The psf system provides a facility to evaluate the enclosed energy function for a pariticular psf at any point on the detector. The default way to do this is to evaluate the psf using the data routine in thin strips starting at the psf position on the detector. This technique is fine as long as the normalisation of the psf is rigorously computed in the data routine. If this is not the case the profile routine can return garbage or even loop indefinitely. If this happens with a new psf then either look at your normalisation in the data routine carefully, or explicitly supply a profile routine of your own. The latter option is often preferrable, as most instrument calibrations produce a parameterised energy profile early on in the calibration programme.