Next Previous Directory
Folder: ALEPH.MEETING
Subject: Minutes of the luminosity meeting 17.5.93
From: TFE@CERNVM
Date: 21-MAY-1993 11:20:06, Expires: 31-JUL-1993 23:59:59
MINUTES OF THE LUMINOSITY MEETING 17.5.93
=========================================
1. REVIEW OF SICAL 1992 RESULTS (B. Bloch)
---------------------------------------
Brigitte Bloch gave a brief summary of the SICAL 1992 luminosity
analysis, already presented in detail on May 7th during the ALEPH
week. The results are now regarded as final. 613 runs (EW-selected
with SLUMOK) have been analysed. Using Method 2, now the standard
method, the luminosity is measured to be 8748.8 inverse nanobarns.
The combined error on this measurement is 3.15 permille, comprising
statistical (1.16 permille, 737042 M2 events), MC statistical
(1.20 permille, more than 2 M events), experimental systematic error
(0.93 permille), and the theoretical error (2.50 permille).
The number of hadrons used for the EW fit is the average between
the TPC and calorimeter count (267337.6 events, 30.557 nb). The
statistical error is 1.93 permille and systematic error is
2.0 permille.
The EW fit (M.Martinez) then gives sigma(0) = 41.559 +- 0.179 nb,
and 2.980 +- 0.038 neutrinos.
2. STATUS OF LCAL 1992 SICAL-PERIOD LUMINOSITY ANALYSIS (T.Fearnley)
------------------------------------------------------------------
T.Fearnley gave a brief status report on the LCAL luminosity analysis
for the SICAL-period 1992. The aim of this analysis is to calculate
the additional systematic error due to SICAL shadowing, and compare
directly the LCAL and SICAL luminosities.
An update from the ALEPH week was presented. 250 K events have
been processed so far by KINGAL, GALEPH and JULIA. The MC is done on
SHIFT6. Another 750 K events can be generated within 3 weeks.
The need for 1 M events was however questioned; an argument in
favour of this number is that 1 M events would give the same
systematic error as for pre-SICAL 1992, only with the addition of
the extra SICAL-related error which is being calculated.
The Bhabha selection will use Method 9 with an additional cut on the
energy fraction in storey 1 (removing events with particles showering
through the SICAL edge). The SICAL-related error is being estimated
by varying this cut. The other "traditional" systematic errors
are expected to be the same as in the pre-SICAL period.
3. LCAL AND SICAL GLOBAL ALIGNMENT ERRORS (T.Fearnley, B.Bloch)
------------------------------------------------------------
The global alignment uncertainty for LCAL is assumed to be 1 mm,
which is the error quoted by the survey. Its contribution to the
luminosity uncertainty is only 0.1 permille, which can be neglected.
Even if the alignment uncertainty were to be revised upwards to 2 mm,
the contribution to the luminosity uncertainty would only grow to
0.4 permille, still negligible.
SICAL is much more sensitive to global alignment errors, because
the inner radius of the fiducial region is only half of LCAL's.
It is at the moment not clear what the survey uncertainty for SICAL
is. This should be clarified. However, the global alignment of SICAL
is determined offline by the asymmetry distribution. The position
uncertainty is estimated to be approximately 0.5 mm. Assuming
this error, the contribution to the systematic error should be around
0.1 permille, as for LCAL.
4. ESTIMATES OF LCAL SYSTEMATIC ERRORS AND CORRELATIONS FOR
PRE-SICAL 1992, 1991 AND 1990 (J.D.Hansen)
--------------------------------------------------------
The NBI group has determined a consistent set of systematic errors
for the LCAL luminosity for the years pre-SICAL 1992, 1991 and 1990
(ALEPH 93-006 PHYSICS 93-03). The error estimates are final.
The fluctuations of the fiducial side cut error
(which contributed most to the systematic error in 1990 and 1991)
are thought to be manifestations of a common systematic error of
statistical nature, which fluctuates with a sigma of 0.21 % around
a mean value of 0.04 +- 0.16 %. These values give a chisquare of
3/3, when fitting the measured fiducial side cut errors for the
different years and generators. The theoretical expectations for this
error has been accounted for. The source of this error is not known;
however, different amplifiers and background are candidates.
For the smaller contributions to the systematic error, the mean
between 1991 and 1992 has been used.
This statistical analysis of the various LCAL systematic errors
has led to new values for the experimental systematic errors and
for the correlations between the years. Also, the "rounding up" of
certain errors has been corrected for. J.D.Hansen argued strongly
for using the new LCAL estimates in future EW fits, and for
using these fits in future publications. There were different
opinions on this. One criticism, for instance, was that the source of
the large fluctuation of the fiducial side cut error (0.21 %) is not
yet identified.
No decision was made concerning the implementation of the
new error estimates. However, Manel Martinez will perform an EW fit
using the measured (not fitted) experimental systematic errors
for pre-SICAL 1992, 1991 and 1990, using the FINAL value for
pre-SICAL 1992 (0.32 %).
5. THE ROLE OF LCAL IN 1993 (J.D.Hansen)
-------------------------------------
The SICAL Bhabha cross section is approximately 5 times larger than
that of LCAL, so the SICAL statistics is 5 times the LCAL
statistics. Moreover, the LCAL experimental systematic error is
about a factor sqrt(5) larger than that of SICAL. This means that
in EW fits, the weight of SICAL relative to LCAL will be 5/1,
approximately, and consequently LCAL will improve the SICAL
measurement with only about 8 %. Such a limited improvement, in fact
within the uncertainty of the SICAL systematic error, is
small compared with the effort needed to measure the luminosity in
1993 and thereafter. J.D.Hansen therefore proposed, on behalf of
the NBI group, to put LCAL on STANDBY for luminosity, with immediate
effect. This means that LCAL will not measure the luminosity in 1993.
However, the NBI group is willing to resume the luminosity
measurement if, and only if, requested to do so by the collaboration.
LCAL will continue to act as an electromagnetic calorimeter.
Bhabha triggers will be needed for calibration purposes.
The question of downscaling of LCAL triggers was then reviewed.
The single-arm trigger (LW_A+BVH, now in bit 7) can be downscaled.
The factor will be decided later, after discussion with other
potential users of the trigger (are there any ?). Concerning the
main Bhabha trigger (LW_ET_HI, bit 6), the downscaling depends on
the degree of STANDBY which LCAL will go to: For an ACTIVE STANDBY
(LCAL capable of measuring the luminosity for data already taken),
the best would be not to downscale at all. For a PASSIVE STANDBY
(LCAL capable of measuring luminosity for data not yet taken), a
downscaling factor of maximum 10 could be used, since the downscaling
factor could always be set back to 1.0. A decision on the degree of
STANDBY has not yet been taken.
THE LUMINOSITY GROUP AGREED TO RECOMMEND TO THE COLLABORATION TO
PUT LCAL ON STANDBY.
Tom Fearnley
Next Previous Directory