BCAL Readout Segmentation Task group Meeting: Jun 9, 2011

From GlueXWiki
Jump to: navigation, search

Meeting Info

Time:

  • 11:30 JLab
  • 9:30 Regina

Location:

  • CC F228

Vid-con:

  • ESNet 8542553

Agenda

  • Review action items from previous meeting
  • BCAL Calibration (David L.)
  • Assign tasks for next week

Minutes

Participants

Andrei S., Irina S., David L., Elton S.

Action Items from Last week

The action items from last week were briefly reviewed. Status summaries are below:

  • Establish new tagged release that includes all recent changes to BCAL code and calib (David)
  • Create new energy calibration function that includes pedestal offset due to dark hits (Irina/Andrei ; David)
    • Ongoing. See slides shown at today's meeting
  • Repeat resolution studies for two segmentation schemes using new calibration (David)
    • Initial pass completed, but issues observed and currently under study
  • Repeat Sascha's M.C. calculation of BCAL contribution to event size (David)
    • Ongoing. May report on next week

BCAL Calibration

David showed some slides outlining the status of his effort to calibrate the BCAL.

Slide 2

Recent code modifications to the area of the repository Dave set up for the current calibration studies were reviewed. Some discussion was had over the addition of the Poisson sampling to the number of photo-electrons derived from the deposited energy. Andrei noted that the currently existing sampling fluctuations may already include this effect so adding it in may be adding it twice. Dave agreed to send an e-mail to Andrei pointing out the code where this is being done so they can have a look to verify it is doing the correct thing.

Slide 3

The calibration method used previously was reviewed. It was noted that there were two inconsistencies with the method implemented in the code and that outlined in Blake's thesis

  • The offset term "B" was not implemented in the current code
  • The current code does not have separate values for the ends of the module in the calibDB as described in the thesis.

Slide 4

The fit parameters as a function of z were shown for the KLOE algorithm along with fits. The fits contained significantly more parameters than the 4 used for a 3rd order polynomial.

Slide 5

Same as slide 4, but for Matt's algorithm which is currently under development. It was noted that these were produced first, in error. The apparently more significant structure in this eventually motivated the decision to switch to a lookup table for the calibration rather than use three ~20 parameter functions. It was only after the table mechanism was virtually complete that it was realized that these were not produced by the KLOE code.

Slide 6

Gaussian fits to slices in reconstructed z (n.b. *not* z_entry) were used to derive the systematic and statistical errors shown. It was pointed out that the upper left plot deviated significantly from the flat y=1 plot expected. This will require further investigation.

  • Andrei noted that they have been seeing non-trivial structure in the reconstructed values so a profile histogram which uses an average may be problematic. Dave agreed to look at this.

Slide 7

Pull distributions derived from the same data used for slide 6 were shown. Rather, the results of Gaussian fits to slices in z of the pull distributions. The pulls were calculated using the "nominal" energy resolution given by the equation on the slide.

General Discussion

There was quite a bit of discussion as to whether the cell-level sampling fluctuations were being implemented adequately. In particular, Andrei was concerned that the 4.1%/sqrt(E) + 1.3% of intrinsic resolution currently used was not adequately treating the case of forward going particles with small angles where most of the energy was deposited in only a few inner cells. We eventually came to an agreement on a course of action:

  • Elton will write up how he understands the current smearing to be implemented, in particular the sampling fluctuations. He will also give his recommendation on how this should be done (if any different)
  • Andrei will think about how one would realistically determine and then include the proper resolution effects of the sampling fluctuations in the hdgeant/mcsmear/hd_root system. He will circulate this to the task group, possibly after Elton has circulated his thoughts so Andrei's thoughts can be merged in.
  • David will run some simulations that have the sampling fluctuations turned off to measure the intrinsic resolution of the hdgeant/mcsmear chain. This will be compared to the 2006 beam test data and if needed, the parameters in the calibDB (currently 4.1% and 1.3%) will be updated. (n.b. the simple two parameter function may also get replaced with something more complex, depending on the outcome of the above two bullets.

It was also noted that Irina is close to having some plots. She hopes to circulate them next week.

Finally, it was mentioned that the lack of documentation of the C++ classes in the reconstruction code makes it very difficult for those unfamiliar with the code and not physically at Jefferson Lab to navigate it efficiently. Dave acknowledged that this is a known issue. It is one that will always exist when the main code authors are focused on code development as a higher priority than documentation. Simon Taylor was known to have started some effort in this area a couple of weeks ago, but how far that progressed is unknown. Dave volunteered to check on it.

Action Items

  • Send e-mail to Andrei and Irina pointing them to the BCAL smearing code where the new Poisson sampling was introduced (David)
  • Measure inherent energy resolution of hdgeant/mcsmear chain without sampling fluctuations turned on. (David)
  • Write description of how we should incorporate sampling fluctuations in the smearing code (Elton)
  • Write description of how we should incorporate sampling fluctuations in the smearing code (Andrei)
  • Check and document input distributions to profile histograms to make sure tails and second peaks don't significantly affect the mean. (David)
  • Follow up on status of GlueX code documentation. (David)