User name Templon
Log entry time 22:16:28 on October12,2000
Entry number 50228
This entry is a followup to: 49935
keyword=Notes on trigger setup for PY-2DThe notes in an earlier halog need to be modified somewhat. The new
tc variable (generated by running the new version of espace) is related
to the old tc variable by tc_new = 360 - tc_old. So the notes in
the old halog entries need to be adjusted for this, namely that making a
trigger adjustment moves the gate in the opposite direction as previously
stated.
.
However this should be simple for PY-2D ... check to see if the peak is
well centered (within 8 ns of center) in the gate. I expect it will be. See
the notes in the earlier halog referenced above. If it is not centered, there is only
one action you can take, which is to choose the next-highest momentum
setting when running trigsetup. Right now, we are running the highest
momentum setting in trigsetup.
.
To check, first go to the espace directory, then the subdirectory "header",
and create the necessary header file for PY-2D. You can use the one already
there for PY-2E as a template. You need to modify the two spectrometer angles, and the two spectrometer fields (take the locked magnet screen value and
multiply by 10.) Left is lepton, right is hadron.
.
The run espace, kumac file "ek/e97111.kumac". 100,000 events should
be enough.
.
After espace finishes, fire up paw and execute the macro "pk/checktc".
The cuts already in there should be OK. You may need to adjust the
cut on tc if you change the trigger. What you are looking for is the
peaks in the lower part of the plot (see below), relative to the accidental
plateau. You want the reals peak to be close to the center. The example
below is fine.
.
DO NOT BE FOOLED by the "hana" program or other paw kumac
files. There is a pulser which is always coincident and makes a
nice time peak. This can be seen in the upper left plot below ... that
apparently nice peak is just the pulser and tells us nothing about
the real event timing. This disappears if you ask for events which
reconstruct properly in the spectrometer (e.g. reactz or x_rot have
reasonable values), but will not disappear if you ask for e.g.
reaction vertex difference cuts (the "nonsense" value for each is
the same for pulser events, so the difference is zero).
.
ALSO, do not base your judgements on the corrected time
spectrum (tc_cor) ... we need to be in the middle of the _hardware_
bucket.
FIGURE 1
