Extended Source Single Spectrum

Extended sources have higher background and so care needs to be taken when performing the background subtraction. To help reduce the background present in the data it is possible to "clean" the data using the Master Veto Rate parameter, contained in the eventrate file. The program xrthk may be used to generate a text file of time slots, for when the MVR value is between 0 and 170. This typically results in a reduction in exposure time of about 10 % but can significantly reduce the background. The particle model of Snowden et al 1991, used in the background subtraction assumes that the MVR is no greater than 170. The text file generated by xrthk can be used directly in xrtsort to specify the time ranges wanted (see Example 2). The same time ranges must be used when sorting the background file for the observation.

When analysing part or all of an extended source, e.g. the inner 0.1 degrees of a cluster of galaxies, there are no PSF considerations when choosing the source box. However, the source box must be small enough that a single vignetting correction (at a single off-axis angle) is appropriate for the file. A reasonable size source box would be a circle of radius 0.2 degrees within the inner ring, or an annulus of width 0.1 degrees outside the central ring. If a larger box is required, see the section on "radial_spectra".

It is not always obvious which off-axis angle to use for the vignetting correction of a large source box, centred on the optical axis. By default, the off-axis angle is set as the half radius of a circle or the mid-radius of an annulus. For a point source with a small source box this is not important, however, with a large source box this may not be appropriate. For instance, if a circular source box was used with a uniform distribution of photons, then the photon weighted mean off-axis angle would be $R / \sqrt{2}$, where R is the radius of the box. On the other hand if the photon distribution is strongly centrally peaked, then an off-axis angle less than the half radius would be appropriate. In general with a smooth decrease in the source radial profile, the half radius is the best value to use for the mean off-axis value, however this default can be overridden on the command line if necessary with:

      > xrtcorr offax1=8.2

Where the off-axis value is in arcminutes.