How to Get 10 MeV Resolution from BLAST

1
How to Get 10 MeV Resolution from BLAST

• A Linear Calibration of
• the Wire Chambers
• Chris Crawford
• June 25, 2004

2
Outline

• Introduction and motivation
• linear least squared fits
• example TOF calibrations
• WC linear calibration method
• survey of raw WC data
• extraction of stub peaks
• extraction of vertex peaks
• Test Calibration of few cells
• Results outlook

3
Introduction

• method of combining all useful calibration data
  into one consistent calculation
• flexible
• any calibration peak can be added
• missing peaks can be omitted
• can use any model of the t2d relationlinear /
  analytic / numerical
• built in measure of self-consistency

4
MotivationMC Resolution
intrinsic wire resolution
Momentum Resolution (MeV)
5
Motivationc2 MC vs data
blue mc (ideal) red mc
(500m) magenta data (p) green data
(e)
6
MotivationReal Data
7
Least Squares Solution

• Design matrix X mdatanparm (normalized by sy)
• SV Decomposition (ignore infinite eigenvalues)

8
ExampleCosmics
9
WC Stub Ntuple

• wcb_ntuple.C
• track info (p,q,f,z)
• stub coordinates
• t0,1,2 x0,1,2, d0,1,2
• qtrk in cell, B field

10
WC raw data
11
Stub Peak Extraction

• old method
• 1 equation in t0,t1,t2
• 7 different stub types (need 4, only 2 are
  linear!)
• modification use qtrk ! (d-2)(n1)6 parms,
  d4, n2

12
Robust Linear Fits

• fit for lowest sigmaas a function of projection
  slope

• alternate methods
• fit-slices ! linear fit
• 2d gaussian
• lsq mode instead of mean

13
Stubs Plane Fit
14
Strange ftrk dependence

• related to fefp deviation?

D x - tan qstub
ftrk
15
Vertex Fit

• initial profile fitto find range
• robust fit each side of vertex separately

16
Least Squares Fit

• implemented in wc_offsets.C

17
Hyperbolic t2d function
18
Test Calibration

• test of1 cell adjacentsides
• continuitybetween cells

run 6992, event 14746
re-crunched runs 6992 6999
19
Hit Distribution in Cells
20
Comparison with Garfield
21
Verification

• room for improve-ment
• nonlinear deviations

cm
cm
22
Results c2
72 33
1mm resolution
23
Results Dp
28 MeV 12 MeV
24
Problems to Investigate

• 5 stereo angle
• rotation or skew?
• (small effect)
• B-field map
• problem with padding, fgt10
• use harmonic interpolation to complete grid

25
Help Wanted!

• extend calibration to all cells
• finish complete calibration
• add variable resolution to TBLFitTrack
• add new peaks using enhanced wcb-ntuple
• double hits (for boundary position)
• segment ftrk (for perpendicular tracks)
• cosmics (to connect chambers, sectors)
• combine this method with nonlinear t2d
• use linear fit to get realistic Garfield
  parameters
• empirical t2d, relaxation method (Chi, Taylan)