One often wants to detect sources in parallel in several variously-aligned images, made perhaps using several EPIC cameras, or observed at different times. If the images are not spatially aligned, a mask created for one exposure map will not necessarily be adequate for another. However, one naturally wants each simulated image in the set to draw from the same population of source positions and fluxes. For this reason, a super-region is created which is at least as large as the intersection of all the masks created in the previous section. The most convenient way to do this is to calculate the direction and width of that cone which just encompasses the set of masks.
The algorithm first calculates the position and radius of the cone which is just large enough to include the pointing directions of all the exposures; then adds 15 arcmin to the cone radius to account for the approximate radius of the EPIC fields; and finally adds a final radius pad of size specified via parameter padsizearcsec.
The RA and dec of the cone central ray, and the cone half-width or angle, are written to the SRCSPECS extension of the source template set (parameter srcspecset) in the keywords CONE_RA, CONE_DEC (both in decimal degrees) and CONE_RAD (arcsec). The angular area in deg of the cone is also given in the keyword CONEAREA. This last is done just for convenience - this keyword is not required in any later stage of eimsim.
It is proposed at some later stage to add other alternative shapes of boundary region, to cater for mosaics which are extended in one direction for example.