Return-path: <boiarino@CEBAF.GOV>
Received: from cebafb.cebaf.gov (cebafb.cebaf.gov)
by CEBAF.GOV (PMDF V5.0-5 #9103) id <01I74V349R2O90N65T@CEBAF.GOV> for
BROOKSW@CEBAF.GOV; Tue, 16 Jul 1996 08:51:14 -0500 (EST)
Received: by cebafb.cebaf.gov (1.37.109.16/16.2) id AA222371469; Tue,
16 Jul 1996 08:51:09 -0400
Date: Tue, 16 Jul 1996 08:51:09 -0400
From: Serguei Boiarinov <boiarino@CEBAF.GOV>
Subject: hi, it's my report
To: BROOKSW@CEBAF.GOV
Message-id: <01I74V34CG3690N65T@CEBAF.GOV>
Content-transfer-encoding: 7BIT
Hi Will,
my work during last time in ITEP:
1. Installation all CLAS software on HP.
2. BOS output implementation in EC.
3. Set of tests development for EC.
Last point in progress now. I remember from my previous experience that
for better understanding we should to compare reconstruction energy for
particular cluster not with (initial energy)*(sampling fraction) but
with real deposited energy. For this I changed some GSIM calorimeter
routines and ECH bank has additional variable EDEP. It's not a real deposit
energy for this particle, it's energy, deposited in some area around
entry point ( 3*(molier radius) for example, should be GSIM parameter ).
Present algorithm (as well as any simple algorithms) can't reconstruct
energy which deposited far from main shower as result of some shower
branch, so we can try to reproduce energy deposited closely to entry point.
It's good enough for main part of events. It's important also that
GEANT threshold not important if we are using EDEP for testing.
New version of GSIM are working in ITEP, I wanted to install it at CEBAF
from here but then decided to do this when I'll be there.
Taking into account this changes I'm using EDEP for program testing.
First test - 1 e- hit per sector - clear program test, to check program after
every change. This test contains histograms: EDEP-Erec, X-Xrec, Y-Yrec where
first values from ECH bank and second ones from reconstruction
program. GSIM file for this test are produced without any other detectors
and avoiding boundary effects ( only central part of sector ). It's enough
to use one electrons energy, 1 GeV for example.
Second test - double hit situation with the same simple conditions. For such
cases we'll never have 3 hits in one strip and simple energy recognition
algorithm can be used.
Third test - 3 hits per sector. More complicated energy reconstruction
method should be used but sometimes it can give us reasonable result.
Set of these 3 tests - first step for EC testing. If something wrong we
will obtain it here, simple initial conditions permit us to figure out
what the problem.
Second step - we are changing one of initial conditions: particle energy,
particle type, permit hits to be closely to sector boundary. I'm not
expecting to obtain any errors here but these tests can show us what
could be improve in program ( some additional analize etc ).
Third step - real situations : include real geometry, real background, real
physical task ( pi zero decay, e/pi rejection etc). Here we can just obtain
that our program good enough or not. Detail program analize under such
complicated conditions not very effective. We can just say for example:
50% of our events reconstructed very good. And then try to increase this
value changing program and going to step 1.
Of course working time should be taking into account also. And time from
calorimeter, I skiped it temporary.
I'm changed GSIM and begin to test program using test 1 and planes to come
to CEBAF in 2 weeks with 3 working tests and some changes in EC program.
Regards, Serguei