XEM analysis instructions

From HallCWiki
Jump to: navigation, search

If you have any questions or comments, please email - cmorean@jlab.org, or call cell # on whiteboard

  • We would like the detector check replay to be done on first 50,000 events of every run, and compare plots with the standard ones in the binder. In addition, we want physics replay to be done on every run in its entirety.
  • Do not start the full analysis until after the run ends!
  • The standard.kinematics file should update automatically. If you notice a dramatic change in plots, the background process may have terminated. See the appropriate section below.
  • Shift leaders are asked to keep a running total of the number of e- events; This will help us better estimate when to move to the next setting.


  • ssh to cdaql1 or cdaql2. Do not use cdaql3 (currently unavailable).
  • go_analysis
  • Do a 50k replay for each arm as soon as you get 50k events
  • ./run_shms.sh
    • Execute the SHMS 50k replay on the latest run


  • ssh to cdaql1 or cdaql2. Do not use cdaql3 (currently unavailable).
  • go_analysis
  • Do a 50k replay for each arm as soon as you get 50k events
  • ./run_hms.sh
    • Execute the HMS 50k replay on the latest run

Information for runsheets

  • Tracking efficiency: will appear in emacs window after 50k: "E SING SING FID EFF"
  • Computer live time: will appear in emacs window after 50k: "Pre-scaled PsX HMS/SHMS Computer Live Time", where X is the trigger leg enabled for that run (1,2, or 3).
  • The HMS and SHMS 50k analysis scripts will give you an estimated number of "good electrons" out of the 50k you just analyzed.
    • Use the good electron fraction ("good electrons"/50,000) to estimate the total number of good electrons for a given run using the total number of events in that run.

Good electrons.png

  • If you forget to write down the total number CODA events (needed to extrapolate the total number of events) you can find it in the SHMS/HMS end of run logbook entries ("Event number")
  • The tracking efficiency and computer live time can also be found in MON_OUTPUT/HMS/REPORT/summary_output_xxxx.txt (or the equivalent SHMS directory)

Full analysis

  • The shift crew will not normally perform the full run analysis, but in case you need to, simply do:
  • "./run_full_hms.sh <run number>"
  • This will by default analyze all events in the run

Notes on the 50k Replay

  • By default, the hms and shms 50k replay scripts process the first 50k events from the most recent .dat file
  • You can run more events or process a different run number if needed by providing different arguments
  • To process a different run number -
    • Provide a run number as an argument, for example, to process run 12384 through the HMS -
      • ./run_hms.sh 12834
  • To process more events -
    • Provide the run number and number of events you want to process, for example -
      • ./run_hms.sh 12834 100000
    • Would process 100k events for run 12834 on the HMS
  • When processing carbon hole runs, you may wish to run more than 50k events

Automated script that writes to database

  • This script should always be running, and is typically seen running on the 2nd desktop of the SHMS DAQ.
  • To run, ssh to cdaql1 or cdaql2, type go_analysis and then launch_runinfod
  • This script will need to be restarted any time the cdaql6 machine is power-cycled.
To add the previous run when a run has already started
caput hcSHMSRunInProgress 0 followed by caput hcSHMSRunInProgress 1
Repeat for the HMS if needed.

Carbon Hole Runs

  • See the instructions on the beam checkout page
  • Broad instructions are included below
  • Adjust the prescale settings accordingly, set a target rate of ~1kHz for the HMS and SHMS 3/4 triggers (PS1 and PS3)
    • The rates you'll see will depend heavily upon setting, the step above is just to get reasonable rates to disc
  • Once the prescale values are set, start a run
  • When you have 50-100k events, run a HMS and SHMS 50k
    • Check the HMS/SHMS Fast Raster page of the output plots, check the hole position. If it looks off, ask MCC to adjust the position.
    • You will need to stop and start a new run when they adjust the position
    • This may take some trial and error to correctly centre the beam
    • You can also try running more than 50k events in the replay if needed (see above)
    • More information is located here

Debugging - General

If any of the plots look odd/not as described:

  • The standard.kinematics file should update automatically, but you can double-check that standard.kinematics has been updated and is CORRECT.
  • Watch out for silly typos, e.g. gbeam rather than gpbeam.
  • Note that standard.kinematics affects things you think it really shouldn't! If things look wrong this is absolutely the first thing you should check.
  • Make sure you have the correct particles going into each spectrometer

You should not -

  • Panic
  • Start editing scripts and deleting things without knowing what you are adjusting
  • Delete ANYTHING in the scripts without also leaving a comment explaining what you have deleted or changed

If you cannot run the replay and see things about not being able to get gpbeam, for example, you might need to restart the script that writes to the database.

  • See instructions above for running the launch_runinfod script.