Документ взят из кэша поисковой машины. Адрес оригинального документа : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0160-1-0.ps.gz Дата изменения: Mon Apr 5 14:35:02 2004 Дата индексирования: Mon Oct 1 22:25:09 2012 Кодировка: Поисковые слова: южная атлантическая аномалия

XMM-Newton CCF Release Note
XMM-CCF-REL-160
OM PSF
Bing Chen
November 28, 2003
1 CCF components
Name of CCF VALDATE List of Blocks
changed
CAL VERSION XSCS ag
OM PSF1DRB 0008 2000-01-01T00:00:00 PSF-U No
PSF-B No
PSF-V No
PSF-UVW1 No
PSF-UVM2 No
PSF-UVW2 No
2 Changes
This CCF le provides a new measurement of the OM UV PSFs. Since most of the sources in the
UV lters are relatively faint, only a single PSF has been available for each of the UVW1, UVM2,
and UVW2 lters in the previously existing CCF. Alice Breeveld (MSSL) has analysed several very
crowded elds and now derived the UV PSFs for dierent count-to-framerate ratios (CFRRs).
3 Scientic Impact of this Update
These new PSFs have been derived from in- ight data on stars of a wide range of brightnesses,
taking coincidence loss and CFRR into account. This update will improve the accuracy of the OM
UV photometry.
1

XMM-Newton CCF Release XMM-CCF-REL-160 Page: 2
Figure 1: Growth curves as a function of radius for PSFs in the UVW1, UVM2, and UVW2 lters,
respectively. The solid lines are the old PSFs, and the symbols trace the new PSFs. The dierent
symbols represent the dierent CFRR groups.
4 Estimated Scientic Quality
In Figure 1, we show the growth curves as a function of radius (in pixels) for, respectively, the
UVW1, UVM2, and UVW2 lter PSFs. (For OM, one pixel is about 0.48 arcsec). The solid lines
are the old PSFs, and the symbols trace the new PSFs. The dierent symbols from the new PSFs
represent the dierent CFRR groups. There are 5 CFRR groups for UVW1, 4 for UVW2 and 3 for
UVM2. For example, the asterisks and diamonds represent the brightest and faintest star groups
for the UVM2 lter.
There is a clear trend to narrower PSFs for brighter stars. The PSF widths for the UV lters
do not increase with an increase in energy as expected. The UVW2 and UVW1 cases seem very
similar, with the UVM2 PSF possibly being slightly narrower.
While this new CCF is stable and suфciently good, a eld-position-dependent PSF canot be
made with available data. Therefore, these new PSFs are eld-position-independent.

XMM-Newton CCF Release XMM-CCF-REL-160 Page: 3
5 Test Procedures
This new CCF has been tested using SAS version 5:4:1. SAS omichain tasks have been run through
several ODFs with dierent modes and no error message has been detected.
In order to check the in uence of the new PSFs on the nal count rates, two tests described in the
next section have been made to compare the count rates derived from the old (om psf1drb 0006.CCF)
and new (om psf1drb 0008.CCF) PSFs for the calibration star GD153 (test 1) and all stars in another
calibration eld in full frame mode (test 2).
6 Summary of the test results
6.1 Test 1:
The observation used is of an OM calibration target, GD153 in rev. 561. In Table 1, the measured
count rates in the UVW1, UVM2, and UVW2 bandpasses are given. It can be seen that the count
rates with the new PSFs are smaller than those with the old PSFs.
Table 1: The count rates for white dwarf GD153 using the old and new UV PSFs.
Filter Old PSFs New PSFs
UVW1 343.058 323.944
UVM2 167.119 147.580
UVW2 86.334 79.977
For the UV lters, XMM-Newton SAS measures counts, corrected for coincidence loss within a
12 pixel aperture, then extrapolates the counts to a radius of 35 pixels using the PSF les. The
smaller counts with the new PSFs in Table 1 arise because the new PSFs are narrower, thus the
aperture correction is smaller. The results are what we have expected.
6.2 Test 2:
A calibration observation from rev. 705 has been used. SAS has detected 2113, 256 and 166 sources
for UVW1, UVM2 and UVW2 respectively. Since the eld is crowded in the UVW1 lter, we rst
consider the results for the UVM2 and UVW2 lters. In Figure 2, the relative dierence in the count
rates refers to the dierence between the count rates measured with the old PSF and the new PSF,
divided by the count rates measured with the old PSF. It can be seen that the relative dierence in
the count rates has 2 values for UVM2 and 4 values for UVW2, depending on the brightness of the
sources. This is because the new PSF is CFRR dependent, including several dierent CFRR groups
(see Figure 1).

XMM-Newton CCF Release XMM-CCF-REL-160 Page: 4
Figure 2: The relative dierence in the count rates (see text for details) as a function of the count
rates for the UVM2 and UVW2 lters.

XMM-Newton CCF Release XMM-CCF-REL-160 Page: 5
Figure 3: Same as Figure 2, but for UVW1 lter.

XMM-Newton CCF Release XMM-CCF-REL-160 Page: 6
For UVW1, the situation is more complicated. Since the eld is crowded, in some cases we
cannot use the default aperture radius (12 pixels) as was done above for UVM2 and UVW2, a
smaller aperture radius has to be used. The background-subtracted counts at rst are extrapolated,
using the PSF, to the coincidence-loss area corresponding to a circle of 12 pixels. A correction is
made for coincidence-losses. Then a further correction is made (as in the case of UVM2 and UVW2)
to extrapolate the counts to a radius of 35 pixels using the PSF again. In Figure 3 the sources which
do not lie on the horizontal lines are processed using a smaller aperture radius. All results show
that this new CCF has been updated correctly.
7 Expected Updates
It is felt that a further update for UV PSFs will not be needed soon unless the OM response
changes signicantly. However, current OM throughput (fudge factors) and zero-points are based
on the measurement of GD153 and several other white dwarfs, using the old PSFs, so all of them
need to be revised soon using the new PSFs.
8 Acknowledgements
Thanks to OM team members, especially Alice Breeveld (MSSL) for her inputs.