Fast Simulation of the Dirc with Pravda

1
Fast Simulation of the Dirc with Pravda

• Rolf Andreassen / B. Meadows
• University of Cincinnati

2
Outline

• Dirc Photon Ring Dictionary
• Some performance Plots
• Adaptability
• Summary

3
What Do We Simulate?

• The Dirc must provide PID information.
• This means (?, ?, N?) for each track.
• These quantities can be combined into a
  probability
• where
• lt?gt is expected Cherenkov angle (depends on
  particle hypothesis, of course) and
• ? is the expected number of photons.
• ? is the hardest quantity to simulate!

4
Dirc Ring Dictionary

• This was invented to provide the efficiency for
  photons to be detected and reconstructed given
  (p,?,?) at entry to bar.

• Why is this important?
• Without this, the Dirc efficiency shows dips in
  efficiency just above the Cherenkov thresholds.
• Measurement of number of photons N? is important
  in these regions

5
Value Added by Ring Dictionary?

• Full Geant4 simulation used
• Tracks swum out to Dirc
• Photons generated during swim through bar
  (multiple scattering)
• Photons tracked through bar and H2O stand-off box
  (SOB) to PMTs
• Event fully reconstructed using BaBar code
• Effect of event environment is included.
• ? of reconstructed photons

? cos?/sin2?
? (radians)
? (degrees)
Tabulated by azimuth ?, dip ? at bar and
Cherenkov ?.
6
Use in Pravda

• Given (p, ?, ?)
• Compute bar number and ? at bar
• Then lt?gt cos-1 1/(? n)
• and ?2 ?track2 ?ms2 ? ?geom2 ?n2
  / ?

Ring
1.473
--? ?
Dictionary
Photons
Due to tracking (? p/p1 for now)
Multiple scattering
PMT bar Size 4 mrad
Dispersion
Smear ? as Gaussian G(lt?gt, ?) Smear N? as Poisson
with mean ?
7
Performance First Look
8
Performance First Look
Average Photons
Cos ?
9
Performance First Look
Average Photons
Cos ?
10
Performance First Look
Photons
Azimuth ? (radians)
11
Performance First Look
Cherenkov Angle (radians)
Momentum p (GeV/c)
May be a bug here ?
12
Performance

• Rolfs plots here

13
Adapt to Different Dirc ?

• Change ? detector
• If no change in time resolution, can simply scale
  ? using new QE
• In general, event environment in new time window
  will differ significantly

Probably need to make ring dictionary using (NEW)
full Geant4.
14
Summary

• We have a simulation that is fast and (with work)
  somewhat realistic.
• Experience in physics studies is still required
• A few questions in implementation remain.
• It works with the present Dirc (with H2O SOB and
  PMTs)
• Upgrade in Dirc readout will require a new, full
  Geant4 production of new ring dictionary.