\documentstyle[12pt,a4]{article}
\title{On line analysis software requirements}
\author{Ph.Dessagne}
\date{Edition 1.1\\janvier 1991}
\begin{document}

\begin{titlepage}
{

\hoffset=1truein
\hsize=5.25truein
\vsize=10.25truein
\font\small=cmssbx10 at 14.4truept
\font\medium=cmssbx10 at 17.28truept
\font\large=cmssbx10 at 20.74truept
\hrule height 0pt
\parindent=0pt
%\parskip=0pt
\hskip 3.9truein
\large
EDOC109\par
\vskip .5truein
\large
Projet EUROGAM\par
\vskip 1.5truein
\hrule height 2pt
\vskip 20pt
\large
On line analysis software requirements\par
\vskip .5truein
\par
\vskip 20pt
\hrule height 2pt
\vskip 1truein
\rm
\small
\begin{tabbing}
edition \=: 1.1\\
date    \>: janvier 1991\\
auteurs \>: Ph.Dessagne \& C.Schuck\\
\end{tabbing}
\vfill
\bf
\medium
Centre Recherches Nucleaires Strasbourg \par
\vskip 5pt
C.S.N.S.M. Orsay
\vskip 5pt
CNRS-IN2P3 France
\par
\vskip .5truein

}
\end{titlepage}

\vskip 15cm
\begin{abstract}
  This note presents the preliminary reflexions on the on line event
 analysis of the french community involved in the EUROGAM project.
 It outlines the main requirements in that field and gives
 an indication on the conditions the event processor unit has to
 fulfill.
\end{abstract}

\vskip2cm
\vfill\eject
\bf
Description of typical on line analysis procedure during an experiment
\vskip0.7cm
\rm
The on line analysis has three main purposes:
\vskip0.3cm
\begin{itemize}
\item 1 Control of the various parameters of the experimental set up
\vskip0.3cm
\item 2 Checking of the quality of the data.
\vskip0.3cm
\item 3 Getting maximum available informations for the subsequent off-line
analysis.
\end{itemize}

\vskip0.7cm
 
The first stage can be done with the histogrammer by creating
individual raw unconditioned spectra (e.g. gamma-spectra
unsuppressed/suppressed, time spectra) as well as raw and Compton
Suppressed Ge folds and Ge and BGO Sum Energy.
(The energy spectra should be done on source, pulser, then with
on-line data).
 
In addition a selection of the Ge counter pattern may be of interest
 to the users.
 It is also necessary to have access permanentely to the rate of events
 at the aquisition stage.
 
Functions such as creating, displaying, zeroing, storing into memory
should be available.
 
The second and third stages should be done on the sorter and imply
the creation of a set of one and two dimensional spectra, depending on a
specific experiment.
It is essential to be able to build in direct mode and with multiple
conditions on
 the ADCs (energy, time, fold, sum energy...) different distributions
 like germanium energy, time, fold, sum-energy, fold versus
 gamma energy and sum- energy versus gamma energy. These informations
 should be available for different reaction channels selected either
 on criteria previously mentionned or by means of additionnal counters
 (recoil separator, particle detectors...).
 
It is clear that one should have during the experiment the possibility
 of analysing high fold events.
 We have in mind that this sorting processes only after gain matching
 for the Ge and selection by fold tresholds.
 
Due to a high number of spectra it is necessary to have a clear interactive
 display of the various sorting conditions based on lookup tables.
  First priority must be given to the constitution and the graphic
 representation of direct and conditionned one dimensional spectra.
  The second priority would be the constitution and the display of
 two-dimensional distributions on which we want to be able to define energy
 gates in order to obtain the corresponding one dimensional spectra.
 
\vfill\eject
\vskip0.7cm
A typical analysis procedure might imply the following different steps
on the sorter:
\vskip0.5cm
\rm
i)  Creation of two matrices Sum energy/Ge energy and Fold/Ge energy
which dimensions could be 64 x 4096 channels both. (2 Mbytes total on 4
bytes/channel)
 
\it It is to be noticed that those spectra could be done with only
one (or some already gain fit) Germanium detector.
\rm
\vskip0.3cm
ii) - It is important to know the time versus energy relationship
in order not to squeeze the low energy lines during the sorting. Therefore
45 128 x 1024 channels matrices should be created at some stages of the
experiment (e.g. begining and end). Those matrices do not have
to be stored in permanent memory but should be sent to a disk.
 
\vskip0.3cm
iii) - Create 45 spectra gated on Fold, Sum energy and time. Those
spectra should be stored and erased regularly in order to check the
data.
 
- The fitting of $\sim$ four peaks and calculating the corresponding gain
matching parameters with different options (e.g. linear, quadratic, including
or not the Doppler correction dependant on the different angles)
 might be done on a parallel work
station.
The resulting parameters should be sent back to the sorter in
the form of tables which will be used for creating the next spectra
which all imply gain matching. (Those spectra and control tables should include
identification tags from a counter
storing the block number informations at the begining and end of the
spectra construction. The results of the fit and new calibration parameters
should be stored in lookup tables which will constituted the logbook of the
experiment. These informations would be used for the final sorting.
\vskip0.3cm
iv) Sum of projected spectra for 4, 5, 6, 7,....folds Ge events
\vskip0.3cm
v) A set of Ge energy spectra (gated on Fold, Sum-energy, time
plus one or a sum of transitions energies)and the corresponding background gates.
\vskip0.3cm
vi) Different set of Ge energy spectra gated on various combinations of 
double or triple gates on Ge energies (e.g the energies of SD band).
It is to be noticed that a triple condition on the Ge
energies would imply considering only folds 4 and higher (such a
threshold would be asked anyway at the Event-builder stage for most
experiments) all these spectra should be created as a function of the
remaining Ge fold.
\vskip0.3cm
vii) One Gamma energy versus Gamma energy gated matrices (4096 x 4096
channels = 32 Mbytes on 2bytes/channel).
\vskip0.8cm
 
\bf
N.B.
\rm
The Ge spectra should be stored in subgroups according to their 
angular position.

The available memory for the sorter (actually 64 Mbytes) allows storage
for one-dimension spectra (2 to 16 Mbytes), both matrices in section i) 
(2 Mbytes) plus one Ge energy * Ge energy (4K x 4K =32 Mbytes) matrix
(section vii).
\vskip 0.8cm
 
 
 For all the spectra (one or two dimensional) obtained during
 the measurement we require the possibility to store both on disk and tape.
 A fast hard copy of the screen is also needed.
 
  It is clear that the data processing on and off line are strongly correlated
 but in a first stage it is crucial to have access to a minimum of graphic
 tools to be able to perform a reliable software analysis in order to monitor
 the quality of stored data.
 


 
\end{document}