Z Segmentation Update (improving reconstruction of non-prompt tracks with the SiD)

1
Z Segmentation Update(improving reconstruction
of non-prompt tracks with the SiD)

• Lori Stevens (SLAC/UCSC)
• August 21, 2007
• ILC Simulation Reconstruction Meeting

2
AxialBarrelTrackFinder1 Bug

• Phi requirement phi coordinate of all hits must
  be within p/2 of each other
• Bug in my version of code caused phi requirement
  to be too restrictive
• Fixed bug and found improvement in both
  efficiency and number of fake tracks when using z
  segmentation

3
Comparing Z Segmentation Before and After Fix
Sample pythiaZPolebbbar-0-1000_SLIC_v1r9p3_sidau
g05.slcio
None None no bug 30cm 30cm no bug 10cm 10cm no bug 5cm 5cm no bug 1cm 1cm no bug
MCPs 304 304 304 304 304 304 304 304 304 304
Found with 5 hits 145 155 144 161 157 162 156 157 126 126
Found with 4 hits 112 115 110 113 115 115 117 117 106 107
Missed 47 34 50 30 32 27 31 30 72 71
Fake 158 65 156 55 48 17 22 10 4 1
4
Some other studies (collisions at 500GeV)

• panpyuds_nobeam_nobrem-1-500_SLIC_
• v2r0p6_sidaug05.slcio
• panpyuds_nobeam_nobrem-3-500_SLIC_
• v2r0p6_sidaug05.slcio
• panpyuds_nobeam_nobrem-7-500_SLIC_
• v2r0p6_sidaug05.slcio

5
panpyuds_nobeam_nobrem-1-500_SLIC_v2r0p6_sidaug05.
slcio
No segments 30cm segments 10cm segments 5cm segments 1cm segments
MCPs 33 33 33 33 33
Found with 5 hits 15 15 15 16 14
Found with 4 hits 13 11 15 16 14
Missed 5 7 3 1 5
Fake 59 61 37 30 18
6
panpyuds_nobeam_nobrem-3-500_SLIC_v2r0p6_sidaug05.
slcio
No segments 30cm segments 10cm segments 5cm segments 1cm segments
MCPs 169 169 169 169 169
Found with 5 hits 79 84 80 81 64
Found with 4 hits 66 66 69 71 73
Missed 24 19 20 17 32
Fake 262 255 161 128 52
7
panpyuds_nobeam_nobrem-7-500_SLIC_v2r0p6_sidaug05.
slcio
No segments 30cm segments 10cm segments 5cm segments 1cm segments
MCPs 150 150 150 150 150
Found with 5 hits 69 73 73 72 57
Found with 4 hits 46 44 53 54 56
Missed 34 33 24 23 37
Fake 252 239 157 100 52
8
Combining the 3 data sets
No segments 30cm segments 10cm segments 5cm segments 1cm segments
MCPs 352 352 352 352 352
Found with 5 hits 163 172 168 169 135
Found with 4 hits 125 121 137 141 143
Missed 63 59 47 41 74
Fake 573 555 355 258 122
9
ZPole vs. 500GeV collisions
10
Conclusion

• The use 4 hit tracks may depend on the jet energy
• Will 10cm segmentation make a difference?
• Tyler Rice and Chris Meyer are doing a study
  using variable segmentation in Dima Onoprienko's
  Garfield Trackerthis may help find some answers

11
HelicalTrackFitter improvements

• Does not currently fit hits in tracker barrel
• Improving to fit hits in all regions
• Will use various fitting methods such as circle
  fitting, line fitting, 2D hit consistency checks,
  and Rich Partridges z
  segmentation code
• Presentation next Friday (8/31/07)