BCAL Readout Segmentation Task group Meeting: Aug 5, 2011

From GlueXWiki
Revision as of 15:48, 24 February 2017 by Marki (Talk | contribs) (Text replacement - "http://argus.phys.uregina.ca/cgi-bin/private" to "https://halldweb.jlab.org/doc-private")

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Meeting Info

Time:

  • 3:00 JLab
  • 1:00 Regina

Location:

  • CC F326-327

Vid-con:

  • ESNet 8542553

Agenda

Review action items from previous meeting

  • Continue timing studies based on electronic pulse shapes (David)
    • See today's slides
  • Check with Matt on meeting time change (David)
    • Done


New Agenda items

Minutes

Participants

David L., Andrei S., Irina S., Zisis P. George L., Elton S., Beni Z.

Review action items from previous meeting

The 2 action items were quickly reviewed, but no discussion was needed.

New Agenda items

There was brief discussion of the document Zisis sent out earlier in the day that he and Andrei authored, summarizing the work of the task group since it's formation and even before then back to March. No one had had a chance to fully digest it yet so the discussion was brief and mostly deferred until they had. Elton did make one comment on how much emphasis we should place on the earlier studies that were known to use reconstruction software lacking optimization for the course segmentation scheme. This discussion did not fully develop, however for the reason noted above.

Zisis encouraged everyone to read the document and send any feedback. He will be gone to Greece for the next 2 weeks, but will have contact via e-mail.

BCAL Signal Timing Distributions

Dave showed some slides indicating the current status of his ongoing work to study the timing resolutions of the various BCAL segmentation schemes.

Slides 3 and 4: These show the total electronic pulse (black) and the contributions from 100ps bins of the input signal distributions (colored). also shown is the point at which the electronic pulse crosses the 44.7mV threshold. The point of these plots is to give an indication of the spread in input signal times that make a significant contribution to the electronic pulse height at the point when it crosses threshold.

  • These use the slower, 15ns rise time pulse shape.
  • Dave noted that on the top plot on slide 4, there are more or less equal contributions to the electronic pulse from the green and yellow bins at threshold crossing time, even though the light from those bins arrived ~1ns apart in time.


Slides 5-7: More slides showing the breakdown of single cell, single event responses to hdgeant generated showers. These include the new 5ns rise time pulse shape extracted from Fernando's note posted the previous day.


Slides 8-11: These showed time vs. fADC value distributions before and after a timewalk correction was applied. Only 12o and 20o were shown.

  • 90o were also produced, but not shown
  • Dave noted having problems getting timewalk functions to fit well for all layers, all angles, all segmentation schemes.
  • Andrei noted that it is not uncommon to require multiple passes to complete a timewalk calibration
    • Dave concurred and noted that some of the wiggles at low energy in the corrected plots may indicate additional degrees of freedom are needed


Slide 12: This slide is incomplete. It was intended to illustrate the 5 different segmentation schemes that will be studied.

  • Code is in place that implements all 5 schemes, but finalizing the timewalk parameters for each is still needed.
  • Zisis asked if the "fine" segmentation scheme would be included as a point of reference. Dave indicated that it would.
    • Dave brought up that the "fine" segmentation scheme itself is basically off the table as an option due to cost constraints on readout channels and available real estate on the electronics boards. This was generally agreed to, but the consensus was that maintaining its inclusion in further studies was worthwhile as a point of reference.


Toy Monte Carlo Studies

Elton has posted GlueX-doc-1796 on the DocDB. He did not show any slides since the conclusions drawn in the note are basically what was shown at the meeting last week. Mainly, that for tracks at θ<20o, one essentially gets an extra measurement from having a 1-2-3 configuration for summing the inner layers as opposed to a 3-3. In addition he has:

  • Added a 12o angle so that he has events for 90o, 20o, and 12o
  • Added a 2-2-2-4 configuration
  • Asked Fernando to look into ramifications of adding another layer (no word back yet)

Elton also reiterated the possibility discussed at the last meeting of using jumpers to select between a 1-2-3 and a 3-3 configuration for the inner layers. This would allow the board design and procurement to proceed while additional studies are done.

Everyone is encouraged to read Elton's note and send feedback to him. An opportunity to comment on/discuss the note will be presented at the next meeting.

Preliminary look on the energy resolution in BCAL readout cells

Andrei showed some slides from a study they are doing to try and extract the single cell energy resolution for the BCAL using FLUKA. This is motivated by the fact that the same parameterization (based on deposited energy) is used to smear the energy for each cell in mcsmear. This is considered incorrect due to the parameters being derived from the whole shower resolution and not the single cell resolution.

There was some discussion on the slides:


Slide 4: This slide shows the energy deposition in the fibers for individual cells over many events with 1GeV γ's incident on the same point of the array (indicated at the top of slide).

  • Beni pointed out that there appears to be a spike at E=0, even for cells that should see a large fraction of the shower (e.g. second column from left, 3rd row from top). Andrei agreed that this seemed strange and would look into it.
  • The RMS values from these plots are used to represent resolutions on the next slide


Slide 5: This slide shows the single cell relative energy resolution as a function energy deposited in the cell.

  • Cell resolutions come from RMS values on previous slide
  • Total energy in cell is estimated from energy in fibers and sampling fraction (11.88% for this study)
  • Dave noted that the RMS values are driven by shower fluctuations which are already included in hdgeant
    • (Andrei noted that he thought Dave indicated shower fluctuations were not included in the simulation at
      an earlier meeting. This may have been due to a misunderstanding during that earlier discussion as to what
      is included in hdgeant and what is in mcsmear)
  • Andrei's main conclusion here was that the single cell energy resolution (in %) is worse than that of the total shower.
    Therefore, we should not be using the total shower parameters when doing energy smearing on the individual cells.


Slide 6: This plot shows the total energy deposited in the fibers for a 2GeV incident photon. The plot shows a Gaussian fit as well as the RMS. This emphasizes the difference between the σ of the fit (9.35MeV) and the RMS (17.76 MeV).


Slide 7: Preliminary conclusions. The lookup table refers to the likelihood that the single cell resolution function/parameters will depend on it's proximity relative to the shower. The significance of this dependence is not yet determined quantitatively.


Dave noted that one of the earlier discussion points was the sampling fluctuation as a function of angle, depth, energy,... This is coupled with the current study that also includes shower fluctuations. If the current Univ. of Regina study is able to extract these, Dave will help incorporate them into the sim-recon source tree. Andrei agreed to keep looking into this.


Action Items

  • Read Zisis/Andrei's note summarizing the status of the task group's activities and progress towards it goal. Send comments to Zisis. (All)
  • Read Elton's Note on parametric M.C. and send comments (All)
  • Continue detailed timing resolution studies on BCAL (David)
  • Continue development of single cell resolution functions (Andrei/Irina)