Subject: aource spectra

Hi SDAST+

Using the most recent version (4.0.1) of source extraction and
Niels Jörgens new ARF (0017) I have analysed the Crab spectra 
in rev. 170 and compared the XSPEC fits for JemX-1 and 2.
In particular I have looked at the dependences on Off-axis position.
With XSPEC I fitted slope and normalization for a power law
keeping nH fix (0.265 x 10^22).

The attached tar file contains the following plots:

crab170SpecSlopeJ1.ps   - JemX-1 slope as function of off axis distance
                          (XSPEC error estimates).
                        
crab170SpecSlopeJ2.ps   - Same for JemX-2

crab170SpecSlopeJ1J2.ps -  JemX-1 & 2 in same plot

(+ 2 Norm-plots as described below))

Weight mean slope for all spectra are for 
JemX-1:  2.1369 +/- 0.003  JemX-2:  2.0967 +/- 0.003

There is still a small dependence on off-axis angle amd fitting
a line give lines with

JemX-1:  Slope = 2.1013 + 0.0140*OffAngle
JemX-2:  Slope = 2.1008 - 0.0016*Offangle


The behaviour of the spectrum normalization is similar.
The systematic drift in spectral flux from on-axis to
4.5 degrees is something like 5%. The point-to-point
scatter is still dominated by the "vignetting variation".
(as we have seen before they appear to be correlated with
position).

Norm plots: crab170NormFitJ2Line.ps, crab170NormFitJ1Line.ps

I also attached a plot with reduced chi square values for the fits.

I think these results, with regards to spectral slope and 
off axis dependence are now quite reasonable. Further
improvement will require a better handling of the
"vignetting variations".

Regards,

Stefan