SoLID Ecal Weekly 20230413

Discussion on Run List

Mike:

• Runlist: in progress, suggest 4 plots – run condition; Scintillator; Shower; Preshower
• also showed alignment group survey report for 18 deg.
• suggestion: put alignment report and Jimmy’s CAD drawing together to form a technical note.

Discussion on Simulation

• Presentation by Ye
• LASPD waveform study: Note that LASPD covers nearly the full top Shower, but only half of left or right Shower
• 5uA LD2: compare waveform of Shower-Top, LASPD-T and LASPD-B, Some events have LASPD >0, some =0 when pulse is out of time or no pulse.
• 10uA LH2: many pulses overlap
• Q&A (Tim): TS4 timing is determined by ShowerSum, but Ye only plotting Shower-Top, would that make a difference? Jixie: Showers are aligned within 2 windows (8ns)

Discussions on Tracking

• Presentation by Xinzhan
• with grid option, cooking speed is 40 times faster than no-grid option. Grid now using GEM1 and GEM4 (bruteforce scan) to define grid for GEM2 and GEM3. Speed-wise it’s good enough.
• 2D clusters are probably still um-level “off” (incorrect) because chisq is peaked at 0, and xp, yp are tilted for low rate runs but symmetric for high rate runs.
• List of possible optimization
  • using target+GEM4 (or 3) instead of GEM1+GEM4
  • using ECal and/or other detector for constraining so we do not rely completely on chisq minimization
  • correct all offsets
  • could we quote both accuracy and efficiency?

Discussion on Cooking

• Jixie did another round of 18 deg data, next will be after GEM tracking is fully opitimized
• Jixie will cook 82 and 7 deg data, question: 82 deg had 2 working GEMs, could we use target as a 3rd point to reject/select the 2-GEM “tracks”? – need to think about this, would the effort be worth the outcome? Maybe not?

Summary of Analysis Meeting Monday April 10

• GEM: (Xinzhan)
  • current estimate of tracking efficiency, considering “good” charged particles that should form tracks, is at 10-30% depending on cuts;
  • Eventually, we might expect 30%. the 4 GEM chambers are 4m apart, so acceptance already does not match. 30% is relative to all charged particles hitting the first two GEMs. If we take the last two layers of GEMs, efficiency should be close to 100%.
• Runlist (Mike)
  • going through all runs to extract MIP positions, etc, will get this information wrapped up soon. XZ: can we have pyroot or similar tools to make plots where we can use mouse hover over to get (run#,MIP#), etc? → yes
  • Note that we can also use the graph to track conveniently which detector and which runs had saturation problem.
  • could also add target/current info, so information is easy to pull out from graph.
• Cooking: (Jixie)
  • implemented Xinzhan’s latest tracking code, cooking 50 golden runs now, 50k cooking is below 96 hours. ifarm testing speed shows 405sec/1k events (so should be 3-4 hours/50k?) for a very high occupancy run (run4818).(Wait, didn’t Xinzhan say 100k events needs only 2837 sec?)
  • Alexandre commented: ifarm speed varies and farm nodes should be much faster.
  • Anyways, the bottom line is that all jobs should run okay without exceeding the farm job time limit.
• Cherenkov: (Darren)
  • looked at Cherenkov for 4680 with GEM cuts to select tracks hitting the center. In the Nch vs. Npe plot, TS1 seems to show two blobs (while TS2 shows just one). Could the higher Npe blob be used to define good electron cuts?
  • Other cuts used: Nch coincidence cut. For example, if 3 channels fire but only 2 are adjacent to each other, the 3rd one (which is away from the 2-cluster) is set to zero.
  • need to add high-energy electron cut to see if this second blob is from electrons. Ye sent suggested cuts:
    • Based on the run 4588 with the highest TS4 threshold (see figure), the high-energy electron cut is ShowerSum >12000;
    • Besides, the run 4577 (40uA high statistics) and run 4344 (10uA) are the good runs for high energy electron study. However, Mike said these runs did not have good GEM latency.
  • Suggestions:
    • (Zhiwen): The Nch coin cut may unnecessarily rejects good events; (Xiaochao) we do not need to use identical cuts as Cherenkov paper.
• General discussions on the analysis:
  • We could consider: To identify good electron response, use only TS1 to plot Nch vs. Npe, apply a subset or all of the following cuts, or apply them one by one and monitor spectra
    • high-energy electron cut ShowerSum>13000 (from Ye)
    • signal in all scintillators > 0.5 MIP leading up to the ECal
    • no signal in SC-B
  • To identify good pion sample, we need TS2 to study Cherenkov response – any golden run on this?
  • Furthermore or alternative, use GEM and Cherenkov information to determine:
    • hit (x,y) position in Cherenkov must “match” projected GEM track projected on Cherenkov. (Mike showed something about finding weighted (x,y) position in Cherenkov).
    • could we use position cuts between GEM and Cherenkov to select good electrons?
    • could we use position cuts between GEM and Cherenkov to select good pions? (For example, that within a circle of track position, there is no Cherenkov hit)?
  • Our goal of Cherenkov study is to define PID samples using Cherenkov. Potentially:
    • electrons = GEM center cut, high Cherenkov yield (some combination of Nch, Npe, position, etc), plus SC cuts??
    • pions = GEM center cut, no Cherenkov light (for all channels or within a certain radius of track). – need TS2 events to confirm. ??
    • PID sample for photons, potentially: photons = response in ECal (all 3 modules), no GEM track (or track wildly mismatch ECal cluster position), no signal in Cherenkov; we could also add anticuts of all scintillators SC-A,C,D, or using two of them and study photon rejection of the 3rd (one of A,C,D or SPD).