Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://xmm.vilspa.esa.es/external/xmm_sw_cal/sas_sci_val/reports/SAS52_report.ps.gz Äàòà èçìåíåíèÿ: Tue Oct 8 14:47:38 2002 Äàòà èíäåêñèðîâàíèÿ: Sat Dec 22 14:58:56 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: îáâìàäåîéñ íåôåïòîùè ðïôïëï÷

XMM
XMM-Newton SAS scientic validation - add for SAS V5.2
XMM-SOC-USR-TN-0007 Issue 1.0
C.Gabriel 1 , C.Erd 2 , M.Guainazzi 1 , M.Denby 3 , J.Osborne 3
November 8, 2001
1) XMM-Newton Science Operation Center, VILSPA, ESA, Apartado 50727, E-28080 Madrid,
Spain
2) XMM-Newton Science Operation Center, ESTEC, ESA, 2200AG Noordwijk, Netherlands
3) XMM-Newton Science Survey Center, Department of Physics & Astronomy, University of
Leicester, Leicester LE1 7RH, UK
i

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-USR-TN-0007
Issue/Rev.: Issue 1.0
Date: November 8, 2001
Page: ii
Contents
1 Introduction 1
2 Input data 1
3 Backward compatibility 1
4 RGS response matrix generation checks 1

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-USR-TN-0007
Issue/Rev.: Issue 1.0
Date: November 8, 2001
Page: 1
1 Introduction
SAS V5.2 was released publicly on September 20, 2001. Detailed release notes pointing to all
the dierences to V5.1 are given in the SAS documentation page
(http://xmm.vilspa.esa.es/sas/documentation/releasenotes/xmmsas 5.2.0.shtml). The main aims
of the new release were:
 ability to process PN data in extended Full Frame mode,
 backward compatibility of RGS processing, so that "old" event les produced with
former SAS versions than V5.1 could be used as entry point,
 xing of the problems related to the RGS response matrix generation, aecting
largely RGS data above 1 keV.
This document addresses the checks done previously to the public release of SAS V5.2. For all
the items not covered by the present document, the reader is referred to Gabriel et al. 2001.
2 Input data
Several datasets with PN data (under proprietary rights) in extended Full Frame mode have
been processed through the pipeline and found OK by data screening.
In addition, Mkn421 RGS data processed with pipeline SAS V5.0 were taken for interactive
analysis to prove the backward compatibility of SAS V5.2.
Mkn421 RGS data processed with pipeline V5.1 were taken for checking the response matrix
generation. Since this happens outside the pipeline processing it was not necessary to re-process
the raw data with SAS V5.2.
3 Backward compatibility
The results using the recipe provided in the watchout pages of the SAS documentation http://xmm.vilspa.esa.es/sas/documentation/watchout/rgs backwards.shtml
were completely satisfactory.
The basic steps are to convert all the event lists using a new task "rgsevconvert", to convert all
the source lists using the task "rgssources", to delete ("dsrm") and regenerate the background
region extension using "rgsregions" and nally to run "rgsspectrum" on the corresponding source
in the source list for extracting spectra for the orders desired.
A comparison of the results achieved using converted SAS V5.0 data and SAS V5.1 produced
data of the same Mkn421 observation showed remarkable dierences in the response matrices.
This was expected due to slightly dierent data selections as a consequence of new calibration
information in the pipeline processing. However, the nal spectra were almost identical, both
for rst and second order of RGS1 and RGS2 data, proving the backward compatibility.
4 RGS response matrix generation checks
As a rst test, we have checked the new version of the rgsrmfgen task against carefully con-
structed and checked reference matrices. These were derived with a development SAS version
based on SAS V5.0. The overlap of these and the interactively derived response matrices run-
ning rgsrmfgen on a dierent data set was very good for both RGSs in rst and second order,
as shown in Figs.?? and 2.

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-USR-TN-0007
Issue/Rev.: Issue 1.0
Date: November 8, 2001
Page: 2
The Mkn421 data reduced with the pipeline corresponding to SAS V5.1 were re-analyzed with
XSPEC using response matrices derived with SAS V5.2. The results of tting simultaneously
RGS1 and RGS2 data with a broken power law model leaving normalisation free between both
are shown in Fig.3. In the best t a 10% larger normalisation factor for RGS1 data is forced by
the data corresponding to the low energy end. This is re ected in the large discrepancies seen
in the high energy part of the spectrum. This is consistent with ndings of the RGS instrument
team. To solve the RGS1-RGS2 dierences a phenomenologically  dependent correction to the
RGS1 data has been proposed, ranging from 1.00 at the high energy end of the spectrum to 1.25
at the lower end, which should compensate for the observed lower eôciency. This correction is
expected to be part of the improvements in the SAS V5.3 version.
In the detailed comparison of the low energy part of the spectrum we see that there is no
discontinuity in RGS2 data between E = 0:37keV and E = 0:44keV . This discontinuity, shown
in the SAS V5.1 validation report (Gabriel et al. 2001) was due actually to the use of a wrong
calibration le. The calibration defect behind this artifact (as seen and reported in the validation
of SAS V5.0, Ehle et al. 2001) could be understood as an additional SiO layer on CCD2 of RGS2
and actually xed previously to the SAS V5.1 release.
Another defect clearly seen in the plotted data is the instrumental O-edge at 0.55 keV. This has
been by now calibrated in detail by the RGS instrument team and it is also expected to be xed
in the SAS V5.3.
References
Ehle et al (2001), XMM-SOC-USR-TN-0003, available at
ftp://xmm.vilspa.esa.es/pub/odf/data/sv/0099280101/0099280101 report.pdf.gz
Gabriel et al (2001), XMM-SOC-USR-TN-0004, available at
ftp://xmm.vilspa.esa.es/pub/odf/data/sv2/SAS51 report.ps.gz

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-USR-TN-0007
Issue/Rev.: Issue 1.0
Date: November 8, 2001
Page: 3
Figure 1: Comparison of reference and interactively derived 1st order spectral responses. In
white and red the RGS1 and RGS2 reference eôciencies, in green and blue the corresponding
distributions derived from the Mrk 421 data set

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-USR-TN-0007
Issue/Rev.: Issue 1.0
Date: November 8, 2001
Page: 4
Figure 2: Comparison of reference and interactively derived 2nd order spectral responses. In
white and red the RGS1 and RGS2 reference eôciencies, in green and blue the corresponding
distributions derived from the Mrk 421 data set

european space agency
agence spatiale europÈenne
XMM Science Operations Team
Document No.: XMM-SOC-USR-TN-0007
Issue/Rev.: Issue 1.0
Date: November 8, 2001
Page: 5
.
Figure 3: Mrk 421: RGS1 and RGS2 data / model ratio from simultaneous tting, covering the
whole 1st order spectra (above) and the low energy part zoomed in (below)