Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

[10GMMF] Notes for TP2 call 1/13/05

Attendees
  • Nick Weiner
  • Jan Peeters Weem
  • Piers Dawe
  • Greg Lecheminant
  • Tom Lindsay
  • Norm Swenson
  • Lew Aronson
  • Ali Ghiasi
  • Abhijit Shanbhag
  • Others who did not ID?
Agenda discussed and approved (see below)
Previous minutes/notes from 1/6 approved
 
Technical progress
  • Objective - resolve as many details as possible by/before the Vancouver meeting Jan '05.
  • TWDP simulation tool
    • Status
      • Norm has sent MATLAB code for review to those who requested it. If others are interested in reviewing the code, contact Norm Swenson. Discussion of the code will be an agenda item for 1/20.
      • Jan ran the code with different waveforms (and now Piers with simulated noise), so it appears to be working for others.
      • Tom submitted comments to adopt the code into the next draft, when ready, and to include the normative information Norm discussed previously.
    • Presentation - see http://grouper.ieee.org/groups/802/3/aq/public/upload/ResultsandcomparisonsforTWDP_MATLABcode.pdf by Tom.
      • Comparison to PIE-D
        • Preliminary comparison to John Ewen's PIE-D results across 108 fibers showed agreement to within ~0.04 dB.
      • Filter configurations
        • Previous discussions questioned whether a Bessel-Thomson filter would be a better approach for the anti-aliasing filter for the noise, with no additional filtering for the signal.
        • ClariPhy ran the code in both modes on a couple of laser waveforms and determined that this filter configuration gives slightly better results (~0.15 dB) than PIE-D. This can be explained by the fact that PIE-D assumes white noise, whereas the new approach filters (colors, reduces) the noise, making it better than PIE-D. In the other filter configuration, the code will treat noise as white with the Butterworth anti-aliasing filter being simply part of the matched filter.
        • Tom read an email from Lew saying that their results generally show TIA noise spectra as similar to the signal frequency response.
        • This all led to a discussion of whether the definition of PIE-D should change, since perhaps real receivers may benefit from noise filtering.for the signal. My opinion is that we should stick with the existing definition of PIE-D due to it's duration and extent of usage, and leave any benefit beyond that a receiver implementation choice.
      • Reference results
        • Reference eye patterns and reference penalty CDFs are provided for comparison to previous MATLAB code results.
    • ACTIONS - new
      • ClariPhy to test sensitivity of filter configurations with 47.1 psec Gaussian pulse shape
      • Update a per fiber comparison with PIE-D for both filter configurations using 47.1 psec waveform
  • Pattern definition
    • Little progress since last call. Actions can be found in previous meeting notes.
  • Waveforms & penalty plots
    • Not discussed, but more waveforms from more companies are desired to scope the pass/fail requirements for TP2.
  • D1.0 comments
    • Nick asked if we could pre-resolve some comments relating to TP2.
    • The comment list is in the agenda. All comments were discussed.
    • A separate email will attempt to capture progress.
 
Next call
  • Date: Thurs, 1/20/05 (regular day/time)
  • Time: 9:00 AM Pacific
  • Duration: 1:30 max
  • Number: 401-694-1515
  • Access code: 421721#
 
----- Original Message -----
Sent: Wednesday, January 12, 2005 7:18 PM
Subject: [10GMMF] Reminder for TP2 call 1/13/05

All -
 
Next call
  • Date: Thurs, 1/13/05 (regular day/time)
  • Time: 9:00 AM Pacific
  • Duration: 1:30 max
  • Number: 401-694-1515
  • Access code: 421721#
 
Agenda
  • Attendance
  • Agenda
  • Approve old minutes
  • Progress on tasks and distribution of work (see within previous minutes, below)
    • Progress on TP2 test methodology and code development
      • Document & code status
      • Comparisons to PIE-D
      • Filter configurations - results from ClariPhy and/or Lew?
    • Test patterns - any progress?
    • Nick has asked if we can pre-resolve TP2 comments
      • Easy? - 49,50,81,83,104,105,106,109,111
      • Less easy? - 107,108,110
    • Other?
  • Plans for next call (1/20/05)
Tom