Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0055-1-0.ps.gz Äàòà èçìåíåíèÿ: Wed Jan 3 15:10:24 2001 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:50:16 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ðåð ð ð ð ð ð ð ð ð ð ð ð ð ð

XMM­Newton CCF Release Note
XMM­CCF­REL­55
OM Astrometry
R. Much
December 11, 2000
1 CCF components
Name of CCF VALDATE List of Blocks
changed
CAL VERSION XSCS flag
OM ASTROMET 0008 2000­01­01T00:00:00 FILTER­U Yes
FILTER­B Yes
FILTER­UVW1 Yes
FILTER­UVW2 Yes
FILTER­UVM2 Yes
FILTER­WHITE Yes
FILTER­MAGNI Yes
FILTER­GRISM1 Yes
FILTER­GRISM2 Yes
POLYNOM MAP No
2 Changes
Compared to the previous version the distortion maps of the U­, B­ and the three UV­filters have
been updated. The old distortion maps of these filters were replaced by the V­filter distortion map.
The magnifier, white light filter and the two 1st order grism distortion maps were replaced by the
V­distortion map as well. The extension holding the paramtrization of the distortion maps was
updated accordingly.
The CCF file describes the redundant detector chain.
Analysis has shown that the offsets between source positions measured in the different filters
are reduced when applying the V­filter distortion map to all filters. This is not suprising, as the
main component of the distortion is caused by the fibre taper of the detector, which introduces a
distortion independent of the filter selection. At the moment the V­distortion map has the highest
accuracy, because it was computed based on the positions of more than 230 sources.
1

XMM­Newton CCF Release XMM­CCF­REL­55 Page: 2
Although the comparison of source position offsets was limited to the V­, U­, UVW1­ and UVW2­
filter, the results can also be applied to other filters, as the main distortion component introduced
by the fibre taper is indepedant of the filter selection.
The approach to use the V­filter distortion map throughout all filters won't account for the small
global offset (at the pixel level) of the source positions between the different filters. This small global
source position offset will be subject of further calibration and will be included into later versions
of the ASTROMET file.
3 Scientific Impact of this Update
After update of the distortion maps the source positions measured in different filters match up better.
The offsets between source positions measured in different filters are reduced, which is illustrated in
Fig 1 and 2.
The average offset of the source positions is reduced between the V­filter and the UVW1­ and
UVW2­filter respectively, when using the V­distortion map throughout the calculation. It becomes
obvious when comparing the scatter of the offset positions between Figure 1 top and bottom (V­ vs.
UVW1­positions) and between Figure 2 top and bottom (V­ vs. UVW2­positions). The data points
in the lower diagram do no longer show the systematic offset and the data points are less scattered.
The offset positions are centered around (0,0). The improvement was expected for two reasons:
ffl the computation of all distortion maps except the V­distortion map was based on only a few
(several tens) sources and thus had an intrinsically lower accuracy.
ffl the main distortion component is caused by the fibre taper and thus is independant on the
filter selection. As the V­filter distortion has the highest accuracy, it describes the fibre taper
distortion best.
The performed study confirms the expectations and justifies the usage of the V­distortion map for
all filters.
4 Estimated Scientific Quality
The positional accuracy of the V­filter is unchanged, which means the source positions in the V­
filters are internally accurate at the one arcsec level.
The bottom diagram of Figure 1 and 2 should be representative enough to estimate the positional
accuracy of all other filters. The average difference between the V­filter and any other filter is
estimated as 2 arcsec from these two diagrams. The resulting overall accuracy of any filter is
calculated as the internal unertainty of the V­filter plus the filter­to­filter accuracy of 2 arcsec
(except for the V­filter, where the filter­to­filter accuracy is 0). From this the overall positional
accuracy is estimates as 1 arcsec for the V­filter and 3 arcsec (=1 arcsec + 2 arcsec) for any other
filter.

XMM­Newton CCF Release XMM­CCF­REL­55 Page: 3
Figure 1: The offset vectors between the source positions of the Coma PV observatons in the V­filter
and the UVW1­filter along the Ra­ and Dec­direction (in units of arcsec). The offsets calculated with
the old UVW1­distortion map are shown at the top and the offsets with the new UVW1­distortion
map (which is identical to the V­distortion map) are shown at the bottom. After introduction of
the new distortion map the average offset is reduced and the offsets are centered around (0/0).

XMM­Newton CCF Release XMM­CCF­REL­55 Page: 4
Figure 2: The offset vectors between the source positions of the Coma PV observatons in the V­filter
and the UVW2­filter along the Ra­ and Dec­direction (in units of arcsec). The offsets calculated with
the old UVW2­distortion map are shown at the top and the offsets with the new UVW2­distortion
map (which is identical to the V­distortion map) are shown at the bottom. After introduction of
the new distortion map the average offset is reduced and the offsets are centered around (0/0).

XMM­Newton CCF Release XMM­CCF­REL­55 Page: 5
At the time of writing the positional RMS accuracy was found to be 1.0 arcsec in the V­filter
and ú1.5 arcsec in the other filters. The analysis was performed outside the SAS.
Inside the SAS only one field was studied sofar. The mean scatter was about \Sigma1 arcsec both in
right ascension and declination, however a global pointing offset of 11 and 7 arcsec was seen in right
ascension and declination direction respectively.
Note that the statement on the absolute positional uncertainty made in the previous ASTROMET
release note is still applicable: The calculated offset between OM and STR boresight derived for
different observations (=the OM­STR boresight offset was determined in a fit whereby the offsets
between catalogued and detected positions were minimized) has a large scatter, which led to the
conclusion that the uncertainty of the STR­OM offset calibration is limited by the accuracy of the
AHF files (or the OM­STR offset is variable with time). Based on the observed scatter of the STR­
OM boresight offset the absolute positional accuracy is conservatively estimated to be better than
8 arcsec.
5 Expected Update
There is a small (less than 2 pixels) global positional offset between the source positions in different
filter elements. These offsets are subject of further calibration analysis and will be included in later
versions of the astromet CCF file.
6 Acknowledgements
Thanks to Jon Mittaz (MSSL) who verified in his analysis that the V­filter distortion map improves
the positional accuracy when applied to other filters.