Trigger Monitoring Tasks

Last update: May/16/05

L1 Fast Monitoring

• Trickle injection monitoring.    
  • This should include items not already covered in L3 fastmon which don't require full statistics. Which histogram going to were is still subject  to discussion.
  • Weaver plot and projections for input L1 to monitor events leaking out of inhibit box due to L1 timing bias in very messy events.
  • L1 line pattern for trickle sampling triggers. The absolute rate, line pattern and line timing yield detailed info on the probability of fluctuating injection bunch triggering L1.
• Various enhancements to DCZ monitoring.
  • New vs old TSF detailed pattern bit comparison. (Jamie Boyd.  DONE)
  • TSF->DCZ transmission test using TSF and ZPD input data.
  • DCZ hardware status: GLINK errors, frame bit errors, internal errors etc. in DAQ head CSR1 data   (Selina Li)
  • ZPD output decision.
• L1 odf damage statistics.
• All L1 component odf damage summary histogram  (Selina Li).
• GLT vs FCT bit checking.
• Needs clean up to avoid cases of damaged events.  (Kris: DONE)
• Need to avoid FCT only trigger jumping ahead just before GLT trigger to flag only genuine bit mismatch.
• DCT (DCZ)/EMT->GLT cable transmission error check.   (R&D: Su Dong,  Implementation: Jim Burke/Paul Jackson. DONE)
  • Need to apply corrections to BLT,PTD,ZPD DAQ data to form  predictions for GLT. Need to apply time window cut for the checking.  
    
• Event time history within a run.
• Histogram event counts vs time since run start for L1 pass and logged L3.  Technical issue: best way to get run start time (note: monitoring are running on many different nodes) ? 
• Labeling L1 lines by name for line rate histograms.
  • Rainer has the KUMAC technology, but needs to connect up to FMon and OPRmon output packaging.
• Automated monitoring alarms to flags errors for some critical observables.
• Some initial discussions are needed to determine which set of variables are the reliable critical observables  to go on the auto alarm list, and how to define the alarm. 
• L1Fmon house keeping
• When fastmon moves to 14.x.x release, needs to check plots, as Itaru split the L1TFmon into a few separate modules and DCZ mon will also show up. (Kris. DONE)
• Integrate and imprive DCZ Fmon into L1FMon JAS. (Kris: DONE)
• Migration of Hbook to Root for OprMon  (Kris: DONE)
• We are still running crontab for making L1FMon archive. Looking forward to the endrun driven production from OEP !    (Kris:  DONE)
• L1FMon archive PS file directory should be partioned into several directories ?
  

L3 Fast Monitoring

• Trickle injection monitoring.
  • A set of critical monitoring which rquires full L1 input statistics:  (Prafulla Behera)
  • injection trigger/marker separations
  • Weaver plot for full L1 input sample
  • Weaver plot for all Bhabhas and clean L3 physics events, and physics rate vs injection cycle for injected bunch.
  • Other usful info from L3
  • L3 T0 and EMC cluster mean time for a few different zones in weaver plot.
• Labeling L3 lines by name for line rate histograms. (Prafulla Behera) 
• Automated monitoring alarms to flag errors for some critical observables.
• Similar to L1, we need some discussions first as to what we should include here.
• Archive browsing capability like L1Fmon with a few fixed pages of plots per run.

General Online Monitoring

• Re-establish and enhance the existing weekly monitoring summary facilities.
• Need to work on convenient ways of autoupdating, archiving and displaying the information. Current L1+genenral  and L3 parts are done quite differently.
• Enhance the weekly monitoring facility with strip charts for some critical trigger performance observables. We need to re-examine what observables we need and create the relevant histogram in Fastmon to serve these variables in the end run histograms for strip charts in any case.
• Longer time scale (e.g. monthly) monitoring to catch subtle changes.

L1 OPR Monitoring

• Various technical details to fix in OPR mon. (Itaru. DONE)
• L1TOprMonGlt has a bug. Needs to check the existence of an L3 line before using it.
• There was one incident of simple track extrapolator crash/infinite loop not yet understood.
  
  
• Event history and basic occupancy plots.
• OPR mon needs to be more selfcontained for convenient Q/A. Some gross conditions variations during run cannot be picked up by existing histograms. This requires the event time history vs run time.  Current efficiency histograms use DCH/EMC reco as base, but they cannot tell clearly what was going on if DCH and EMC had major problems. Some basic TRG occupancy plots will help. 
• DCZ performance histograms.
(Kris. DONE)
• Efficiency and resolution vs reco tracks.
• Labelling L1 lines by name for line rate histograms.
• Add more strip charts for critical observables.

L3 OPR Monitoring

• L3 efficiency and resolution vs Reco quantities.   (Prafulla Behera. DONE)
• Labelling L3 lines by name for line rate histograms.    (Prafulla Behera)
  • Add monitoring of efficiencies to seleced set of known lines ? 
• Establish strip charts for critical observables.  (Prafulla Behera)

• Mini version of OPR monitoring which can run on the TRG Mini.    (Nicolas Berger)
• DCT/EMT efficiency from GLT input Mini data.
• Stripcharts  for MiniMon.