We found a bug in faking BI spectra with XSPEC12. Errors of faked BI spectra become too small compared with the FI spectra, when you fake AFTER including background and response like
data ***.pi resp ***.rmf arf ***.arf back ***.bkg model ... fakeit ...You can obtain a correct result If you fake BEFORE including background and response like
model ... fakeit ***.bkg ***.rmf ***.arf ...XSPEC11 do not have this bug.
The XIS team has released a set of non-Xray background (NXB) files suitable for the preparation of AO-5 proposals.
The spectra of a blank sky, which include both the cosmic X-ray background and the non-xray background, are not provided. The XIS team would like to ask proposers to simulate the blank sky spectra by themselves.
The NXB spectra were extracted from a circular region of a radius of 6 mm centered at the XIS/HXD nominal positions. Because the unit of the BACKSCAL keyword changes with the coordinate system, the spectra were created for both the DET and SKY coordinates.
The XIS team also prepared the NXB spectra for the entire filed of view. In this case, users should note that regions irradiated by the onboard calibration sources were excluded.
Two spectra, one for the BI sensor (XIS1) and the other for the sum of FI sensors (XIS0,3), have been extracted for each coordinate system and for each extraction region. Note that the spectra for the FI sensor is summed, not averaged.
The exposure times are ~350ks. Proposers may think the exposure is not enough for simluations of long expososure observations. However, in most cases, users usually make use of "xisnxbgen" in XIS spectral analyses. This NXB generator extracts events from the NXB database of ~300 days long by default, which results in NXB spectra with an exposure of ~350ks. Thus the XIS team concluded that proposers should use the provided NXB spectra, if they will analyze their data in the ordinary way.
This release consists of the following 12 files.
These (either the sky or detector coordinate versions) are suitable
for simulation purposes. Use with a matching set of
and ARF files.
Caution in simulating XIS spectra
using the XSPEC "fakeit" command
We would like to draw your attention on the importance of the keyword "BACKSCAL" (background file scaling factor) in both the background fits files and the simulated fits files. As written in the XSPEC manual, the observed spectrum C(E) is calculated from the data spectrum D(E) and the background spectrum B(E) as follows:
C(E) = D(E) - (b_D/b_B) B(E),
where b_D is BACKSCAL written in the spectral fits file D(E) and b_B is that written in B(E). (This is a simplified explanation; see the XSPEC manual for details.) Therefore, unless correct BACKSCAL values are used, background subtraction could become largely erroneous. The NXB files for the 6mm region have BACKSCAL values of 0.187 (detector coordinate) or 0.0833 (sky coordinate). The NXB files for the entire field of view have BACKSCAL of 0.901 (detector coordinate) or 0.419 (sky coordinate). On the other hand, simulated data files with the "fakeit" command have a BACKSCAL value of 1.0. Thus, a correct BACKSCAL value, which corresponds to the extraction region you simulate, needs to be set to the data fits file manually. Note that the BACKSCAL values depend on the coordinate system assumed.. Otherwise, the background subtraction could be wrong by almost an order of magnitude.