Hi,
it seems to me that we have about 1ns difference in the delay in left and
right sectors.
atatched are plots of reconstructed mass spectra from unpolarized
deuterium runs. red curve is positive charges into left sector, blue curve
is positive charge into right sector.
the bottum plot is the mass spectra reconstructed by lrn. there is huge
shift in the two (proton)mass peaks, 190MeV away.
since momentum and track length reconstruction in the two sectors are
rather symmetric, the most possible reason for this shift is a different
delay in the two sector.
note that for a positive charge into left sector for example, mass is
calculated by: m = p/(gamma*beta) where beta = L/T and in turn T = ttl -
ttr + reconstructed electron time of flight.
the difference in delay enters into ttl - ttr. 1ns difference corresponds
to ~10% of electron time of flight and ~5% of proton time of flight. so
mass reconstruction is marginally sensitive to 1ns errors.
the figure on the top shows the mass spectra after subtracting an extra 20
channels offset in one sector. the two (proton)mass peaks are aligned.
there is some funny shape which could come from the fact different paddle
combinations have different energy loss for instance, so the big peak is
in fact the sum of many small peaks at different location.
The impact of this difference is more than just a shift in mass
reconstruction. Reconstruction and particle ID work at their best when
coincident trigger is activated by an electron. when the delays in left
and right are different, program could be misled to consider a pi+/pi- as
the trigger particle instead of the electron.
Please let me know what you think.
Chi
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:29 EST