Difference between revisions of "BCAL Reconstruction Meeting 2015-10-01"

From GlueXWiki
Jump to: navigation, search
m (Text replacement - "http://argus.phys.uregina.ca/gluex" to "https://halldweb.jlab.org")
 
Line 40: Line 40:
 
## Layer efficiencies (Ahmed)
 
## Layer efficiencies (Ahmed)
 
## Up/down gain ratios for spring data (Mark)
 
## Up/down gain ratios for spring data (Mark)
## [http://argus.phys.uregina.ca/gluex/DocDB/0028/002815/001/mich_elect_oct1.pdf Muon decays] (Will)
+
## [https://halldweb.jlab.org/DocDB/0028/002815/001/mich_elect_oct1.pdf Muon decays] (Will)
 
# Reconstruction
 
# Reconstruction
 
# Simulations
 
# Simulations
 
## CPPsim (Tegan)
 
## CPPsim (Tegan)
## [http://argus.phys.uregina.ca/gluex/DocDB/0028/002812/001/bcal_recalibration_151001.pdf Re-Calibration of BCal with Cosmics] (Andrei)   
+
## [https://halldweb.jlab.org/DocDB/0028/002812/001/bcal_recalibration_151001.pdf Re-Calibration of BCal with Cosmics] (Andrei)   
 
# Any other business
 
# Any other business
  
Line 61: Line 61:
 
## Layer efficiencies (Ahmed): 1st draft of report is almost ready to be uploaded to DocDB.  Couple of variables are still to be studied in the future. ES: mod 7 vs 8?  Ahmed, no big smoking gun.  MD: simrecon emulation faithful? DL: no.  This is an issue for our own emulation.  Fernando's stuff: it will be deployed by April. Need to have ability to have more than one emulation code. Mike might be able to work on it in Nov. There is a group discussing this but emulation is nobody's top priority.  Mike's name on it.
 
## Layer efficiencies (Ahmed): 1st draft of report is almost ready to be uploaded to DocDB.  Couple of variables are still to be studied in the future. ES: mod 7 vs 8?  Ahmed, no big smoking gun.  MD: simrecon emulation faithful? DL: no.  This is an issue for our own emulation.  Fernando's stuff: it will be deployed by April. Need to have ability to have more than one emulation code. Mike might be able to work on it in Nov. There is a group discussing this but emulation is nobody's top priority.  Mike's name on it.
 
## Up/down gain ratios for spring data (Mark): there was a replay which used improved code by adding Mike's cut for late and low hits but it did not fix everything, as it was done completely with TDCs for which the calibration is not completely done by Mike.  Mark could ask for a new replay.  Construct z position from ADC times instead of z position reported from Points.  Study comparing z-point vs z-track in ADC vs TDC.  (George?)  Ask Kei to replay the five interesting runs used in the BCAL calibration for the pi0 sample (10 runs). Mark does not expect a huge improvement; not a big change from spring to fall.  ES: look at LED data with both ADCs and TDCs.  Mark's AttenLen are not loaded into ccdb but are in a log entry.  Will has not seen any improvement in pi0 width with new ratios and old AttenLen.  He will ook at new+new and see what happens.   
 
## Up/down gain ratios for spring data (Mark): there was a replay which used improved code by adding Mike's cut for late and low hits but it did not fix everything, as it was done completely with TDCs for which the calibration is not completely done by Mike.  Mark could ask for a new replay.  Construct z position from ADC times instead of z position reported from Points.  Study comparing z-point vs z-track in ADC vs TDC.  (George?)  Ask Kei to replay the five interesting runs used in the BCAL calibration for the pi0 sample (10 runs). Mark does not expect a huge improvement; not a big change from spring to fall.  ES: look at LED data with both ADCs and TDCs.  Mark's AttenLen are not loaded into ccdb but are in a log entry.  Will has not seen any improvement in pi0 width with new ratios and old AttenLen.  He will ook at new+new and see what happens.   
## [http://argus.phys.uregina.ca/gluex/DocDB/0028/002815/001/mich_elect_oct1.pdf Muon decays] (Will): time diff between early and late hits gives 1st plot; 2nd plot unattenuated energy in cell, some cuts were needed to clean up things; mean at 33MeV but no cutoff at 52MeV; study cuts more. Plot 3, 4: with pi0 minimization not much energy collected in Layer 4.  Plots 3 and 4 are not all the data, some runs are still being processed.  Look at tail more closely with more data; layer 4 tail looks much stronger, which might be related to Layer 4 calibrations being too high. Last plot: red - most energy in layer 1 (leader of the pack); the distributions do not line up with each other; strange: Layer 3 in middle of caller gives lowest energy; layer 1 can have a lot of energy at forward angles.  Two channel cut min requirement on the energy of the electron; but Layer 4 is so big that things may be lost based on this cut.  Look at noon and 6 vs 3 and 9 o' clock; there should be no azimuthal effect although track length through Layer 4 is big. Sector width in layer 4 is bigger.  Much improved over distributions from two weeks ago.  Difference? Will does not know as he accidentally deleted the plugin and rewrote it. :-)
+
## [https://halldweb.jlab.org/DocDB/0028/002815/001/mich_elect_oct1.pdf Muon decays] (Will): time diff between early and late hits gives 1st plot; 2nd plot unattenuated energy in cell, some cuts were needed to clean up things; mean at 33MeV but no cutoff at 52MeV; study cuts more. Plot 3, 4: with pi0 minimization not much energy collected in Layer 4.  Plots 3 and 4 are not all the data, some runs are still being processed.  Look at tail more closely with more data; layer 4 tail looks much stronger, which might be related to Layer 4 calibrations being too high. Last plot: red - most energy in layer 1 (leader of the pack); the distributions do not line up with each other; strange: Layer 3 in middle of caller gives lowest energy; layer 1 can have a lot of energy at forward angles.  Two channel cut min requirement on the energy of the electron; but Layer 4 is so big that things may be lost based on this cut.  Look at noon and 6 vs 3 and 9 o' clock; there should be no azimuthal effect although track length through Layer 4 is big. Sector width in layer 4 is bigger.  Much improved over distributions from two weeks ago.  Difference? Will does not know as he accidentally deleted the plugin and rewrote it. :-)
 
# Reconstruction
 
# Reconstruction
 
# Simulations
 
# Simulations
 
## CPPsim (Tegan): worked eta-prime decay after the eta plugin.  Plugin did not work;  emailed it to Justin who will have a look after the DIRC TDR is done.
 
## CPPsim (Tegan): worked eta-prime decay after the eta plugin.  Plugin did not work;  emailed it to Justin who will have a look after the DIRC TDR is done.
## [http://argus.phys.uregina.ca/gluex/DocDB/0028/002812/001/bcal_recalibration_151001.pdf Re-Calibration of BCal with Cosmics] (Andrei): Objectives are to attempt to improve calibration of cosmic muons using external trigger scintillators and also calculate the crossing length of the tracks and normalize the energy deposition.  New work: fit particle track to signal profile so as to clean data set and make corrections (colour scale goes from yellow to bright red for energy deposition).  Data is from the runs Tegan took in December 2014 (non-zero external triggers, comics trigger done by Sasha, no magnetic field). Slide 3: there are not a lot of events like the ones on the right; left plot is a shower with 5 tracks. Slide 4 middle plot is very weird; sometimes track only on top of BCAL sometimes only on bottom hits are seen. Slide 5 good tracks with BCAL data only, through its sides, without chambers: how precise are those tracks? Answer, tracking is pretty accurate (at few mm level). Fits are done in 2D cause the external triggers narrow down acceptance. Slide 6: red - old method just with energy deposition, blue - new method using tracks. Left: module 1 (9 o' clock) don't expect much change, right is module 13 (10 o' clock) large improvement seen. Statistics are low because only files with external trigger info used (many runs had zeros for their channels in the DAQ). Slide 7: module 11 (around 11 0' clock) and improvement here.  Calibration/normalization can be extracted more reliably (this is the hope) and better estimation of efficiencies by using the tracks.  Efficiency is low, but perhaps related to "gaps" in Edep, where some cells are visibly inefficient.  More simulation work needed and comics stats at real-operation conditions (temp, bias, threshold) for spring 2016.  The runs analyzed were only a few hours of running (70% in 6 hours). It might be worth to look at spring 2015 conditions comics (i.e. take new ones).  ES: efficiency depends on orientation. AS: cannot have inner layer hits w/o outer layer hits, but the reverse can occur.  Method of few-mm  track uncertainty affects outer layers less, so method is complimentary to old Edep method that favours inner layers. MD: forget external counters, which would open up huge statistics and can stuff z-dependence.  Do fit in 3D which requires timing, more complex.  We have three competing calibration techniques: this one can allow z-study in a symmetric way.
+
## [https://halldweb.jlab.org/DocDB/0028/002812/001/bcal_recalibration_151001.pdf Re-Calibration of BCal with Cosmics] (Andrei): Objectives are to attempt to improve calibration of cosmic muons using external trigger scintillators and also calculate the crossing length of the tracks and normalize the energy deposition.  New work: fit particle track to signal profile so as to clean data set and make corrections (colour scale goes from yellow to bright red for energy deposition).  Data is from the runs Tegan took in December 2014 (non-zero external triggers, comics trigger done by Sasha, no magnetic field). Slide 3: there are not a lot of events like the ones on the right; left plot is a shower with 5 tracks. Slide 4 middle plot is very weird; sometimes track only on top of BCAL sometimes only on bottom hits are seen. Slide 5 good tracks with BCAL data only, through its sides, without chambers: how precise are those tracks? Answer, tracking is pretty accurate (at few mm level). Fits are done in 2D cause the external triggers narrow down acceptance. Slide 6: red - old method just with energy deposition, blue - new method using tracks. Left: module 1 (9 o' clock) don't expect much change, right is module 13 (10 o' clock) large improvement seen. Statistics are low because only files with external trigger info used (many runs had zeros for their channels in the DAQ). Slide 7: module 11 (around 11 0' clock) and improvement here.  Calibration/normalization can be extracted more reliably (this is the hope) and better estimation of efficiencies by using the tracks.  Efficiency is low, but perhaps related to "gaps" in Edep, where some cells are visibly inefficient.  More simulation work needed and comics stats at real-operation conditions (temp, bias, threshold) for spring 2016.  The runs analyzed were only a few hours of running (70% in 6 hours). It might be worth to look at spring 2015 conditions comics (i.e. take new ones).  ES: efficiency depends on orientation. AS: cannot have inner layer hits w/o outer layer hits, but the reverse can occur.  Method of few-mm  track uncertainty affects outer layers less, so method is complimentary to old Edep method that favours inner layers. MD: forget external counters, which would open up huge statistics and can stuff z-dependence.  Do fit in 3D which requires timing, more complex.  We have three competing calibration techniques: this one can allow z-study in a symmetric way.
 
# Any other business: ES: Andrei generated Michel plots.  Can they be produced with threshold?  AS: yes, but what threshold?  Irina produces these. ES will send suggestions (e.g. 1 MeV/channel), and possibly on selection criteria. Elton, Will, Andrei will do this offline.
 
# Any other business: ES: Andrei generated Michel plots.  Can they be produced with threshold?  AS: yes, but what threshold?  Irina produces these. ES will send suggestions (e.g. 1 MeV/channel), and possibly on selection criteria. Elton, Will, Andrei will do this offline.

Latest revision as of 12:40, 14 March 2017

Video Conferencing Information

Meeting Time: 11:15 a.m. EDT

  1. To join via a Web Browser, go to the page [1] https://bluejeans.com/907185247.
  2. To join via Polycom room system go to the IP Address: 199.48.152.152 (bjn.vc) and enter the meeting ID: 907185247.
  3. To join via phone, use one of the following numbers and the Conference ID: 907185247.
    • US or Canada: +1 408 740 7256 or
    • US or Canada: +1 888 240 2560
  4. Upon connection all microphones are automatically muted. To unmute your mike on a Polycom or equivalent unit, enter *4. Unmuting on a computer is trivial as there is a microphone button than can be clicked.
  5. More information on connecting to bluejeans is available.

Participant Direct Lines

  • JLab Phone: in CC F326 is 757-269-6460 (usual room)
  • JLab Phone in CC L207 is 757-269-7084
  • Phone in the Regina Video-conference Suite is 306-585-4204

References

  1. BCAL Reconstruction Issues
  2. BCAL Reconstruction Algorithms
  3. Will's reconstructed energy info

Action Items

  1. Make MC plots that show E vs Egen and resolution (Tegan)
  2. Low priority items
    1. z-coordinate determination from up/down amplitude ratio
    2. Sampling fraction tables (Tegan, Andrei)
    3. Code Cleanup

Agenda

  1. Announcements
  2. Action Items
  3. Commissioning Efforts
    1. Some comments on effective speed (George).
    2. Time-walk corrections (Noemi)
    3. Layer efficiencies (Ahmed)
    4. Up/down gain ratios for spring data (Mark)
    5. Muon decays (Will)
  4. Reconstruction
  5. Simulations
    1. CPPsim (Tegan)
    2. Re-Calibration of BCal with Cosmics (Andrei)
  6. Any other business

Minutes

Attendees: Mark, Elton, David (JLab), Will M., Curtis, Mike (CMU), Sean (NWU), George (Athens), Ahmed, Tegan, Noemi, Andrei, Zisis (Regina)’’

  1. Announcements:
    1. Sean: sims of spring data with new BCAL code. Folks should look at it.
    2. Elton: rain/hurricane at NN this weekend.
    3. Zisis: reminder that Mark is presenting the BCAL work from JLab/CMU and Zisis from Athens/Regina. Everyone: please prepare relevant plots and info.
  2. Action Items
  3. Commissioning Efforts
    1. Some comments on effective speed (George). Non-linearity in z for comics run 3221, +/-50cm spread at z<200, which gives an ellipse on the fitting! Above z>280 it is linear. Different structure see in Run 3134 B=0, linear up to 300cm. For calibration constants George plans to restrict extraction to linear regions. Mike: for comics do you use pTrackFind cosmics=1 or do zero field tracking? George uses the latter, but that one assumes that tracks come from the target which may explain weird track seen (Mike makes no claim to this effect!). The problem is in the CDC. Comics use a different fitter; no FDC tracks are used, just CDC. Mark: pTrackFind is parametrized in an odd way, so the track matching may have to be done by George: develop procedure. George sent time offsets to Mike who is uploading them to ccdb. and is now looking at Physics Analysis.
    2. Time-walk corrections (Noemi): Mike's algorithm being used, uncertainties in the data (apply cuts) to clean up things.
    3. Layer efficiencies (Ahmed): 1st draft of report is almost ready to be uploaded to DocDB. Couple of variables are still to be studied in the future. ES: mod 7 vs 8? Ahmed, no big smoking gun. MD: simrecon emulation faithful? DL: no. This is an issue for our own emulation. Fernando's stuff: it will be deployed by April. Need to have ability to have more than one emulation code. Mike might be able to work on it in Nov. There is a group discussing this but emulation is nobody's top priority. Mike's name on it.
    4. Up/down gain ratios for spring data (Mark): there was a replay which used improved code by adding Mike's cut for late and low hits but it did not fix everything, as it was done completely with TDCs for which the calibration is not completely done by Mike. Mark could ask for a new replay. Construct z position from ADC times instead of z position reported from Points. Study comparing z-point vs z-track in ADC vs TDC. (George?) Ask Kei to replay the five interesting runs used in the BCAL calibration for the pi0 sample (10 runs). Mark does not expect a huge improvement; not a big change from spring to fall. ES: look at LED data with both ADCs and TDCs. Mark's AttenLen are not loaded into ccdb but are in a log entry. Will has not seen any improvement in pi0 width with new ratios and old AttenLen. He will ook at new+new and see what happens.
    5. Muon decays (Will): time diff between early and late hits gives 1st plot; 2nd plot unattenuated energy in cell, some cuts were needed to clean up things; mean at 33MeV but no cutoff at 52MeV; study cuts more. Plot 3, 4: with pi0 minimization not much energy collected in Layer 4. Plots 3 and 4 are not all the data, some runs are still being processed. Look at tail more closely with more data; layer 4 tail looks much stronger, which might be related to Layer 4 calibrations being too high. Last plot: red - most energy in layer 1 (leader of the pack); the distributions do not line up with each other; strange: Layer 3 in middle of caller gives lowest energy; layer 1 can have a lot of energy at forward angles. Two channel cut min requirement on the energy of the electron; but Layer 4 is so big that things may be lost based on this cut. Look at noon and 6 vs 3 and 9 o' clock; there should be no azimuthal effect although track length through Layer 4 is big. Sector width in layer 4 is bigger. Much improved over distributions from two weeks ago. Difference? Will does not know as he accidentally deleted the plugin and rewrote it. :-)
  4. Reconstruction
  5. Simulations
    1. CPPsim (Tegan): worked eta-prime decay after the eta plugin. Plugin did not work; emailed it to Justin who will have a look after the DIRC TDR is done.
    2. Re-Calibration of BCal with Cosmics (Andrei): Objectives are to attempt to improve calibration of cosmic muons using external trigger scintillators and also calculate the crossing length of the tracks and normalize the energy deposition. New work: fit particle track to signal profile so as to clean data set and make corrections (colour scale goes from yellow to bright red for energy deposition). Data is from the runs Tegan took in December 2014 (non-zero external triggers, comics trigger done by Sasha, no magnetic field). Slide 3: there are not a lot of events like the ones on the right; left plot is a shower with 5 tracks. Slide 4 middle plot is very weird; sometimes track only on top of BCAL sometimes only on bottom hits are seen. Slide 5 good tracks with BCAL data only, through its sides, without chambers: how precise are those tracks? Answer, tracking is pretty accurate (at few mm level). Fits are done in 2D cause the external triggers narrow down acceptance. Slide 6: red - old method just with energy deposition, blue - new method using tracks. Left: module 1 (9 o' clock) don't expect much change, right is module 13 (10 o' clock) large improvement seen. Statistics are low because only files with external trigger info used (many runs had zeros for their channels in the DAQ). Slide 7: module 11 (around 11 0' clock) and improvement here. Calibration/normalization can be extracted more reliably (this is the hope) and better estimation of efficiencies by using the tracks. Efficiency is low, but perhaps related to "gaps" in Edep, where some cells are visibly inefficient. More simulation work needed and comics stats at real-operation conditions (temp, bias, threshold) for spring 2016. The runs analyzed were only a few hours of running (70% in 6 hours). It might be worth to look at spring 2015 conditions comics (i.e. take new ones). ES: efficiency depends on orientation. AS: cannot have inner layer hits w/o outer layer hits, but the reverse can occur. Method of few-mm track uncertainty affects outer layers less, so method is complimentary to old Edep method that favours inner layers. MD: forget external counters, which would open up huge statistics and can stuff z-dependence. Do fit in 3D which requires timing, more complex. We have three competing calibration techniques: this one can allow z-study in a symmetric way.
  6. Any other business: ES: Andrei generated Michel plots. Can they be produced with threshold? AS: yes, but what threshold? Irina produces these. ES will send suggestions (e.g. 1 MeV/channel), and possibly on selection criteria. Elton, Will, Andrei will do this offline.