Difference between revisions of "20180912-Analyzer-Mtg"

From PREX Wiki
Jump to: navigation, search
(Agenda)
(Agenda)
Line 12: Line 12:
 
== Agenda ==
 
== Agenda ==
  
# item
+
# n.b. Next week's meeting is cancelled.  Next meeting is scheduled for Sept 26.
#* comment
+
#* Cameron and Tao will email questions to this group, or notify if there is some progress/output to be looked at.
#* Carl and Bob's Two-ET solution for online analysis, diagram: [https://prex.jlab.org/wiki/index.php/File:Two_et.pdf#file]
+
# Carl and Bob's Two-ET solution for online analysis, diagram: [https://prex.jlab.org/wiki/index.php/File:Two_et.pdf#file]
#* BPM Pedestal root scripts update: [https://ace.phys.virginia.edu:80/HAPPEX/3538#file]
+
#* Bob will make a test script when he returns to fully qualify this system
 +
#* JAPAN compiles and connects to a test ET server, so Paul thinks it should be ready to test off of the new ET system
 +
#** Kent: probably first priority is to get online monitor working (since this is most likely to be used in beam studies). Then feedback, because that requires more software porting/development. (Of course, feedback will also be used in injector studies, so it is still on our priority list!)
 +
#*** possible 1st test: run analysis off ET server, then off disk data file.  Compare root outputs for completeness.
 +
#*** possible 2nd test: have JAPAN make a memory map rootfile.  Testing this could be done with client software (ported from qwanalysis, or panguin), or perhaps(?) an interactive root session.
 +
#*** possible 3rd test: move on to feedback.
 +
#** Paul will add an issue in github to cover this.
 +
#* Bob suggests that we commit to a non-blocking ET solution (so analysis will never cause deadtime)
 +
#** This is agreed, but Kent points out that we need very visible alarms so we don't fail to notice if online monitor or (especially) feedback don't receive full event set.  This is especially critical because poor charge feedback would doom MOLLER, so we need to demonstrate sufficiency of our feedback technique during PREX.
 +
# Bob suggests we make a list of milestones for next collaboration meeting
 +
#* We agree this is a great idea but none of us commits to mapping this out for the next couple of weeks.  We all agree to think in spare time, send suggestions for this list to other members of this group.
 +
# Cameron is starting to document the naming convention for devices used in JAPAN.
 +
#* Sorted through BPM names.
 +
#* will look at PAN db files for Injector and CountingHouse to find more data types (bcm, dets, sams, scandata words, helicity /pattern flags, synch monitor, linephase monitor, dithering/BM words, etc)
 +
#* HRS detectors will likely follow det<l/r><name>, with l/r for left/right and name indicating up/down stream, or AT det.
 +
# Tao: BPM Pedestal root scripts update: [https://ace.phys.virginia.edu:80/HAPPEX/3538#file haplog:3538]
 +
#* has documented bpms/bcms rather completely, in slides in the log entry.
 +
#* working pedestal calibration scripts from JAPAN, reasonable looking results
 +
#* will email Amali to ask her to run PAN scripts to produce comparable output.  may need to look through PAN scripts to figure out source of any discrepancies.
 +
#* Suggestions from Kent
 +
#** should add parameters for lower and upper limits on current used in calibration
 +
#** a useful option would be to either
 +
#***include points that are outside limits on plots (exclude them from fits and use a different data marker, like an open circle)  OR
 +
#***exclude points that are outside limits entirely (so we can focus on the quality of points that are used)
 +
# Cameron asks about DAQ problems that were seen in test beam.
 +
#* Data was somehow corrupted, with vqwks reporting same value for many events in a row.
 +
#* Problem would spontaneously(?) fix itself between runs.
 +
#* Paul suggests checking the number of samples on the vqwks to see if this was correct, or somehow corrupted.
 +
#* Kent will ask Amali to send run numbers with this problem to Paul for investigation.
  
 
== Present ==
 
== Present ==

Revision as of 10:35, 12 September 2018

Back to Main Page >> Analyzer_Meeting

previous meeting << >> following meeting

Logistic information

 BlueJeans calling instructions:
 Toll-Free Number (U.S.&  Canada):  888-240-2560
 International toll number:         408-740-7256
 Bluejeans CODE:                  684 884 269
 Bluejeans link: https://bluejeans.com/684884269

Agenda

  1. n.b. Next week's meeting is cancelled. Next meeting is scheduled for Sept 26.
    • Cameron and Tao will email questions to this group, or notify if there is some progress/output to be looked at.
  2. Carl and Bob's Two-ET solution for online analysis, diagram: [1]
    • Bob will make a test script when he returns to fully qualify this system
    • JAPAN compiles and connects to a test ET server, so Paul thinks it should be ready to test off of the new ET system
      • Kent: probably first priority is to get online monitor working (since this is most likely to be used in beam studies). Then feedback, because that requires more software porting/development. (Of course, feedback will also be used in injector studies, so it is still on our priority list!)
        • possible 1st test: run analysis off ET server, then off disk data file. Compare root outputs for completeness.
        • possible 2nd test: have JAPAN make a memory map rootfile. Testing this could be done with client software (ported from qwanalysis, or panguin), or perhaps(?) an interactive root session.
        • possible 3rd test: move on to feedback.
      • Paul will add an issue in github to cover this.
    • Bob suggests that we commit to a non-blocking ET solution (so analysis will never cause deadtime)
      • This is agreed, but Kent points out that we need very visible alarms so we don't fail to notice if online monitor or (especially) feedback don't receive full event set. This is especially critical because poor charge feedback would doom MOLLER, so we need to demonstrate sufficiency of our feedback technique during PREX.
  3. Bob suggests we make a list of milestones for next collaboration meeting
    • We agree this is a great idea but none of us commits to mapping this out for the next couple of weeks. We all agree to think in spare time, send suggestions for this list to other members of this group.
  4. Cameron is starting to document the naming convention for devices used in JAPAN.
    • Sorted through BPM names.
    • will look at PAN db files for Injector and CountingHouse to find more data types (bcm, dets, sams, scandata words, helicity /pattern flags, synch monitor, linephase monitor, dithering/BM words, etc)
    • HRS detectors will likely follow det<l/r><name>, with l/r for left/right and name indicating up/down stream, or AT det.
  5. Tao: BPM Pedestal root scripts update: haplog:3538
    • has documented bpms/bcms rather completely, in slides in the log entry.
    • working pedestal calibration scripts from JAPAN, reasonable looking results
    • will email Amali to ask her to run PAN scripts to produce comparable output. may need to look through PAN scripts to figure out source of any discrepancies.
    • Suggestions from Kent
      • should add parameters for lower and upper limits on current used in calibration
      • a useful option would be to either
        • include points that are outside limits on plots (exclude them from fits and use a different data marker, like an open circle) OR
        • exclude points that are outside limits entirely (so we can focus on the quality of points that are used)
  6. Cameron asks about DAQ problems that were seen in test beam.
    • Data was somehow corrupted, with vqwks reporting same value for many events in a row.
    • Problem would spontaneously(?) fix itself between runs.
    • Paul suggests checking the number of samples on the vqwks to see if this was correct, or somehow corrupted.
    • Kent will ask Amali to send run numbers with this problem to Paul for investigation.

Present

Excused

Cip