s#c#g#p#e#_#.rmf & s#c#g#p#e#_#_#.rsp

caldb/data/asca/sis/cpf/93dec20

Unknown

A total of 312 files containing various redistribution & response matrices for the SIS. The SIS, CCD chip, and data-selection etc for which each matrix is valid is encoded within the filename (as well as by internal keywords) as follows:

s$i$c$j$g$X$p$k$e$m$_$nn$.rmf & s$i$c$j$g$X$p$k$e$m$_$nn$_$xx$.rsp

• $i$ is the SIS sensor ($i$ =0 or 1)
• $j$ is the CCD chip ($j$ = 0,1,2 or 3)
• $X$ is the grade(s) which have been combined ($X$ = 0, 02 or 0234)
• $k$ is the split threshold in use ($k$ = 20 or 40)
• $m$ is a flag indicating whether the echo effect has ($m$ = 0) or has not ($m$ = 1) been corrected
• $nn$ is the number of channels ($nn$ = 512 or 1024).

The files ending in .rmf are detector redistribution matrices only; the files ending in .rsp are detector redistribution matrices which have been multiplied by an ARF (ie include the XRT effective area etc). In the latter case, the location for which the ARF was constructed is encoded using 2 characters, $xx$, where:

• The first character denotes the nominal SIS pointing position, with 1 indicating that used whilst operating in 1-CCD mode, 2 that used during 2-CCD mode, and 4 that used during 4-CCD mode.
• The second character denotes the phase of the mission, with p indicating the PV phase, and a the AO phase.

Geoff Crewe (MIT) in OGIP format, 1993 Dec 17.

OGIP format RMFVERSN = 1992a (see CAL/GEN/92-002).

• All the above matrices were generated using sisrmg version 0.4
• A circular extraction cell of 4 arcmin radius was used to generate the ARFs

• CAUTION: The following matrices contain an erroneous 'spike' at 11 keV which can severely effect data analysis.
  s1c2g#p40e1#.# where # represents a wild-card.
  • The spike is at an energy not usually used in data analysis
  • NO other files are thought to be effected.
  • The ASCA GOF at NASA/GSFC is currently deciding what should be done regarding these files.
  Geoff Crewe (MIT) reports:
  .... there is indeed an evil bug at 11keV in all the s1c2p40e1 response matrices. The files are probably ok if you toss all photons above 10 keV (which is customary for a variety of reasons).

User Comments & Potential Problems

The following comments & potential problems associated with these datasets have been reported by users. The points are currently under investigation by the GOF & h/w teams, and have not been officially confirmed/denied. Users who suspect they see these effects should contact the GOF for further information.

• Nick White (1994 Jan 22): For SIS0 there is a strong line feature at 2.2 keV which appears to be caused by a mirror M edge not being modelled correctly in the ARF file. The feature is also evident in SIS1 but is a factor of 2-3 times weaker. Keith Gendrau has created his own response matrices not using JBLDARF (from Kunieda's areas) and using these matrices the 2.2 keV feature is not seen in SIS0. I suspect a problem in JBLDARF in the interpolation of the mirror effective area.
• Nick White (1994 Jan 22): As a matter of policy the mirror team are delivering the areas with a 100 eV resolution. This does not seem fine enough, especially around the Gold M-edges. An energy step of at least 10 eV would be better at critical energies, such as around the Gold M-edges.
• Nick White (1994 Jan 22): The silicon fluorescence line for the SIS detector is too big in the current response matrix. Adding a negative line at 1.74 keV with an EW of 10 eV (s0) and 7 eV (S1) gives a better fit.
• Nick White (1994 Jan 22): There is still the problem of the increased effective mirror area above 6-7 keV, causing excesses in sources. This makes it impossible to fit sources like Cyg X-1, to see if their is a broad iron line. This excess was calibrated out for BBXRT, can this calibration be applied to ASCA ?
• Nick White (1994 Jan 22): There seems to be a problem with the low energy (less than 1 keV) response of SIS1. For joint fits of S1 and S0, there seems to be an excess which translates to lower column densities given by S1. The S0 column densities seem to be better.
• Richard Mushotzky (1994 Apr 12): The s1c2p40e1 matrices have problems at energies greater than 7 keV - not just at energies greater than 11 keV. This is most likely an analysis artifact of the 11 keV spike reported above.