SAS watchout page

I get strange edge-like features in my pn spectrum. Is therefore something I need to know?

A bug has been found in arfgen, which affects pn spectra extracted from annuli. The energy interpolation in the encircled energy correction employed by arfgen is incorrect. This can lead to a saw-tooth response between 2 and 7 keV, and then again between 7 and 12 keV. This may mimic an absorption edge in the spectral fitting.

This has been fixed in SAS 5.4; in the meantime the problem can be solved by specifying the extra parameter eegridfactor on the arfgen command line, e.g.:

arfgen [your_parameters] eegridfactor=10

This bug does not noticeably affect MOS spectra, nor pn spectra extracted from circular regions.