[Prev][Next][Index][Thread]

[Fwd: Tagging problems]

Ciao Stefano,
    questo e' cio' che sono riuscito a recuperare dell'analisi di Kevin
sul problema del tagging: purtroppo sembra che i plot a cui fa
riferimento non ci siano piu' ma penso che eventualmente lui li possa
recuperare.....spero che ti possa essere utile.

Colgo l'occasione per fare a te, alla tua gentil consorte....ed alle
birbe gli auguri di Natale e di buon 2002!!!

                                                        Gianluca

Return-Path: <e831-error@fnal.gov>
Received: from infnmi1.mi.infn.it by axpmi.mi.infn.it with SMTP;
          Thu, 29 Jul 1999 4:41:34 +0200
Received: from FNAL.FNAL.Gov (fnal.fnal.gov [131.225.9.8])
	by infnmi1.mi.infn.it (8.9.1a/8.8.8/infnmi) with ESMTP id EAA12914;
	Thu, 29 Jul 1999 04:41:29 +0200 (MET DST)
Received: from fnpx08.fnal.gov ([131.225.53.34])
 by FNAL.FNAL.GOV (PMDF V5.1-12 #3998) id <01JE42KYXRGG000PDX@FNAL.FNAL.GOV>
 (original mail from stenson@fnal.gov); Wed, 28 Jul 1999 21:41:16 -0500 CDT
Received: from fnpx08.fnal.gov ([131.225.53.34])
 by FNAL.FNAL.GOV (PMDF V5.1-12 #3998)
 with ESMTP id <01JE42KYJ09O000RCO@FNAL.FNAL.GOV> for
 e831-expand@reprocess.FNAL.FNAL.GOV; Wed, 28 Jul 1999 21:41:06 -0500
Date: Wed, 28 Jul 1999 21:41:06 -0500
From: Kevin <stenson@fnal.gov>
Subject: Tagging problems
To: E831 Collaboration <e831@fnal.gov>
Errors-to: e831-error@fnal.gov
Reply-to: stenson@fnal.gov
Message-id: <Pine.SGI.4.05.9907281654280.24472-100000@fnpx08.fnal.gov>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII


At the last group meeting there was a question raised about
the validity of the beam silicon tagging data.  The problem
observed is the energy and sign of the particle found by the
beam tagging system seems to be uncorrelated with the sign
observed in the RESH/POSH or the total energy found be adding
the energy in the RESH (or POSH), BGM and J/Psi candidates for
the J/Psi -> mumu events from Skim 2.

One possible explanation is related to the trigger timing.  The
beam tagging system relies on the constant timing of the 2nd
level trigger.  Unfortunately a few days before the end of the
run it was discovered that the timing of the 2nd level trigger
depended on the trigger (cf.
http://www-focus.fnal.gov/dario/archive/focus/1244.html).  
The only triggers unaffected are trigger 1 (standard hadronic
trigger) and trigger 2 (dielectron trigger).

After performing a brief study on raw data from file 2 of runs
14061, 14062, 14080, and 14081, I have concluded that this is 
correct explanation.  The study can be found at 
http://www-focus.fnal.gov/stenson/timing_bad.ps.
The first page shows the number of reconstructed tagging tracks
by trigger.  For triggers 1 and 2 the distributions look fine.
The remaining triggers show much too few tracks. Following this
page are 8 more pages (one for each trigger).  There are four 
plots on each page.  Clockwise from the top left we have:
a) Energy in RESH (excluding RESH0 and RESH1) for events which 
have 1 beam-tagged negative track and a RESH signal consistent 
with 0 or 1 electron.
b) Energy in POSH (excluding POSH0 and POSH1) for events which
have 1 beam-tagged negative track and a POSH signal consistent
with 0 or 1 positron.
c) Energy in RESH (excluding RESH0 and RESH1) for events which
have 1 beam-tagged positive track and a RESH signal consistent
with 0 or 1 electron.
d) Energy in POSH (excluding POSH0 and POSH1) for events which
have 1 beam-tagged positive track and a POSH signal consistent
with 0 or 1 positron.

If the system is working correctly we expect energy in the RESH
to be correlated with electrons and energy in the POSH to be
correlated with positrons.  Thus, we would expect plot (a) to
show more non-zeros than (b) and plot (d) to show more non-zeros
than plot (c).

This is exactly what is seen for trigger 1 and trigger 2.  However,
the other triggers (4,5,6,8,9,11) show essentially the same RESH
(POSH) distribution independent of the charge of the incoming lepton
identified by the beam silicon tagging system.

As a further confirmation, the file
http://www-focus.fnal.gov/stenson/timing_fix.ps shows the exact same
plots for data taken from file 2 of runs 14520, 14521, 14522, 14523
which is after the problem had been fixed for triggers 4, 8, 9, and 
11.  However, the intensities are quite different (factor of 3) for
the two run periods so you comparisons between them should be made 
with care.  Also, it looks like trigger 9 did not get fixed.

The upshot of all this is that the beam silicon tagging information
is effectively unusable for any triggers other than the hadronic 
trigger and the dielectron trigger.  This explains the problems seen 
by Peter G (Peter K) who were using triggers 4,9,11 (5) for their
studies on J/Psi->mu+mu- (e+e- pairs).  However, the J/Psi to e+e- 
is effectively triggered on by the dielectron trigger.  Therefore,
studies of J/Psi which REQUIRE good beam energy resolution (a few %) 
should use the J/Psi -> e+e- mode.  If a 15% incoming beam energy
resolution is acceptable, the J/Psi -> mu+mu- can still be used with 
the average beam energy of 282 GeV.

I close with a heavy heart over the sorrow I'm sure this news
brings.  I hope we can emerge from this tragedy with a new sense
of purpose and that can-do winning attitude.

Or not.

Kevin



P.S. An interesting side note.  The timing was "fixed" for triggers 9 
and 11 by using the UNprescaled outputs instead of the normal prescaled
outputs.  Therefore, for this data the standard trigger bits for 9 and
11 are always off.  The unprescaled outputs are not saved on DST.
In addition, trigger 8 (prescaled MG1) was moved to trigger 3 during 
this time (run 14452-end).