Hello,
there was a lot of excitment about some weird results shown over the
weekend. A number of things went wrong.
The target polarization for runs taken Friday->Saturday (after gap studies)
was ok (1st attachment). Starting from saturday afternoon we used again
the spin flipper. Things are not so good (2nd attachment)
The problem has to do with the fact that the spin flipper/compton assumes
that a new run starts with a new fill, in other words it resets the "flip"
variable to its even state. However this is not correct when a run is
started in the middle of a fill, after a flip. This is not really a bug:
it is hard to define what the beam-helicity flip status was before the
current run, given that two neighbouring runs may start on different days !!
The simplest solution is to start a new run with a new fill, or to
* DUMP THE BEAM BEFORE STARTING A NEW RUN *
then everything will be fine.
Such procedure would also good for the fill number variable, which would
then become "local" to a given run only. An example is shown in the 3
attachment: during fill number 7317, between times 3057.1 and 3057.2
(x10^4 seconds since jan 01 2003....), the beam helicity is flipped at ca.
t = 3057.155, has shown by the "helicity flip" variable in the bottom
panel (in black) which goes from + 1 to -1 (actually shown is from +0.9 to
-0.9). In red, the beam helicity from the polarized source, which does not
change value for the whole fill number. In blue the beam current
profile (scaled by 1/500.)
At t = 3057.18 run # 4133 ends, and new run #4134 starts. Although this is
still fill # 7317, now the helicity flip variable defaults again to its
"even" value (+1.0 or +0.9) and the charge between t =3057.18 and 3057.2
will end up in the wrong spin state.
We ended up in a bizarre situation, where fill no 7317 is a "good" fill in
run 4133 but is a bad fill in run 4134. Incidentally, the following fill,
fill # 7318, is a good fill and is fully contained in run 4134.
Since a fill is truly the real quantum of our experiment, we should
remember to dump the beam before starting a new run. It should be simple
enough. One can easily end the run right at the end of a fill to maximize
efficiency !
So when I patiently hand-pick (for a few run...) those fills that did not
have the above problem the "right" asymmetry re-appeared (last attachment)
ALSO:
Please see runlist and elog entries 18659, 18563 about what macro/cuts
to use for analysis of runs with the spin-flipper on. I think you MUST
read ALL of the e-logbook before making ANY sense of your analysis.
The analysis directory stays pro2003/analysis
-- ________________________________________________________________________________ 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 - 14:07:30 EST