As of 21:15 Apr. 4 I have updated the EC ADC
calibration bank (ECCA.bfp) in $CLAS_PARMS
with new adc calibration constants derived
from analysis of MIP tracks in the EC.
The previous analysis demanded a single pixel
in both inner and outer calorimeters. This
produced cleaner MIP peaks but with limited
statistics, preventing me from fitting the
attenuation lengths to estimate the PMT gains.
Instead I corrected the 'mean gains' for light
attenuation using test lab cosmic ray measurements.
In the present method I don't demand single pixels,
which gives me enough statistics to do fits, at
the expense of less certainty that I'm looking
at MIP events. Also I've used energy cuts to
clean up some of the low energy junk I mentioned in my
web page.
'Mean' gains and 'Fitted' gains are fairly consistent
for the long strips, with a few systematic
shifts which I don't understand. For short strips
(1-10) in some cases large differences are seen. However,
I believe the fitted results are more accurate.
In any case, having exhausted the statistics of the MIP
runs, the errors on the fitted gains will not improve
beyond 5-10%. Furthermore, there are systematic effects
which are exaggerated by poor gain-matching. Therefore,
several cycles of HV changing and fitting are going to be
necessary, hopefully at the beginning of the next run
cycle. Until then, more cosmic-ray runs.
Cole Smith
cole@apollo.phys.virginia.edu