005), and you have not run batgrbproduct
, then batmaskwtevt needs to be run to produce one: include
auxfile=sw[obsid]bevtr.fits
on the batmaskwtevt
command line to output the ray-tracing results to a file.]
>batupdatephakw >Input spectrum filename: grb.pha >Auxiliary raytracing filename: sw00377487000bevtr.fits
In order to fit the BAT spectrum, finally a detector response matrix needs to be built.
>batdrmgen >Input event PHA file name: grb.pha >Output response matrix file name: grb.rsp
The BAT team recommend fitting the spectra between 15-150 keV, since the BAT mask becomes transparent around 150 keV. Data below 14 keV and above 195 keV should definitely be ignored. The DAP housekeeping file can be left set to NONE.
In certain cases (for example, simultaneously fitting with Fermi-GBM spectra in XSPEC
), a PHA2 format BAT spectrum may be required. In this case, when batbinevt
prompts for "Make light-curve or spectrum:", the response should be pha2. When reading this PHA2 BAT spectrum into xspec
, the command is data grb.pha{1}, for example.