Subject: Re: Jem-X detector resolution

Hello Stefan:

Your analysis of the leakage of the pulsed Vela flux is 
very interesting and I will try to apply it to my light 
curve software as well. I am not sure I understand the 
precise definition of your "PIFCUT". I do not understand 
how you get about 80% of the flux leaking into the 
background signal with a PIFCUT value of 0.5?

In my software I also define cuts, I use the source model
to calculate an expected illumination of each detector
pixel, I quantify the leakage in 250 levels, and then
I sort the pixels according to this quantized illumination
parameter. I then seach systematically for that combination
of an upper level (above which pixels are counted as
source pixels) and a lower level (below which pixels are
used as background pixels) which yields the best signal to
noise for the final source count rate estimation. 

As you write, it is important to estimate what fraction
of the source signal we have caught in the source pixels,
and what fraction we subtract off, when subtracting the
background. Just last week I got a clear demonstration
of the importance of these estimations, when I twisted
the detector resolution parameters too far and the
program returned more Crab signal than we actually
observe. So an independent check of the validity of
the leakage estimate is certainly very valuable.


niels