May 8, 2007 Calorimetry

From GlueXWiki
Jump to: navigation, search


Dial-In Information


BCAL Analysis

Blake discussed the studies he has been doing to measure the time resolution of the BCAL. He noted that the best weighting algorithm for combining multiple cells is to weight based on the energy deposited in the cell. Using this weighting algorithm he studied variance of the mean time and time difference for clusters generated by photons at 90 degrees in the center of the BCAL.

Mean Time

The mean time resolution as a function of energy could be characterized by a statistical and floor terms of 64 and 134 ps. The 134 ps floor term is to be added in quadrature to the estimated 113 ps error introduced by tagger/trigger timing to get the actual floor term. It was noted that the scatter of the data points about the fit line is not consistent with purely statistical errors. We concluded that there is likely some residual systematic effect in the data due to the timing alignment of different tagger paddles used to span the entire energy range.

Time Difference

Similar plots of the end to end time difference indicated statistical (floor) terms of 81 (21) ps. These plots are nicely fit to the expected parameterization for the resolution and do not show the additional systematic "scatter" about the best fit line as was seen with mean time. Elton and Dave cautioned that some errors may cancel in the time difference causing one to underestimate the final expected TOF error (which, in practice experiment, must be derived from the mean time).


  • The key question is: "How much of the 134 ps floor term seen in the mean time measurements is due to resolution of the BCAL?"
  • Instead of looking at mean time with respect to the tagger, examine mean time with respect to a separate cell in the BCAL. This removes the potential cancellation of errors within each cell.
  • Since the fluctuations in mean time resolution seem to be introduced by various tagger paddles that span the energy range, select just 3 - 4 tagger paddles and various energies and compute the mean time resolutions for these paddles only. Such an exercise might help verify or understand the intrinsic resolution of a single paddle.
  • Carefully Examine the time walk correction. Since both ends should have comparable pulse shape, a systematic error in time walk correction would likely subtract from end to end difference but persist as an additional error when computing the mean time.

BCAL π0 Reconstruction

Mihajlo showed preliminary plots that examine the BCAL photon reconstruction. The probability of conversion for photons uniform in angle seems to be about 10%. About 10% of these converted photons appear to generate multiple showers in the BCAL. This indicates that efficiency loss due to conversions might not be as severe with the BCAL as it was with the FCAL.

Mihajlo also showed plots of E_generated/E_reconstructed as a function of energy. These plots indicate at low energies the reconstruction algorithm systematically reconstructs a lower energy shower than was generated. This error also has some z-dependence. Mihajlo is attempting to generate a z and energy dependent correction to apply to photon energies.

Upcoming Meetings

  • PID/Calorimetry workshop will be a JLab next Monday and Tuesday so there will be no Tuesday meeting
  • On Tuesday May 22, the focus of the meeting will be software
  • The next calorimetry meeting will be Tuesday, May 29