Difference between revisions of "Feb. 22, 2008 Tracking CDC/FDC"

From GlueXWiki
Jump to: navigation, search
(Agenda)
m (Text replacement - "http://argus.phys.uregina.ca/gluex" to "https://halldweb.jlab.org")
 
(6 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 
==Agenda==
 
==Agenda==
  
# Status of MC resolution studies (David)
+
# Status of MC resolution studies [[Media:2008_02_22_TrackingStatus.pdf|PDF]] (David)
 
# EM Background rates (David)
 
# EM Background rates (David)
 
# Status of Kinematic Fitting (Matt)
 
# Status of Kinematic Fitting (Matt)
## Signal Leakage in pi/K (new studies) [http://argus.phys.uregina.ca/gluex/DocDB/0009/000971/002/Bellis_piKseparation_II.pdf pi/K Separation (GlueX doc-971)]  
+
## Signal Leakage in pi/K (new studies) [https://halldweb.jlab.org/DocDB/0009/000971/002/Bellis_piKseparation_II.pdf pi/K Separation (GlueX doc-971)]  
## Previous talk [http://argus.phys.uregina.ca/gluex/DocDB/0009/000971/002/Bellis_piKseparation.pdf pi/K Separation (GlueX doc-971)]  
+
## Previous talk [https://halldweb.jlab.org/DocDB/0009/000971/002/Bellis_piKseparation.pdf pi/K Separation (GlueX doc-971)]  
# Status of PWA anchoring (Paul, Alexander?)
+
# Status of PWA anchoring (Paul, Alexander)
 
##BNL-E852 tracking resolutions
 
##BNL-E852 tracking resolutions
 +
###[http://www.jlab.org/~ostrov/e852resol_pos.gif Pos-Res]
 +
###[http://www.jlab.org/~ostrov/e852resol_neg.gif Neg-Res]
 
##Three pi resonance generation & simulation
 
##Three pi resonance generation & simulation
 
###[http://www.jlab.org/Hall-D/software/wiki/images/b/bc/A2_plots.jpg  gamma p -> a2(1320) n] [[http://www.jlab.org/Hall-D/software/wiki/images/d/d8/A2_plots.pdf pdf file]]
 
###[http://www.jlab.org/Hall-D/software/wiki/images/b/bc/A2_plots.jpg  gamma p -> a2(1320) n] [[http://www.jlab.org/Hall-D/software/wiki/images/d/d8/A2_plots.pdf pdf file]]
Line 15: Line 17:
 
# CDC Hardware Status [[ 22 February 2008 ]]
 
# CDC Hardware Status [[ 22 February 2008 ]]
 
# FDC Hardware Status
 
# FDC Hardware Status
 +
# Update on segment-based track finding (Simon)[http://www.jlab.org/~staylor/track_finding_update_feb08.pdf]
  
 
==Participating==
 
==Participating==
Line 42: Line 45:
  
 
==Minutes==
 
==Minutes==
 +
''minutes by David Lawrence''
  
Attendees: ''TBD''
+
Attendees:  
 +
 
 +
@JLab: David L., Elton S., Tim W., Paul E., Simon T., Alex S., Mark I., Dan C.
 +
 
 +
@CMU: Curtis M., Yves, Matt B.
 +
 
 +
@FSU: Alexander O.
 +
 
 +
@IU: Beni Z., #delta;Matt S.
 +
 
 +
 
 +
===MC Resolutions From GEANT===
 +
David showed plots of the most recent tracking resolution results from the GEANT based Monte Carlo. These showed some noticable improvements
 +
especially in the forward direction. The fits included only CDC and FDC wires (no cathode strips). Plots of the residuals in the CDC and FDC as well as
 +
resolutions for 90<sup>o</sup> tracks for both parametric calculations and geant-based were shown. These indicated something was not understood
 +
for events with no smearing in both the CDC and FDC. David noted that the problem looks as though it may only be manifest in tracks that are nearly
 +
perpendicular to the wires. He will investigate further to track it down.
 +
 
 +
Also shown was a plot of the dE in the CDC for proton, kaon, and &pi; tracks with total momentum less than 1GeV/c. From a set of these, the threshold
 +
for reconstructing particles of the various types as a function of angle was made. This indicated about a 300 MeV/c threshold for protons and around
 +
200 MeV/c for Kaons both with some dependance on angle.
 +
 
 +
===EM Background Rates===
 +
David showed 4 plots indicating the EM background rates in the FDC, CDC, and TOF for both the March 2007 geometry and the current geometry.
 +
Dan noted that the center-most strips would be half-size so the values for the peak rates in those plots would need to be cut in half.
 +
 
 +
It was discussed whether additional studies would be needed prior to the March review. It was agreed that these would suffice with the one modification
 +
mentioned above.
 +
 
 +
===&pi;/K Separation via. Kinematic Fitting===
 +
Matt gave an update on the work he has been doing to study how well kinematic fitting could be used to separate &pi;'s and K's in a pair of similar
 +
channels from f<sub>2</sub>(1270) decays. The study's goal was to see how well events from the &pi;&pi;KK channel could be isolated from a
 +
background of &pi;&pi;&pi;&pi; events which are produced with a cross-section 100 times larger than the "signal".
 +
 
 +
Matt has implemented 2 of the sugestions from the last meeting:
 +
# Assume we cleanly identify the proton
 +
# Apply an "anticut" where the confidence levels of the other possible channels are below some limit (either 1% or 10%)
 +
 
 +
With these, he was able to show that with GlueX resolutions (March 2007) he could achieve a sample purity with only 4% background contaimination
 +
for the 10%CL anticut and <1% for the 1%CL anticut. These were at cost of throwing away about 50% and 65% respectivley of the true signals events.
 +
Note that these were done with no fiducial cuts and no information from the TOF. Matt will work on incorporating the TOF next as well as how
 +
they might reclaim the true events that were cut.
 +
 
 +
===BNL E852 Tracking Resolutions===
 +
Alexander showed some plots of the total and transverse momentum resolutions of positive and negative tracks from E852. The plots
 +
were made by looking at the difference between the track parameters before and after kinematic fitting. The resolutions were about
 +
0.35% for total momentum and 0.41% for transverse momentum (compared to the recent ~2% results from GlueX geant).
 +
 
 +
It was noted that since the tracks were involved in the kinematic fit, they could bias it leading to inaccurate results. Alex recalled that the
 +
average track momentum for the data set was around 6GeV.
 +
 
 +
Curtis suggested that invariant masses and missing masses might be the more appropriate values to compare to GlueX resolutions. Alex
 +
noted that for E852 3&pi; events, the missing mass (of the undetected target baryon) was 0.26(GeV/c)<sup>2</sup>
 +
 
 +
===3&pi; Resonance Generation/Simulation===
 +
Paul gave a nice update on some recent work he has done using some of the older code he and Curtis developed years ago. With it,
 +
he used (March2007) GlueX resolutions to smear a generated data sample that contained 3 background channels and 1 exotic
 +
&pi;(1600) channel. He was then able to include these channels and 14 others (which were not represented in the data set) in
 +
a fit to see if he could extract the mass distributions (and relative contributions) of the input channels. He showed some very nice
 +
mass plots indicating the mechanism was working quite well. The reconstructed channels were populated with mass distributions that had
 +
all of the major features of their inputs and the extra channels had distributions consistent with zero (as they should).
 +
 
 +
Paul also showed a missing mass<sup>2</sup> plot from the &pi(1600)&rarr;3&pi; with GlueX resolutions that had a width of
 +
0.22(GeV/c)<sup>2</sup> which is comparable to the E852 width.
 +
 
 +
Paul will continue to look at this and will incorporate fiducial acceptance as well as the new resolution functions (when available).
 +
 
 +
===Schedule of preparations===
 +
Elton showed a page he set up on the wiki for the March Tracking Review. The page contains a preliminary list of documents that will
 +
be supplied as background material to the reviewers, each with names of the responsible parties. Also included was a preliminary
 +
schedule for dry runs of the talks . He noted this was all preliminary and expected to change, but that we need to start arranging things
 +
and thinking about deadlines to make sure nothing falls through the cracks.
 +
 
 +
===CDC Hardware Status===
 +
Yves reported that they are still tracking down problems with the gain of the shaper boards. A new board is on the way which will hopefully
 +
shed light on the issue.
 +
 
 +
Yves also reported a problem with the drift time spectrum coming out of the prototype. This could be related to the timing of the fADCs.
 +
He will check that and if needed, recheck the Garfield simulation.
 +
 
 +
Curtis noted that they also have plans to remeasure the wire tension which has not been checked for a couple of years.
 +
 
 +
===FDC Hardware Status===
 +
 
 +
===FDC Track Finder Status===
 +
Simon reported one progress on the segment based finder in the FDC. This is now the default track finder for the FDC. He has made
 +
improvements in the efficiency of low momentum tracks. There are still problems with inefficiencies of tracks from 11 to 15 degrees.
 +
Tracks at angles greater than 15 degrees hit only 1 or two packages.
 +
Simon also showed a plot indicating the transverse momentum resolution of track candidates at angle greater than 15 degrees
 +
is more than doubled.
 +
 
 +
It was suggested that a few of us get together early next week so we could decide by Wednesday whether effort needs to be directed
 +
towards a hybrid finder before the review.

Latest revision as of 12:38, 14 March 2017

Agenda

  1. Status of MC resolution studies PDF (David)
  2. EM Background rates (David)
  3. Status of Kinematic Fitting (Matt)
    1. Signal Leakage in pi/K (new studies) pi/K Separation (GlueX doc-971)
    2. Previous talk pi/K Separation (GlueX doc-971)
  4. Status of PWA anchoring (Paul, Alexander)
    1. BNL-E852 tracking resolutions
      1. Pos-Res
      2. Neg-Res
    2. Three pi resonance generation & simulation
      1. gamma p -> a2(1320) n [pdf file]
      2. gamma p > X n where X contains: a1(1170), a2(1320), pi2(1670), pi1(1600) [pdf file]
      3. rebirth of pwa simulations
  5. DC and PID Review Schedule for preparations (Elton)
  6. CDC Hardware Status 22 February 2008
  7. FDC Hardware Status
  8. Update on segment-based track finding (Simon)[1]

Participating

Location

The meeting will be at 11:00am in Cebaf Center F324-325.

Telephone

To connect by telephone: 1.) dial:

800-377-8846 : US
888-276-7715 : Canada
302-709-8424 : International (not a free call!)

2.) enter participant code: 39527048# (remember the "#")

Video Conferencing

The room at JLab does not have video conferencing equipment installed.


Minutes

minutes by David Lawrence

Attendees:

@JLab: David L., Elton S., Tim W., Paul E., Simon T., Alex S., Mark I., Dan C.

@CMU: Curtis M., Yves, Matt B.

@FSU: Alexander O.

@IU: Beni Z., #delta;Matt S.


MC Resolutions From GEANT

David showed plots of the most recent tracking resolution results from the GEANT based Monte Carlo. These showed some noticable improvements especially in the forward direction. The fits included only CDC and FDC wires (no cathode strips). Plots of the residuals in the CDC and FDC as well as resolutions for 90o tracks for both parametric calculations and geant-based were shown. These indicated something was not understood for events with no smearing in both the CDC and FDC. David noted that the problem looks as though it may only be manifest in tracks that are nearly perpendicular to the wires. He will investigate further to track it down.

Also shown was a plot of the dE in the CDC for proton, kaon, and π tracks with total momentum less than 1GeV/c. From a set of these, the threshold for reconstructing particles of the various types as a function of angle was made. This indicated about a 300 MeV/c threshold for protons and around 200 MeV/c for Kaons both with some dependance on angle.

EM Background Rates

David showed 4 plots indicating the EM background rates in the FDC, CDC, and TOF for both the March 2007 geometry and the current geometry. Dan noted that the center-most strips would be half-size so the values for the peak rates in those plots would need to be cut in half.

It was discussed whether additional studies would be needed prior to the March review. It was agreed that these would suffice with the one modification mentioned above.

π/K Separation via. Kinematic Fitting

Matt gave an update on the work he has been doing to study how well kinematic fitting could be used to separate π's and K's in a pair of similar channels from f2(1270) decays. The study's goal was to see how well events from the ππKK channel could be isolated from a background of ππππ events which are produced with a cross-section 100 times larger than the "signal".

Matt has implemented 2 of the sugestions from the last meeting:

  1. Assume we cleanly identify the proton
  2. Apply an "anticut" where the confidence levels of the other possible channels are below some limit (either 1% or 10%)

With these, he was able to show that with GlueX resolutions (March 2007) he could achieve a sample purity with only 4% background contaimination for the 10%CL anticut and <1% for the 1%CL anticut. These were at cost of throwing away about 50% and 65% respectivley of the true signals events. Note that these were done with no fiducial cuts and no information from the TOF. Matt will work on incorporating the TOF next as well as how they might reclaim the true events that were cut.

BNL E852 Tracking Resolutions

Alexander showed some plots of the total and transverse momentum resolutions of positive and negative tracks from E852. The plots were made by looking at the difference between the track parameters before and after kinematic fitting. The resolutions were about 0.35% for total momentum and 0.41% for transverse momentum (compared to the recent ~2% results from GlueX geant).

It was noted that since the tracks were involved in the kinematic fit, they could bias it leading to inaccurate results. Alex recalled that the average track momentum for the data set was around 6GeV.

Curtis suggested that invariant masses and missing masses might be the more appropriate values to compare to GlueX resolutions. Alex noted that for E852 3π events, the missing mass (of the undetected target baryon) was 0.26(GeV/c)2

3π Resonance Generation/Simulation

Paul gave a nice update on some recent work he has done using some of the older code he and Curtis developed years ago. With it, he used (March2007) GlueX resolutions to smear a generated data sample that contained 3 background channels and 1 exotic π(1600) channel. He was then able to include these channels and 14 others (which were not represented in the data set) in a fit to see if he could extract the mass distributions (and relative contributions) of the input channels. He showed some very nice mass plots indicating the mechanism was working quite well. The reconstructed channels were populated with mass distributions that had all of the major features of their inputs and the extra channels had distributions consistent with zero (as they should).

Paul also showed a missing mass2 plot from the &pi(1600)→3π with GlueX resolutions that had a width of 0.22(GeV/c)2 which is comparable to the E852 width.

Paul will continue to look at this and will incorporate fiducial acceptance as well as the new resolution functions (when available).

Schedule of preparations

Elton showed a page he set up on the wiki for the March Tracking Review. The page contains a preliminary list of documents that will be supplied as background material to the reviewers, each with names of the responsible parties. Also included was a preliminary schedule for dry runs of the talks . He noted this was all preliminary and expected to change, but that we need to start arranging things and thinking about deadlines to make sure nothing falls through the cracks.

CDC Hardware Status

Yves reported that they are still tracking down problems with the gain of the shaper boards. A new board is on the way which will hopefully shed light on the issue.

Yves also reported a problem with the drift time spectrum coming out of the prototype. This could be related to the timing of the fADCs. He will check that and if needed, recheck the Garfield simulation.

Curtis noted that they also have plans to remeasure the wire tension which has not been checked for a couple of years.

FDC Hardware Status

FDC Track Finder Status

Simon reported one progress on the segment based finder in the FDC. This is now the default track finder for the FDC. He has made improvements in the efficiency of low momentum tracks. There are still problems with inefficiencies of tracks from 11 to 15 degrees. Tracks at angles greater than 15 degrees hit only 1 or two packages. Simon also showed a plot indicating the transverse momentum resolution of track candidates at angle greater than 15 degrees is more than doubled.

It was suggested that a few of us get together early next week so we could decide by Wednesday whether effort needs to be directed towards a hybrid finder before the review.