Photon Reconstruction in b1pi events 04/25/2012

From GlueXWiki
Jump to: navigation, search

50,000 b1pi events. EM background on. svn rev 9031.

All plots use thrown vertex information to decouple photon analysis from tracking.


mcsmear behavior is modified so that noise hits do not affect timing of non-dark hits. Normal time-smearing still-applied.

Thrown photons

This is the z at which the photon would reach the inner radius of the BCAL.

PhotonEvsZ thrown BCAL.svn9031.png

Reconstructed "photons"

The plot below comes from DNeutralParticleHypothesis. At this stage the only cut that has been applied is rejecting clusters geometrically matched to charged tracks (this matching code is somewhat arcane and probably should be examined more closely. True facts: it does allow more than one cluster match per track, it reduces the number of clusters from 253,000 to the 175,000 below).

The z in these plots is cluster z at the presumed cluster maximum, not the z at the inner radius, so not directly comparable to thrown photon plot above.

PhotonEvsZ recon BCAL.svn9031.png

Removing the lower right bin to get a more reasonable scale:

PhotonEvsZ recon BCAL.better scale.svn9031.png

Reconstructed showers (single particle events)

Generate, simulate, and reconstruct samples of single-particle events with same kinematic distribution. These plots are made from DBCALShowers, this means that matching to charged tracks has not yet been performed, but non-linear energy corrections are applied.


lower rightmost bin removed, as above, to fix scale

Single particle DBCALShower EvsZ.gamma.better scale.svn9031.png

What is the stuff in the at large z, small E?


(pi+ only in this graph) lower rightmost bin removed, as above, to fix scale

Single particle DBCALShower EvsZ.pi+.better scale.svn9031.png

Why this distribution?

  • 79091 BCAL clusters
  • 53979 left after vetoing matches to charged tracks


Single particle DBCALShower EvsZ.proton.better scale.svn9031.png

  • 25357 BCAL Clusters
  • 14062 left after charged track matching

Cluster depth

r coordinate in cylindrical coordinates of cluster center

Single particle DBCALCluster r pi vs gamma vs proton.png

  • blue: pi+
  • red: proton
  • black: photon


In full b1pi event:

  • Timing cut: reduces # of clusters from 175,000 to 71,000.
  • Loose depth cut: reduces # of clusters by additional 7%.


Thrown photons

PhotonEvsTheta thrown FCAL.svn9031.png

Reconstructed "photons"

PhotonEvsTheta recon FCAL.svn9031.png

Single-particle events


After matching to charged tracks.

Single particle DNeutralParticleHypothesis EvsTheta FCAL pi.png

  • 92335 FCAL Clusters
  • 15442 remain after matching


Single particle DNeutralParticleHypothesis EvsTheta FCAL gamma.png

Shower shape

RMS cluster size:

Single particle DFCALCluster RMS pi vs gamma.png

  • blue: pi+
  • black: photon

splash = E_{3x3}/E_{5x5}

Single particle DFCALCluster splash pi vs gamma.png

  • blue: pi+
  • black: photon


On full b1pi events:

  • Timing cut: Cuts out ~16% of showers
  • Loose cuts on RMS size, splash: cuts out additional 10% of showers

only ~1% "good" photons lost in these cuts