[About EXOSAT] [Archive] [Software] [Gallery] [Publications]


The recently issued EXOSAT calibration update file contains a revised set of ME difference spectra. Figure I shows the original Ar Half I and Half 2 difference spectra which were used for correcting both positive and negative offset spectra. Figures 2 and 3 show the new offset difference spectra. These spectra should provide a 'better' background subtraction than the original difference spectra for three reasons:

1. They were accumulated using many more blank field observations and so should be more representative of specific difference spectra than the originals Which were derived using data obtained in 1983 and early 1984 (Aryon) only.

2. Approximately 150 potential ME blank field observations were examined and rejected as having a source present in either the aligned or offset quadrants. The original difference spectra were used to generate background subtracted spectra and check for the presence of sources in the aligned detector data. This more cdreful screening of data, both for the presence of weak sources and for the effects of solar activity means that contamination due to faint source is liKely to be much less of a problem.

The observations used to produce the new difference spectra are:

HER 1/2 84/057 0855+143 84/345 APHCRB 85/081
VV-47 85/296* H2103+095 85/307* Abell7 85/057
CPD-48 84/329 GL876 84/311 IRAS 1 83/328
HD12929 84/233 4C 7.38 84/210 2010+72 84/356
0235+164 85/025 HYADES 83/232 0740+76 84/363
Jupiter 83/213 H0850+13 83/301 H0139-68 84/318
Z Cha 85/186 Nova Vul 85/162 H1622-751 85/070
HD129791 85/016 H2126+56 84/159 Iota Aur 85/063
Delta Tra 85/070 AM Cvn 85/155 Error Box 85/154
Error Box 35/149 1053+70 84/326 Beta Hyd 84/266
IH0900-48 84/352 0027+70 84/335 0008+70 84/355
Mira 85/353* H1028-568* 86/007 H0323+022* 86/007

*except for Detector C.

3. Separate positive and negative offset difference spectra are supplied. Examination of Figures 2 (positive offset) and 3 (negative offset) suggests that there are small differences between them.

The spectra have been smoothed using a 7 point weighted average as before but the Argon spectra have not had the counts in channel numbers 8 set to zero as before since we were hopeful that the stricter checks for the presence of weak sources would enable a reliable estimate of the shape of the spectra in these channels to be made. However, examination of the count rates obtained before and after array swaps in a number of Abell cluster observations where the source is not expected to vary suggest that it is better to set the count rates in Argon channel numbers 8 to zero as before. This has not been done on the files on the calibration update tape and is left to the individual user. in addition it is not clear whether the observed difference spectra for the corner detectors (A,D,E and H) which are very smalI are merely the result of fluctuations in the background and should be set to zero.

In general it is always better to generate a set of individual difference spectra for an observation of interest if at all possible. This can easily be done if data exists for both half arrays in a positive and negative offset direction by subtracting the -ive from the +ive spectrum and dividing the counts in the resulting spectrum by 2. This of course, assumes that the shape and magnitude of the difference spectra are the same in the two directions which is probably OK to first order (see Figures 2 and 3).

Alternatively, if only 1 array swap is present the two sets of slew data can be used to estimate the difference spectra. In this case the resulting difference spectra can be quite noisy and can more successfully be used -to normalise the intensity of the standard difference spectra.

The layout of the new difference spectra on the calibration update tape is:

Record                           Contents
1 Argon detector 1 Positive offset diff. spectrum
2 Argon detector 1 Negative offset diff. spectrum
3 Argon detector 2 Positive " " 4 Argon detector 2 Negative " " . 16 Argon detector 8 Negative " " 17 Argon Half 1 (1-4) Positive " " 18 Argon Negative " " 19 Argon Half 2 (5-8) Positive " " 20 Argon Half Negative " " 21-40 As above for the Xenon detectors

Where a record is 128 words. The values read from the tape should be multiplied by 10-5 to obtain cts/s/bin as with the original difference spectra on the CCF and are for channels 1 through 128 where channel 1 of the Argon detectors is always zero. Note that the supplied spectra are deadtime correction. In order to correct a background spectrum obtained in an offset configuration the relevant difference spectrum shjould be added if the background spectrum was obtained in a negative offset direction and subtracted if the offset spectrum was obtained in a positive offset direction.

A.N. Parmar
C. Izzo

[About EXOSAT] [Archive] [Software] [Gallery] [Publications]