A current worry for an "extended" run (1 week) is the stability of the
atomic source and the issue of spin-dependent normalization. If this is
not constant for different spin states then we are in high seas again.
In the attachment I show normalized tof rates (counts/C) integrated over
a number of ee'p timing peaks (independent of reconstruction) as a function
of run #. As was mentioned some time ago you essentially get this monitor
online by typing ".rate ###". Certainly there are time-dependences which
we should understand at some point! The bottom panel is ligit.
This rate is a monitor of the target "density". Note that here molecules
and atoms are weighted "equal" (i.e. 1 molecule = 2atoms, which does not
have to be true for their polarizations). This yield of d(e,e'p) events
is the fastest counting reaction directly related to the gas in the target
cell. To the extent that we are insensitive to the polarization value with
such a broad selection, we can use this as a monitor also for the ee'p
analysis (note: e,e'p is a monitor in the truer sense only for ee'd)
The color code corresponds to the beam/target polarization states
(nominally, beam = +-, target-tensor = +1,-2)
I then extract the average normalized yield/run in 3 ways:
1) the average norm-yield weighted by the number of counts (this I think
matters the most for the asymmetry analysis)
2) The averate norm-yield weighted by the inverse square of its error bar
3) just the average normalized yield (paw fit, with error bar: it
essentially norm-yield weighted by its sqroot)
Beam-Tgt state 1 2 3
+ + 18.51 18.37 18.37+-0.96
- - 18.33 18.06 18.04+-0.81
+ - 18.26 18.09 18.09+-0.90
- + 18.45 18.31 18.31+-0.81
In conclusion:
the false asymmetry due to different luminosities seems small, although
the uncertainty is at least 5%. There are other systematics involved if the
the *atomic fraction* changed over time, since that changes target density
at a given flow and is may very well change the true polarization.
In the short term Ligit is consistent with small differences in
target densities but in the long run ligit becomes an almost random
number
And oh well, yes, maybe (????????) there is a T=-1, T=+2 dependency
in these very raw yields (of the order of 1%). Hopefully one day it
will be bigger and calibrated.
-- ________________________________________________________________________________ Tancredi Botto, phone: +1-617-253-9204 mobile: +1-978-490-4124 research scientist MIT/Bates, 21 Manning Av Middleton MA, 01949 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:29 EST