Full release note for ppsprod-12.10
===================================

Pipeline Version: 12.10_20140207_1200
Pipeline Name: ppsprod
Release Date: 2014-02-07
SAS Manifest: xmmsas_20121219_1645



Summary
=======

This pipeline represents a major change with respect to the previous
prod8.4_64 pipeline, embodying both a substantial number of changes to
the PCMS processing system itself and incorporation of version 12.0.1
of the SAS (with additional package updates).

The most significant changes to the pipeline include:

i) Use of an empirical 2D PSF. This PSF provides a much improved
description of the polygonal core structures of each of the
instrumental PSFs and also the spoke patterns, compared to the
previous Medium accuracy PSF - see Read et al., 2011, A&A, 534, 34
(http://adsabs.harvard.edu/abs/2011A%26A...534A..34R)

ii) Extraction of source-specific products (SSP) (time-series and
spectra) for individual sources is redesigned. Firstly, the size of
the (circular) source extraction aperture is now optimized (using
eregionanalyse) to yield maximum signal-to-noise - previously a fixed
aperture of radius 28" was used. The aperture is restricted to a
radius (r) where 12" < r < 40". Secondly, the previous annular
background region (60" < r_back < 180", co-centred on the source
position) is replaced by a circular region of comparable area but it
is now offset from the source so as to ensure there is sufficient
background signal in all circumstances. A minimum of 70% useful area
is required, except in exceptional circumstances. This, in particular,
resolves problems found with negligible background data being
available for sources in MOS SW modes in earlier processing.

Spectra and time-series are now extracted for all sources whose
spectra (source plus background) contain > 100 good counts (cf the
original process which demanded > 500 EPIC counts).

iii) The background flare filtering now uses a S/N optimization
approach and operates on an in-band (0.5-7.5 keV) light-curve from
which prominent sources (eboxdetect likelihood > 50) have been masked
out. The algorithm discards, in declining order, high count-rate
time-bins, computing the S/N of the residual data, to find the cut
rate at which the S/N peaks. This can result in significant increases
and decreases of usable exposure time. The flare background
time-series is now exposure corrected and the file contains the S/N v
rate-cut curve. The PDF graphic also shows the S/N v rate-cut curve.

iv) Astrometric correction is now performed by the task catcorr. This
makes use of a number of enhancements over the previous task,
eposcorr. A notable change is its ability to make use of multiple
reference catalogues.  In the pipeline the task catprep now generates
a file that contains upto 3 reference catalogue extracts (SDSS-DR9
where available, USNO B1.0 and 2MASS).  Each, where available, is used
to determine the field offset correction and the 'best' solution is
adopted for the correction if it meets the application criteria. The
task also employs an algorithm that takes account of the local density
of reference sources and performs fitting to find the solution rather
than a grid technique. In principle, use of the 2MASS catalogue allows
the process to find solutions in more fields, particularly in the
galactic plane, than the previous USNO B1 -only approach.

v) A number of astrometry issues have been addressed since 2XMM. In
particular, a time-dependent boresight (CCF) is now employed that
removes a 1-2" time-dependent pointing variation. This is important in
fields that can not be corrected by catcorr.

 vi) Masks are generated to identify OOT and scattered RGA features
from piled-up sources. These masks are used to 1) flag detections
centred within the mask and 2) to avoid locating SSP background
regions on them.

vii) RGS spectra are now computed in wavelength space

viii) RGS time-series are generated for the first time.

ix) The noisy rows in MOS1 CCD#4 caused by the event on December 11,
2013 are excluded from full processing, as they resulted in a
significant number of spurious source detections.

The version of SAS being used is a pipeline-specific manifest
xmmsas_20121219_1645. This is based on the public SAS release 12.0.1,
described at
http://xmm.esac.esa.int/sas/current/documentation/releasenotes/xmmsas_12.0.1.shtml.



Documentation
=============

Documentation for this pipeline can be found at:

http://xmm.esac.esa.int/external/pipeline/

and links therein.