Attending VZ,DH,RA,AM,YX,TA,BT,WX,MK
Topics:
-Luminosity
+Luminosity is beam current times target thickness
+Integrated luminosity is collected charge times target thickness
+In order to account for deadtime, we use "beamgated" current and
charge.
+Just as a comment: when we normalize our yields per combined
beam/target spin state to the collected charge in that state, we
make the assumption that the target density does not change between
those states, i.e. the ABS is assumed to be 100% efficient.
Any deviation from this assumption would show up in false
asymmetries and it is important to check this for every run and for
the sum of all runs.
+2004 Beamgated luminosity 2.7e31 / cm^2s at 95mA average current,
with 20-30% deadtime, 20min lifetime. This number is based on
analysis of quasielastic d(e,e'p) events, normalized MC cross
section. Here, MC-generated phase space with cross section weighting
of each event. Detector efficiencies are 1 in this approach,
i.e. for the real luminosity may be larger by 1/eff.
+The number agrees well with ET's number from ed elastic analysis.
+With a calibrated buffer system one can measure the same cross
section yield with the same detector efficiency for a precisely
known flux into the target cell (or target thickness assuming the
conductance), i.e. one can determine the target thickness
*absolutely* and relate it to the ABS target thickness. This has
never been done conclusively so far.
+ABS flux is ~4.4e16 atoms/s
+The current limitation of the ABS is due to the vaccuum pumping
efficiency. The record (at RHIC) is ~8e16 atoms/s.
-Discussion on comparison of MC and experiment in order to extract
observables:
+In order to extract an observable, e.g. a form factor, we measure a
number of counts in a certain Q2 bin. This yield is proportional to
the *average* differential cross section in that bin, with the
luminosity as a normalization factor. The yield distribution as a
function of Q2 has usually a dramatic dependence on Q2, even within
the narrow Q2 bin size that we usually pick. The reason is the Mott
cross section which goes like 1/Q^4, i.e. the cross section varies
in a bin from 0.1 to 0.2 (GeV/c)^2 by a factor 4. In addition the
real cross section drops even faster because of F(Q2) (e.g
dipole-like in case of quasielastic, or even faster in case of ed
elastic). Therefore the average cross section for a given bin does
in general not equal the exact value of the cross section at the
center of the bin (this is only if the cross section varies linearly
within the bin). In the present versions of generators we are
weighting each generated event with the interpolated differential
cross section for the tossed kinematics of each event, based on
Arenhoevels full model. In this way the cross section is averaged
over the considered bin for ideal kinematics.
+We discussed wether it is better to normalize the measured yields
event-by-event by the Mott cross section (which is calculable for
each single event) and then compare those yields to an MC that does
not weight the events with the cross section but with the form
factors only. The corresponding step can be done in pion production
by dividing out the virtual photon flux event-by-event from the
data. I brought this up because I wonder wether doing it this way
compared to averaging the MC with the absolute cross sections is
less vulnerable to systematic errors introduced by systematic
momentum misreconstruction. I'm not convinced by either approach and
would appreciate your comments.
The difficulty comes e.g. when the reconstruction of the momentum is
systematically off. We then build distributions of normalized yields
that are systematically off. It may be that it won't matter for
asymmetries but only for absolute cross section or form factor
measurements.
-Pion channels
AS presenting new results
Inclusive p(e,e') with improved cuts
MC of AS and TF agree well
Radiative tail effect of elastic events pulls the measured asymmetry
down in trigger type 1 (visible when compared to MC). In trigger type
7 however, elastic events are almost not present and this effect is
absent, i.e. data agrees with MC.
Showed false asymmetry in A_hp, is big in left sector for unpolarized
target (not understood)
(e,e'pi+) results are very clean, agree well with MC
The inclusive pion production also requires the combination of
trig=1,2,7. However, background are different in each channel, and may
even be polarized. Radiative effects are important. All of this should
be handled by the Montecarlo (pion production + radiative tail from
elastic scattering). The signal of one channel may be the background
in another...
-ep elastic:
CC presenting timing results for TOFs, GEp/GMp and hPz vs. Q2 once
with Q2 reconstructed with momentum, once with theta). Less Q2
dependence of hPz when Q2 is defind via angles.
Some deviations from Hoehler, need full MC
Regards,
Michael
--+-------------------------------------+--------------------------+ | Office: | Home: | |-------------------------------------|--------------------------| | Dr. Michael Kohl | Michael Kohl | | Laboratory for Nuclear Science | 5 Ibbetson Street | | MIT-Bates Linear Accelerator Center | Somerville, MA 02143 | | Middleton, MA 01949 | U.S.A. | | U.S.A. | | | - - - - - - - - - - - - | - - - - - - - - -| | Email: kohlm@mit.edu | K.Michael.Kohl@gmx.de | | Work: +1-617-253-9207 | Home: +1-617-629-3147 | | Fax: +1-617-253-9599 | Mobile: +1-978-580-4190 | | http://blast.lns.mit.edu | | +-------------------------------------+--------------------------+
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:32 EST