|
1
|
Oct 24/24 PionLT/KaonLT Analysis Meeting Notes
|
|
2
|
----------------------------------------------
|
|
3
|
(Notes by GH)
|
|
4
|
|
|
5
|
Today: Mark Jones will give a special presentation on
|
|
6
|
hcana tracking improvements resulting from the VCS anaysis
|
|
7
|
|
|
8
|
Present
|
|
9
|
-------
|
|
10
|
Regina - Garth Huber, Nathan Heinrich, Muhammad Junaid, Nacer Hamdi,
|
|
11
|
Alicia Postuma
|
|
12
|
FIU - Pete Markowitz
|
|
13
|
CUA - Tanja Horn
|
|
14
|
JLab - Dave Gaskell, Mark Jones
|
|
15
|
|
|
16
|
Mark
|
|
17
|
----
|
|
18
|
- VCS experiment saw a build up of events between photon and pi0 MM^2 peaks
|
|
19
|
- noticed more than expected Chi-square>100 events in Heep elastics
|
|
20
|
- Heep plot of Chi-Square vs W^2 clearly shows the excess at large Chi-Square
|
|
21
|
|
|
22
|
- tried various hcana algorithm improvements to find what was causing these
|
|
23
|
large Chi-Square events
|
|
24
|
|
|
25
|
1) cut on DriftTime after StartTime correction is applied (removes 10-20ns
|
|
26
|
jitter)
|
|
27
|
- reduces # junk hits, but did not help large Chi-Square problem
|
|
28
|
|
|
29
|
2) modified NewLinkStubs algorithm
|
|
30
|
- loops through Ch1,2 SpacePoints
|
|
31
|
- if Stubs in Ch1,2 pass: xt_track_criterion, yt_track_criterion,
|
|
32
|
xpt_track_criterion, then a track is formed
|
|
33
|
- if after above loops, no track is found, makes tracks from all Ch1,2
|
|
34
|
SpacePoint combos and picks the lowest Chi-Square combo
|
|
35
|
|
|
36
|
3) made a NewFindSpacePoints algorithm
|
|
37
|
- old algorithm was ok, but since SHMS/HMS DC have similar layout, clustering
|
|
38
|
algorithm can be simplified
|
|
39
|
- make clusters of like UU',XX',VV' planes instead of pairs of unlike
|
|
40
|
planes
|
|
41
|
- then make combos of UU',XX' into UX clusters, VV',XX' into VX clusters
|
|
42
|
- have X,Y info from UX,VX clusters
|
|
43
|
|
|
44
|
4) modify TrackFit algorithm to eliminate hit w/largest residual and do a new
|
|
45
|
fit
|
|
46
|
|
|
47
|
- plot of SHMSxp stubs difference (StubXpdiff) of golden track
|
|
48
|
- set xp_track_criterion=0.4 to exclude tracks >0.4, <-0.4
|
|
49
|
- #possible tracks got a lot bigger than old code, since only one golden
|
|
50
|
track is selected per event, total #tracks should remain the same
|
|
51
|
- Result: new code has 10X fewer events with Chi-Square>100, Heep W peak is
|
|
52
|
narrower
|
|
53
|
|
|
54
|
- *NB* These algorithm changes are OFF by default, need to set flag ON and set
|
|
55
|
spacepoint criterion to implement
|
|
56
|
- *NB* most effective one is: TrackLargeResidCut
|
|
57
|
- only does track fit from stubs if residual was large
|
|
58
|
|
|
59
|
- *NB* Mark has also updated the chamber alignment parameters, these are needed
|
|
60
|
for tighter residual cuts applied in the above algorithms
|
|
61
|
- they are available from:
|
|
62
|
https://github.com/JeffersonLab/hallc_replay/blob/master/PARAM/SHMS/DC/pdc_newalign_geom.param
|
|
63
|
and
|
|
64
|
https://github.com/JeffersonLab/hallc_replay/blob/master/PARAM/HMS/DC/hdc_newalign_geom.param
|
|
65
|
|
|
66
|
- Nacer and Junaid will investigate if this makes an improvement to their Heep
|
|
67
|
replays
|
|
68
|
|
|
69
|
Junaid
|
|
70
|
------
|
|
71
|
PionLT SHMS pi+ PID studies
|
|
72
|
- fixed integral calculation, now #events do not drop as much when cuts are
|
|
73
|
applied
|
|
74
|
- used: Aero>2.5pe and 1.2<RF<3.4 cuts
|
|
75
|
|
|
76
|
- check of plot of RFtime vs delta_SHMS
|
|
77
|
- if RFtime is correlated with delta, the correction for pi+ TOF is not
|
|
78
|
applied well
|
|
79
|
- *NB* Junaid will follow-up with Stephen about the TOF correction to RFtime
|
|
80
|
|
|
81
|
- will need different RFtime cuts when beam bunch spacing changes
|
|
82
|
- *NB* also should check the cut for each SHMS setting, hopefully the TOF
|
|
83
|
correction reduces the SHMS momentum sensitivity, but this needs to be
|
|
84
|
confirmed
|
|
85
|
- Junaid will also do some cut dependence study
|
|
86
|
|
|
87
|
Nathan
|
|
88
|
------
|
|
89
|
PionLT pi+ and SIDIS Coin Lumi scans
|
|
90
|
- implemented Random substraction, results now look a lot better
|
|
91
|
|
|
92
|
- LH2 pi+ Runs 14601-14606, now no-track and track plots have consistent drop in
|
|
93
|
yield slope, about 8% @ 70uA
|
|
94
|
- LH2 Runs 15816-15836, no-track and track also consistent
|
|
95
|
- LH2 Runs 16260-16276, also consistent
|
|
96
|
- LH2 RUns 16220-16242, NOT consistent, both show anti-boiling, but different
|
|
97
|
amounts
|
|
98
|
- doing checks, believes the CoinTime cut is letting in randoms
|
|
99
|
- SIDIS Runs 12218-12258 cover less current range than pi+
|
|
100
|
- no-track and track are constent, but slope might different than pi+, need
|
|
101
|
to do a proper fit to know for sure
|
|
102
|
|
|
103
|
- Dave: the purpose of these studies is to compare them to the LH2 boiling from
|
|
104
|
singles and see if there are any extra deadtime effects in the coincidence
|
|
105
|
data
|
|
106
|
- Garth: yes, it seems very likely there is an extra effect, as the pi+
|
|
107
|
14601-6 yield drop is somewhat larger than expected for LH2 boiling
|
|
108
|
- *NB* since extra deadtime effects should be rate-dependent, not
|
|
109
|
current-dependent, we need to disentangle the two effects, can't just apply
|
|
110
|
a larger "effective" LH2 boiling correction
|
|
111
|
- for a given scan, current and rate are directly correlated, it is only by
|
|
112
|
combining different settings at same current but different rate that we
|
|
113
|
can disentangle them
|
|
114
|
|
|
115
|
- *NB* preferred method would be to:
|
|
116
|
1) combine LH2 singles scans to get a more accurate LH2 boiling correction
|
|
117
|
2) apply this correction to the Coin Lumi data
|
|
118
|
3) overlay the different Coin scans, if they are consistent vs current, then
|
|
119
|
probably it is LH2 boiling
|
|
120
|
4) if they are inconsistent vs current, then overlay them vs different SHMS
|
|
121
|
rates to find the one that is most consistent, and use that to determine a
|
|
122
|
second correction to the Coin data
|
|
123
|
|
|
124
|
- overall, the studies look very encouraging, hopefully we can complete them
|
|
125
|
sometime in November
|
|
126
|
|
|
127
|
Alicia
|
|
128
|
------
|
|
129
|
BSA paper updates
|
|
130
|
- PRL sent our response to 2 of 3 referees on Oct 22, still waiting on a
|
|
131
|
response
|
|
132
|
|
|
133
|
Nacer
|
|
134
|
-----
|
|
135
|
KaonLT Heep singles @ 3.8 GeV
|
|
136
|
- last week analyzed HMS singles, obtained R_HMS=1.037 +/- 0.011
|
|
137
|
- found the problem with the SHMS singles replay
|
|
138
|
- needed to modify replay macro kin primary and kin secondary
|
|
139
|
- SHMS was analyzed as proton instead of electron, now get W peak
|
|
140
|
- also checked SHMS PID cuts
|
|
141
|
- ended up using analogous cuts to HMS electron selection: CAL and W cuts
|
|
142
|
- no HGC, Aero cuts applied
|
|
143
|
- obtain R_SHMS=1.017 +/- 0.016
|
|
144
|
|
|
145
|
- Dave: it is very interesting that R_HMS is high
|
|
146
|
- suggest to look at Heep singles runs for other KaonLT energies to
|
|
147
|
investigate any systematic effects
|
|
148
|
- most interested in settings where SHMS/HMS are close to Coin setting
|
|
149
|
kinematics
|
|
150
|
|
|
151
|
Next Meeting
|
|
152
|
-------------
|
|
153
|
- Thur Oct 31 @ 16:30 Eastern/14:30 Regina
|
|
154
|
- KaonLT will go first
|
|
155
|
|
|
156
|
|
|
157
|
|
|
158
|
|
|
159
|
|
|
160
|
|
|
161
|
|