Mattione Update 10312011

Procedure

• Reconstruct π(1600) -> b1π events with t-slope = 2.0 (was 5.0 for X(2000)) using both the 334 and 1234 segmentations
• Plot Δβ vs p for π⁺ candidates to see proton, π⁺ separation.

1234 Segmentation BCAL Time-Smear Details

• The 1234 segmentation is not available in the HDGEANT simulation, so its BCAL shower time was smeared manually during reconstruction
  • The BCAL hit-time smear was disabled in mc_smear
• The shower time was smeared by: σ_smear = sqrt(σ²_{t_1234_Target} - σ²_{t_334_NoSmear}), where σ_{t_1234_Target} = σ_{t_334}*Improvement_Ratio, Improvement_Ratio = σ_{t_1234} / σ_{t_334}
  • These time uncertainties were studied as a function of dE and θ
  • The σ_{t_1234} / σ_{t_334} improvement ratio was obtained from fitting the results of David Lawrence's studies of photons
  • The σ²_{t_334} and σ²_{t_334_NoSmear} variances were determined separately for protons and pions, by comparing the reconstructed shower times to their true values.

Improvements from last week's study

• Looking at π(1600) -> b1π events with t-slope = 2.0 instead of π(2000) -> b1π events with t-slope = 5.0: Higher proton momentum in BCAL for studying separation
• The σ_{t_334_NoSmear} resolution is no longer assumed to be zero: it is non-negligible for large energy deposits (presumably due to clustering?)
• The σ_{t_334} resolution is determined separately for protons and pions (was using the pion resolution for both previously)
• Better fit to photon σ_{t_1234} / σ_{t_334} distributions due to extra data point at 50 degrees
• Low-θ, low-dE time resolutions better-simulated: instead of extrapolating fit, performing linear-interpolation between data-points.
• 2 (presumably canceling) bug fixes, final results on hold until confirmation (tomorrow possibly?).

NOTES

• The photon σ_{t_1234} / σ_{t_334} is assumed to be applicable to protons and pions

David's BCAL Photon Time Resolution Studies

• γ BCAL σ_t vs. dE (GeV):

Time Resolution Segmentation Ratio

• γ BCAL σ_{t_1234} / σ_{t_334} vs. dE (GeV):
  • NOTE: When applying this "Improvement Ratio," the maximum allowed value was set to 1 (shouldn't be worse for 1234 at large angles, right?)

• Fits to γ BCAL σ_{t_1234} / σ_{t_334} vs. θ in dE bins:
• Don't trust fits at low-θ, so linearly-interpolate between 12 and 16 degrees.

BCAL π^- & proton 334 Segmentation σ_t Fits

• When the BCAL time-smear is enabled, the π^- & proton time resolutions are similar.
• When the BCAL time-smear is disabled, the π^- & proton time resolutions are similar, except at high shower energies.

BCAL π^- 334 Segmentation σ_t Fits

• BCAL π^- 334 Segmentation σ_t Fits vs. θ in dE bins:
  • NOTE: For the first dE bin, the values were linearly interpolated between the points instead of fit to a function.
    • Last week, the fit was extrapolated from high-θ over the low-θ range.

σt Linearly Interpolated

BCAL π^- 334 Segmentation σ_t Fits with BCAL Time Smear Disabled

• BCAL π^- 334 Segmentation σ_t Fits with BCAL Time Smear Disabled vs. θ in dE bins:
  • NOTE: These resolutions are negligible except at high dE.
  • NOTE: The bad fit results were ignored.
  • NOTE: For the first two dE bins, the time smear was set to zero.
  • NOTE: For the third and forth dE bins, the values were linearly interpolated between the points instead of fit to a function.

Evaluated as σt = 0	Evaluated as σt = 0	σt Linearly Interpolated
σt Linearly Interpolated

BCAL Proton 334 Segmentation σ_t Fits

• BCAL proton 334 Segmentation σ_t Fits vs. θ in dE bins:
  • NOTE: For the first dE bin, the values were linearly interpolated between the points instead of fit to a function.
    • Last week, the π- fit was used, and was extrapolated from high-θ over the low-θ range.

Linearly Interpolated

BCAL Proton 334 Segmentation σ_t Fits with BCAL Time Smear Disabled

• BCAL proton 334 Segmentation σ_t Fits vs. θ in dE bins:
  • NOTE: These resolutions are negligible except at high dE.
  • NOTE: The bad fit results were ignored.
  • NOTE: For the first dE bin, the values were linearly interpolated between the points instead of fit to a function.
  • NOTE: For the second dE bin, for θ <= 27.0, the values were linearly interpolated between the points. For θ > 27.0, the time resolution was set to zero.

Linearly Interpolated	Linearly Interpolated

π(2000) vs π(1600) b1pi Events, π⁺ BCAL Δβ vs p

Δβ = β_p - β_t = p/sqrt(p*p + m*m) - path/(c*TOF)

• b1pi Events, π⁺ BCAL Δβ vs p, 334 Segmentation:

π(1600) t-slope = 2

π(2000) t-slope = 5