Difference between revisions of "GlueX Start Counter Meeting, September 21, 2017"

From GlueXWiki
Jump to: navigation, search
(Add the minutes.)
 
(No difference)

Latest revision as of 21:05, 1 November 2017

GlueX Start Counter Meeting
Thursday, September 21, 2017
11:30 am EDT
CEBAF Center, Room F326/327

Agenda

  1. Announcements
  2. Review minutes from the last meeting
  3. Calibration and Efficiency
  4. Time-dependent geometry adjustments
  5. Mcsmear procedure and parameters for the start counter
    1. Message from Sean on separating tracking contribution to resolution in mcsmear
  6. NIM/TDR
  7. Action Item Review

Communication

BlueJeans

Slides

Talks can be deposited in the directory /group/halld/www/halldweb/html/talks/2017 on the JLab CUE. This directory is accessible from the web at https://halldweb.jlab.org/talks/2017/ .

Minutes

Present:

  • FIU : Werner Boeglin, Mahmoud Kamel
  • FSU : Sean Dobbs
  • JLab : Thomas Britton, Mark Ito (chair), Simon Taylor, Beni Zihlmann

There is a recording of this meeting on the BlueJeans site. Use your JLab credentials to authenticate.

Review of minutes from the last meeting

We went over the minutes from August 24.

  • Thomas suspects that here may be a clocking problem with the start counter in HDvis that needs to be looked at. Mahmoud expressed interest in playing with HDvis with an event source. Thomas will send him email with instructions.
  • Werner recently sent some final plots to Eric Pooser for the NIM paper. The FIU folks want to get together one more time before general release, but Werner thinks they are pretty close now.

Calibration and Efficiency

Mahmoud brought us up-to-date on his efficiency analysis. Please see his slides for the plots and details.

He has changed the procedure to use the RF signal as a time reference for the start counter hits in his efficiency calculation, cutting the resulting time difference to decide whether a counter is hit or not. He is correcting for the target z position of the event vertex, the time-of-flight from the target to the start counters, light propagation in the counters and z-position of the hit in the counters. The time difference has the expected prominent peak at zero, but with not insignificant (few percent) nearest-neighbor RF bucket peaks and smaller next-to-nearest-neighbor peaks as well. Interpretation of these peaks as accidental background gives a artificially low efficiency after subtraction. There were several suggestions about things to try.

  • Beni advised avoiding RF times that were chosen based on the start counter.
  • Beni provided guidance on which RF factory product should be used.
  • Mark suggested selecting events where the event vertex is well defined, either those with multiple tracks or those with tracks that go transverse to the beamline. That may give events with better RF bucket choices.
  • Mark mentioned calculating efficiency treating the neighboring peaks as signal and getting the accidental correction from RF peaks far away from zero where the time distribution looks flat (just to see what you get).

Separating tracking contribution to resolution in mcsmear

Sean led us through his wiki page where he studies, in simulated data, the dependence of the measured start counter timing resolution on the amount of start counter time smearing input to mcsmear. Roughly speaking, he measures something about 10% broader than the smearing we put in. This is likely due to the contribution to the resolution from tracking that is intrinsic to the measurement.

Sean uses an empirical linear function to capture this effect, on a counter-by-counter basis. We all agreed that this should be implemented into mcsmear at some point. Sean also pointed out that there may be an even bigger effect to include, that of the dependence of time resolution on the z-position of the hit in the start counter.