Документ взят из кэша поисковой машины. Адрес оригинального документа : http://xmm.vilspa.esa.es/docs/documents/CAL-SRN-0130-1-0.ps.gz Дата изменения: Mon Jan 20 12:57:28 2003 Дата индексирования: Mon Oct 1 22:17:26 2012 Кодировка: Поисковые слова: comet

XMM-Newton CCF Release Note
XMM-CCF-REL-130
RGS Bad Pixels
C. Gabriel and C. de Vries
November 29, 2002
1 CCF components
Name of CCF VALDATE EVALDATE Blocks changed XSCS ag
RGS1 BADPIX 0012 2000-01-01T22:00:00 2001-06-12T23:00:00 BADPIX NO
RGS1 BADPIX 0013 2001-06-13T00:00:00 2002-11-13T23:00:00 BADPIX NO
RGS1 BADPIX 0014 2001-08-14T22:00:00 2001-09-20T21:00:00 BADPIX NO
RGS1 BADPIX 0015 2002-11-13T23:30:00 | BADPIX NO
RGS1 BADPIX 0016 2002-11-27T21:00:00 | BADPIX NO
RGS2 BADPIX 0012 2000-01-01T22:00:00 2001-06-12T23:00:00 BADPIX NO
RGS2 BADPIX 0013 2001-06-13T00:00:00 2002-11-03T23:00:00 BADPIX NO
RGS2 BADPIX 0014 2001-08-14T22:00:00 2001-09-20T21:00:00 BADPIX NO
RGS2 BADPIX 0015 2002-11-04T00:00:00 | BADPIX NO
2 Changes
This release comprises two fundamental issues concerning bad pixels:
 a number of so-called "low count" columns was established by an analysis performed at SRON
with science data. Those columns are due to hot / warm pixels / segments by which on-
board processing already removes the events, but no removal registration is done. As a result
a number of spectroscopy channels have a lower eective area than expected, and produce
"outliers" in the spectral tting results,
 in November 2002 the operational temperature of the RGS CCDs was decreased from -80C to
-110C. This was performed mainly to reduce the eects of radiation damages occurred since
XMM-Newton launch. As a result a drastic reduction of hot pixels / columns was observed.
1

XMM-Newton CCF Release XMM-CCF-REL-130 Page: 2
Since the "low count" columns we want to ag as "advisory hot columns" (eg. not to be uploaded
but to be taken as bad in the data analysis) correspond to those observed disturbed permanently
since the beginning of the mission, a whole set of CCFs covering all periods has to be released.
3 Scientic Impact of this Update
The CCFs with indices 0012, 0013 and 0014 cover the whole period from begin of the mission to
the time the RGS instruments were cooled down, so they should replace completely all the RGS
BADPIX CCFs previously released. They contain for each period in time the same bad pixels /
columns they had in the former valid releases with the adding of the "low count" columns as advisory
hot columns. They replace for an actual analysis date the CCFs with indices 0005, 0006, 0007 and
0008 (using the start and end of validity keywords we could reduce the number of les to 3 with
following correspondence: 0012 for 0005, 0014 for 0007 and 0013 for 0006 and 0008). Please refer to
the corresponding former Release Notes for details on the uploaded and advisory hot columns and
pixels, which were already present in the CCFs 0005 to 0008.
Five "low count" columns in RGS1 and one in RGS2 are agged in the new CCFs (0012 to 0014)
as advisory hot columns. They have been determined using bright continuum source observations,
distributed more or less uniformly through the whole operational phase. They agged columns
showed up as "low count" in all of the observations analysed. The eect of such a correction is made
clear in the comparison of plots in Fig.1.
Figure 1: Left: detail of a tted spectrum without correcting for "low count" columns; right: same
spectrum tted taking them into acount
After cooling down the RGS instruments, the number of hot pixels decreased drastically. Hot
columns and pixels were determined using the RGS diagnostic data of two revolutions after cooling.
Results were:

XMM-Newton CCF Release XMM-CCF-REL-130 Page: 3
RGS1 : one hot column and ve hot pixels were found. Three hot columns have been uploaded
so far, one of them being the one found still to be hot after the CCD cooling. From November 27,
2002, only this one hot column will be uploaded. This makes necessary the release of two CCF les,
one covering the period between cooling of the CCDs and the operational change on Nov. 27, and
the other one from that date on. Both contain the same number (5) of advisory hot pixels (instead
of 28 hot pixels plus one hot segment, as prior to cooling). The "low count" columns cannot be
included until a new analysis on this item is performed.
RGS2 : one hot column and two hot pixels were found. The hot column found is the same
which has been uploaded continuously since September 2001. Therefore, no operational changes are
needed and only one CCF is needed, valid from the time the cooling of the CCDs took place for this
instrument. It contains the two advisory hot pixels instead of the 15 prior to cooling. Concerning
"low count" columns the same comment applies as to RGS1. First checks on RGS2 data indicate
that the "low count" columns have disappeared, as expected, due to cooling of the CCDs.
4 Estimated Scientic Quality
The inclusion of "low count" columns in the CCFs 0012 to 0014 should improve the quality of
spectral tting, since the eective area for the channels containing them will be properly calculated.
These channels appeared so far very often as outliers in the tted signal distributions, if a signicant
signal was expected.
Please take into account that:
 the "low count" columns are agged as if they were hot. When the SAS rgsbadpix task runs
(with parameter "withadvisory=yes", which is the default), it will check those advisory hot
columns and issue a "discrepancy" warning that the column was not found hot in the telemetry.
 only the permanent "low count" columns have been agged, eg the ones showing with a high
frequency this disturbance. There is a number of aected columns showing this eect only
from time to time, so that they cannot be agged this way. They show up also as outliers in
the spectral ts because of their reduced eective area.
5 Expected Updates
Updates could be expected after the analysis of bad pixels with science data post-cooling is done.
"Low count" columns should be re-established, if any, as well as those hot pixels / columns which
do not show up in diagnostic mode, but only in spectroscopy mode.

XMM-Newton CCF Release XMM-CCF-REL-130 Page: 4
6 Test procedures
General checks:
 use FV (or another ts viewer) for le inspection. It should contain 2 binary extensions
(BADPIX and BADPIX1)
 use the SAS task CALVIEW to see if the CAL digests and uses the new les.
7 Summary of the test results
The ts viewer fv was used to inspect all 9 CCF les, wrt their structure, validity dates and contents
of the rst extension (BADPIX). Everything OK.
The SAS task cifbuild was run several times using data corresponding to periods covered by the
dierent CCFs in order to check the correct selections. Selections were correctly done.
Finally the SAS task calview was used to prove that all these calibration les are ingested
correctly by the CAL, by pointing to the dierent Calibration Index Files and producing bad pixel
plots.