WARNING TO USERS

The TGS response matrices, in FITS format, available in the HEASARC archive prior August 1997 and on the CD "Spectra and Lightcurves" (June 1996), are incorrect.
The new files has been replaced in the directories (and subdirectories) /exosat/data/tgs/spectra and /exosat/cdrom/vol2/disc2 in the legacy FTP area. The responses retrieved from Browse and in the CD area are the FITS files with the ORDERS SEPARATED (see below for TGS responses further information ).
In addition 13 GSPC responses for GSPC spectra taken simultaneosly with the TGS were re-made for compatibilty with xspec 10. (see below for GS responses further information )

TGS responses problem

The TGS response originally generated at the EXOSAT observatory were in the old SF format. At HEASARC the responses were translated into a FITS format using the program rsp2rmf.
A peculiarity of the grating response is that different orders can overlap in channel space. Therefore at given energy range, the total efficiency in one channel is the sum of the effiencies of all left and right orders which contribute to that channel. The original TGS responses were written such that at given energy range, the efficiencies which contribute to a channel range for each order were kept in separate group. Therefore several groups could overlap in channel space when several left and right orders contribute to the same channel.
The program rsp2rmf assumes that at given energy range each group contributes to a unique channel range and does not account for several groups contributing to the same channel space. The resulting TGS response matrix has therefore a deficency in the total efficiency at energy where several left and right orders overlapped. The wrong version of these FITS matrices were in the HEASARC archive since March 1995, and distributed on the CD 'Spectra and Lightcurves' (June 1996).
The faulty responses were remade using a modified version of rsp2rmf. Two sets of FITS response matrices were remade. The first set preserves the structure that was in the original format, that is for a given energy range, the efficiencies for each order (either with the left and right separated or summed) which contribute to a channel range are kept in separate group (ORDERS SEPARATED). The second set, for a given energy range, has summed together the efficiencies for all orders which contribute to each channel (ORDERS SUMMED).
XSPEC works correctly on both sets of matrices. All the other programs which manipulate response matrices (addrmf,dmprmf,marfrmf,rbnrmf) work currectly only on the second set of files.
In the archive the new responses can be found in the following directories:
 /exosat/data/tgs/spectra        contain the ORDERS SEPARATED FITS file
 /exosat/cdrom/vol2/disc2/TGS    contain the ORDERS SEPARATED FITS file
                                 in each subdirectory 
 /exosat/data/tgs/responses      see below
The "/exosat/data/tgs/responses" directory contain:
 /tgs          responce matrices with the ORDERS SEPARATED
               but all the LEFT order are summed to the RIGHT orders.
               These matrices are appropriate for the spectra attached 
               to the database TGS.
  
 /tgs2         responce matrices with the ORDERS SEPARATED
               The LEFT and the RIGHT order are not summed.
               These matrices are appropriate for the spectra attached 
               to the database TGS2

 /tgs_sum      responce matrices with the ORDERS SUMMED
               but all the LEFT order are summed to the RIGHT orders.
               These matrices are appropriate for the spectra attached
               to the database TGS.

 /tgs2_sum     responce matrices with the ORDERS SUMMED
               The LEFT and the RIGHT order are not summed.
               These matrices are appropriate for the spectra attached
               to the database TGS2.

 /images       image file containing an image of the responce.

 /wrong        contain the old faulty version of the response files.

GSPC Responses problem

There are 13 responses associated with the GS spectra taken simultaneously with the TGS. These responses were availble in the HEASARC archive since march 1995 and are in the EXOSAT CD "Spectra and Lightcurves" (June 1996). These response were re-made because they contain a negative number in the column N_CHAN for the last usable channel. Although the content of the responses is correct, the negative value has been set to zero. Xspec version 10 (and higher) requires the use of the new matrices.
[EXOSAT Home] [About EXOSAT] [Archive] [Software] [Gallery] [Publications]

Page authors: Lorella Angelini Jesse Allen
HEASARC Home | Observatories | Archive | Calibration | Software | Tools | Students/Teachers/Public

Last modified: Thursday, 26-Jun-2003 13:48:33 EDT