XMM-Newton Science Analysis System


badpixfind (badpixfind-1.37) [xmmsas_20070708_1801-7.1.0]

Meta Index / Home Page

Comments

• To quote fhelp cleansis (the ASCA equivalent routine) - Warning: This is an analysis tool, not a program to run blindly: Clean, check, repeat.

• As regards how much of the badpixfind routine can safely be incorporated into the PPS, we believe that the number of bad pixels found by badpixfind may depend sensitively on the input parameters (thresholds etc.). The task has been and is being developed so that an automatic running (and importantly, usage in badpix) can be performed. Tests on many datasets have been performed to find the most favourable parameter settings (which will be different for MOS and pn). The default settings have been set to reasonable values for MOS analysis. They are very conservative for pn analysis however, a more favourable pn parameter setting being: hithresh=0.0045 withsearchbadcolumn=Y hicolthresh=0.00105 backgroundrate=0.0001 loenergythresh=0.14 hienergythresh=10.0 narrowerthanpsf=1.5

• The searching algorithms within badpixfind can be performed within user-defined energy ranges. This will make any automatic usage of badpixfind a little safer, the noiser, very low energy photons being ignorable. For PN, the events are selected via their PHA values, for MOS, via their ENERGYE1 values.

• It is possible for a bad pixel map to be produced (with optionally, areas outside the FOV masked out also). This is described throughout the task description (see especially the user's description and the parameters and examples section). This map can be fed into evselect to create `masked' images, lightcurves etc.

• Searching for `flickering' pixels is possible, as described in the user's description. This, as with all of the mode 1 routines, must not be used blindly - any number of effects could lead to a `flickering' pixel (a bright source/feature passing briefly into the pixel, a genuinely variable source, CCD edge effects etc). It is recommended that a smallish number of flickertimesteps is used ($\sim$5). Larger numbers lead to lots more small number statistics and sometimes (especially for MOS) very large arrays.

• It is believed that this task is inapplicable to timing and burst modes, and will not be accessed.

• In obtaining the exposure information, for the pn, the LIVETIME attribute is read. For MOS, the EXPOSURE attribute is read. In the case of merged, calibrated event lists the LIVETInn attribute corresponding to the appropriate chip is read for both PN and MOS.