XMM Users' Handbook

   
Example for simulating EPIC observations

To produce the EPIC spectrum shown in Fig. 75 in § 3.7.1.2, a 6 keV Raymond-Smith thermal plasma model spectrum created with xspec was imported into SciSim. The energy resolution of the xspec input model was chosen to be 1 eV, i.e. higher than the energy resolution of EPIC. The adopted source flux is 10^-12 erg s^-1 cm^-2 in the energy band from 0.5-4.5 keV. To import the spectrum from the SciSim GUI, users must choose option ``Sources'' in the top menu, then ``new'' in the source editor, then click on ``Spectrum'' and choose the option ``xspec'' from the menu. This will prompt the user for the input file name ( FILENAME.DAT) from the current directory. The absorbing column density can also be provided in the source editor of SciSim (in units of 10²⁴ atoms cm^-2; thus, no absorbing column density needs to be modeled in xspec.

Under the top-level GUI option ``Configure'', which spawns the configuration GUI, the ``RayGen'', ``Mirror'', ``RGA'' and ``EPIC'' modules are switched on by clicking on the icons with the left mouse button. Clicking on the RayGen icon with the right mouse button, the integration time was in the current example set to 30000 [s]. Clicking on the mirror module icon (with the right mouse button), one can choose between different mirror modules. The same procedure can be followed to choose a particular EPIC camera by clicking on the EPIC icon. For a ``standard'' model simulation neither any spacecraft effects nor other options regarding the mirrors and instruments need to be specified. Most of the options offered there are included for instrument calibration purposes. Spacecraft effects are included based on the current best guess of the expected pointing performance.

Once all inputs are specified, the simulation is started by clicking on the ``start'' button at the bottom of the GUI (or ``Simulate'' ``Start'' in the top menu). While the simulation is running, a performance bar will inform the user of its progress.

Upon successful completion the results of the model run are contained in a file named, by default, ``tempe'' (the filename can be changed from the GUI), which is created in the current directory (i.e., the one from which SciSim has been started). To convert the data into a form readable by existing software, tools are provided (reporter2pha and reporter2fits) to produce a PHA format output spectrum. This spectrum can then be transfered into xspec for analysis or for display.

Extended sources can be modeled either by using an input image and using the tool image2sources described above or by using the top-level GUI menu option ``Sources'', choosing option ``Edit'' and then specifying in the folder ``Shape'' an ellipsoidal (either uniformly or non-uniformly emitting) shape (only options available), instead of the default (``point'').

Note: Default configuration files, which have been derived from the best information available before the current release of SciSim, were used exclusively for the plots in this document. Users should be aware that a number of simple changes to the configuration files might be used either to aid some simple scenario modeling, or to investigate more practical problems of science analyses. The scope of data items that might be modifiable seems at first daunting, but we note a few here that are easily changed:

• The charge transfer efficiency (CTE) should be set to 1.0 if spatially variable spectra are to be analysed, as there is no representation yet of the SAS tools which correct for spatially varying gain. On the other hand, if the observer wants to investigate likely errors involved due to such CTE changes and the minor effect on the energy resolution the data values need not be modified

• The dead spaces between EPIC CCDs have been set to representative values in order to demonstrate the cosmetic effect on focal plane coverage. Users requiring to utilise 100% field coverage in order to estimate diffuse fluxes could run simulations with dead space modified to 0.0.

• The simulation assumes perfect set-up of offset tables within the camera, so that there is no gain and offset error predicted for optically bright targets. If users should wish to simulate the effect of a ``thin'' or ``open'' filter position with some light leakage loading, then this is best obtained by modifying the value of system noise. For example, based on the suggestions of maximum visible brightness usable per filter, one can estimate the additional light loading by the following method: For the thin filter, where m_V = 17 is a typical limiting magnitude (for 1 electron per pixel), the observation of an m_V = 12 target would produce a peak signal of about 100 electrons per pixel per full frame in MOS (16 in pn). Therefore, the additional shot noise is 10 (4) electrons rms. This should be added in quadrature to the existing noise and consequently the noise field in the configuration file should be amended to 11 (6) electrons, and the resulting effect on resolution noted. Unfortunately, this does not correctly simulate the energy offset, nor the spatial variation of optical loading.