Katherine and I yesterday have lowered the trigger scintillator thresholds from 37 mV to 20 mV, and raised a few of the high voltages from their earlier values. This was to try to ensure that we are on the efficiency plateau for these counters.
While we are not completely convinced that the present settings are, indeed, at the full efficiency plateau, we do know that the new settings have substantially large beam-off mean-timer rates for both bars than before. The effect is much more dramatic than the modest fractional increase in rates when we have 25 nA of beam on target.
This may mean that we are now partly in the true noise (tube noise, light leaks...), or that we are becoming more efficient for room background and cosmic events that deposit less than does a m.i.p. electron traversing the scintillator at the nominal angle.
Ordinarily, a modest increase in beam-off rates would not be something to be concerned with. But, because we need to run with 50pA scale currents when we have Region 2 on, this means that the triggered events are dominated by cosmics/backgrounds when we are at 50 pA.
I looked at a 20,000 event sample of run 8662, and asked how many of these events had a "MD trigger" in either of the two active octants (2 and 6 in this run). I used the md2m_f1, md2p_f1, md6m_f1 and md6p_f1 histograms (leafs) to determine
The results are:
Octant 2: 1205 events Octant 6: 1157 events Total : 2362 events
This implies that, roughly, 2362/20,000 = 12% of the triggers are likely to be "real" elastic electron candidates.
So, to get 100,000 good tracks, we need close to a million triggers.