The dependence on roll angle

The non uniform distribution of RA and Dec differences is a clear indication that a systematic error is affecting the BeppoSAX NFI position accuracy, at least for the MECS units.

Figures 2 and 3 show the LECS and MECS RA and Dec deviations plotted as a function of the satellite roll angle. Clearly, both the LECS and MECS offsets are strongly correlated with the roll angle.

The amplitude of this effect is  ~15 arcsec in the case of the LECS and ~25 arcsec for the three MECS units.

The fact that RA and Dec differences are correlated with the satellite roll angle is a sure indication of a non perfect (~20 arcsec) calibration of the misalignment matrices of both LECS and MECS instruments.

Note that this correlation with the roll angle has also the effect of producing the radially asymmetric distributions shown in Figure 1.
 
 

Figure 2: The differences in RA and Dec between LECS X-ray positions and optical positions plotted as a function of the spacecraft roll angle. The solid lines are best fits to the data and are discussed in the text.

 
 
 

Figure 3: The differences in RA and Dec between X-ray positions and optical positions plotted as a function of the spacecraft roll angle (top: MECS1; middle: MECS2; bottom: MECS3). The solid lines are best fits to the data and are discussed in the text.