Документ взят из кэша поисковой машины. Адрес оригинального документа : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0155-1-2.ps.gz Дата изменения: Thu Sep 11 20:35:01 2003 Дата индексирования: Mon Oct 1 22:23:40 2012 Кодировка: Поисковые слова: кокон

XMM-Newton CCF Release Note
XMM-CCF-REL-155
OM Astrometry
A. Talavera, A. Breeveld, B. Chen
September 11, 2003
1 CCF components
Name of CCF VALDATE EVALDATE List of Blocks
changed
XSCS ag
OM ASTROMET 0011 2000-01-01T00:00:00 | FILTER-V NO
| FILTER-U NO
| FILTER-B NO
| FILTER-UVW1 NO
| FILTER-UVW2 NO
| FILTER-UVM2 NO
| FILTER-WHITE NO
| FILTER-MAGNI NO
| FILTER-GRISM1 NO
| FILTER-GRISM2 NO
| POLYNOM MAP NO
| POLYNOM MAP2 NO
2 Changes
The new distortion map is based in observations of the G153 eld. The U lter was used and the
positions of 813 stars were measured. The old map was based in a LMC eld with only 230 sources.
Since the main source of distorsion is the detector ber taper, this map can be applied to all
lters, as it was the case with the previous one based in V lter. Tests made, e.g. for the UVW1
lter, indicate that the map is adecuate for all lters.
The previous map had been obtained by tting the measured oset (distorted position minus
linear position, or detector minus sky positions) as a polynomial function of the linear (sky) position.
However, SAS needs to convert the measured detector position into undistorted one so as to assign
1

XMM-Newton CCF Release XMM-CCF-REL-155 Page: 2
proper astronomical coordinates. Therefore what is needed is a function of the measured detector
(distorted) position and this is how the new distortion map OM ASTROMET 0010 has been built.
Another function performed by SAS using the ASTROMET contents is to correct the observed
OM images, which are aected by distortion, into real sky images. To achieve this, omatt SAS task
builds up the nal image pixel by pixel, where these pixels are traced back from their distorted posi-
tion by a time comsuming iterative algorithm. A set of polynomial coeфcients giving the distorted
position as a function of the linear one (as it was done -erroneously- in previous versions) would
allow us to speed up the process by direct computation of the polynomial for each point. Therefore,
a new table POLYNOM MAP2 has been added for this purpose into OM ASTROMET 0011. It
contains the polynomial coeфcients of the transformation from linear sky position to distorted OM
image space as a function of linear position. In other words, it is the inverse transformation of
POLYNOM MAP table.
3 Scientic Impact of this Update
Since the distortion is now much better mapped, the astrometry for the whole eld of view is
better now. The reconstructed positions have now an error around 1.6 pixels (rms) while the errors
(distortion) in the distorted image are 27.0 pixels (rms).
The addition of the new table POLYNOM MAP2, has no scientic impact on the processed
data. It just decreases the computing time.
4 Estimated Scientic Quality
The improvement in the pointing reconstruction accuracy that can be achieved with this distortion
map is from 10% to 20%. For the UVW1 lter we have measured 0.8".
The new table does not aect the quality. It will be used to build up faster the linearized sky
image which is not used for photometric analysis, not for astrometry. This is a cosmetical feature
of SAS omichain.
5 Expected Updates
No more observations will be performed. This map can be considered as nal. However, monitoring
of the instrument in the long term may indicate that some adjustement is necessary in the future.

XMM-Newton CCF Release XMM-CCF-REL-155 Page: 3
Field New map Old map
(rms error in pixels) (rms error in pixels)
G153 1.61 2.43
3C273 3.04 3.1
LMC 1.75 1.9
EXO 1.03 1.27
Table 1: Note. The old map was based on the LMC eld
Filter: U B UVW1
RA Dec RA Dec RA Dec
old map 1.1" 1.2" 1.2" 1.6"
new map 1.0" 0.9" 1.1" 1.3" 0.8" 0.8"
Table 2: Pointing reconstruction accuracy
6 Test procedures
The tests in this section and in the next one refer to OM ASTROMET 0010 whose contents is exactly
the same than OM ASTROMET 0011, except for the new table which has been veried internaly for
consistency. An estimation of the computing time savings will be provided in a forthcoming revision
of this release note.
The map has been veried for internal consistency, by comparing corrected positions of a number
of stars in dierent elds with their sky values.
Then, SAS 5.4 was used to reprocess the G153 eld, and the nal positions, RA and Dec. were
cross-correlated with the USNO catalogue.
7 Summary of the test results
Table 1 gives a comparison between old and new distortion maps for several elds.
Table 2 shows the errors in the coordinates determined by SAS (cross correlation with USNO
catalogue) using the new map
References