A ~20 arc sec systematic error that affects LECS and MECS coordinates has been identified and in this web page we give a brief summary of the work and provide analytical formulae and a specific task (saxposcor) to correct the error.

A detailed discussion on LECS and MECS positioning accuracy has been published by Astronomy & Astrophysics (Perri & Capalbi 2002, A&A 396, 753) and is available here. Further details can be found in the technical report "An improvement of BeppoSAX LECS and MECS source positioning accuracy".
 

1. The misalignment problem

The accuracy of LECS and MECS source positioning has been studied using a sample of bright X-ray sources with known accurate optical positions. We have computed the deviations in RA and Dec between the X-ray and the optical positions and the results are shown in Figure 1.
The measured 90% error circle radius values are 29'' for the LECS, 53'' for MECS1, 42'' for MECS2 and 43'' for MECS3.

In Figure 2 the LECS and MECS RA and Dec deviations are plotted as a function of the satellite roll angle. Clearly, LECS and MECS offsets are strongly correlated with the spacecraft roll angle, indicating a non perfect calibration (~20 arc sec) of the misalignment matrices of both instruments.

This non perfect calibration of the misalignment of the NFI instruments is due to a wrong spacecraft attitude reconstruction that affects the observations used to compute LECS and MECS misalignment matrices.
 

2. Correction equations

The observed dependence of the LECS and MECS RA and Dec deviations on the spacecraft roll angle can be used to correct the NFI source positions. We have fitted the RA and Dec deviations as a function of the roll angle obtaining the following best fits:
 

where  is the value of the roll angle expressed in degrees.

The coordinate correction is obtained adding the RA and Dec offsets to the LECS and MECS RA and Dec values:

In  Figure 3  the deviations in RA and Dec between LECS and MECS positions and the optical positions after the coordinate correction are plotted. As can be seen, the accuracy of LECS and MECS positions has significantly improved.

The 90% error circle radius after the correction is 16'' for the LECS, 38'' for MECS1, 18'' for MECS2 and 22'' for MECS3, which represent an improvement by a factor of 2 in the positioning accuracy of the NFI instruments.

Our study has been conducted analyzing independently the sky images of the four X-ray telescopes on board BeppoSAX (LECS, MECS1, MECS2, MECS3).

We have also studied the coadded MECS1+MECS2+MECS3  ("MECS") and MECS2+MECS3 ("MECS23") images. The corresponding RA and Dec deviations from optical positions (before and after the coordinate correction) are shown in Figure 4. The 90% error circle radius after the correction is 38'' for MECS and 17'' for MECS23.

The above equations can be used to correct analytically the source coordinates derived from any LECS and MECS event file processed or retrieved from the BeppoSAX archive at SSDC before March 12 2002 (see next section).
 

3. The task "saxposcor"

A specific task (saxposcor) that corrects the coordinates of the LECS and MECS event files has been prepared at the SSDC and is available here .

The task updates the TCRVL keywords of sky coordinate at reference pixels in the event files and can be used by BeppoSAX users to correct their own LECS and MECS data. After updating an event file with saxposcor the derived source coordinates are corrected according to the above equations.

The task saxposcor has been used to correct all LECS and MECS event files in the BeppoSAX archive at ASDC.
The SSDC archive correction has been performed on March 12 2002 and the coordinates derived from any NFI event file retrieved after this date are corrected.

The task saxposcor has also been included in the SAXDAS software (since version 2.3.0). The new version of  the SAXDAS software is available via anonymous ftp from the BeppoSAX ASDC ftp server.

Starting from March 12 2002, all  BeppoSAX observations are processed at the SSDC with the new version of SAXDAS.

This work has been motivated from the evidence of a ~40'' discrepancy between LECS and MECS source positions, as noticed by various BeppoSAX users. A previous discussion on MECS positioning accuracy can be found in Fiore et al. 2001, MNRAS, 327, 771.

We thank F. Fiore for useful discussions and a careful supervision of the work and Tim Oosterbroek for his contribution to the early phase of this work. We also thank F. Tamburelli and B. Saija for their help in the development of the script saxposcor.

Please send questions/problem reports to the helpdesk at www.ssdc.asi.it

Document authors: Matteo Perri and Milvia Capalbi