XMM Users' Handbook

   
SciSim performance issues

Several features of XMM are currently not included in SciSim. The most important are:

1.

General

• CAVEAT: Only limited validation of SciSim was performed using ground calibration data!

• SciSim run times for large numbers of rays or complicated input scenarios (as e.g., an input spectrum produced with xspec) can be very long (up to orders of hours).

• SciSim contains no information on the (expected) relative instrument alignment. This leads to shortcomings in some simulations as, for example, seam losses in between EPIC and RGS CCDs of one detector which will in practice be covered by the other unit.

• The CCD frame times for all instruments are not adjustable (from the GUI). They can only be influenced by choosing a certain instrument mode.

• For those who are interested in calibration issues, it might be important to know that at present the instrument temperatures cannot be chosen.

• Input of a sky image (emission distribution) is via a tool, which is at present still rudimentary and causes extremely long run times of the simulations.

• The pointing direction of the model run is not propagated to the output ODF.

2.

EPIC

• SciSim does not account for pn columns that might be discarded because of the hit of a minimum ionising particle (MIP; i.e., a cosmic particle).

• There is no SciSim routine to handle optical/UV loading of the EPIC cameras and, accordingly, also none to perform EPIC MOS offset calculation to correct for this effect.

• There is no EPIC pn event reconstruction in SciSim, because this is an offline functionality performed by the SAS. Thus, the SciSim tool ``cluster'' should be used at interim (see § A.4).

3.

RGS

• The scattering model that is used has not yet been optimised for the high energy end of the bandpass, which causes an overestimation of the amount of scattered light. The instrument's performance with respect to scattering wings at these energies is slightly better than predicted by SciSim.

• See above, regarding seam losses between CCDs. This is visible in Fig. 54.

4.

OM

• OSim cannot simulate OM grism observations.

• OSim does not allow the modeling OM fast mode observations.

• There is no OM window selection option.

• OM detector distortion is not taken into account.

• OM detector response non-linearity is not included.