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

Re: [10GMMF] Notes for TP2 call 10/28

Hi Martin - we missed you on the call. We assumed/hoped you would participate in this work.
We didn't get into your question too much in the call. I recommended that anyone interested should consider a range of transmitters (drivers, coupling circuits, lasers, etc.) that might reflect their own objectives of performance & margin and yield & cost, even while perhaps not fully knowing the relationship to overall LRM penalties. Each company may have a different view. Everyone knows that faster and more linear transmitters will produce lower penalties and the budget is tight, so I'm expecting reasonableness to play. I am reluctant to narrow it down beyond this at this point.
This raised a question about willingness to show data. Certainly ClariPhy, upon advisement to do so, would keep any results of its work anonymous. I'm hoping that some at least some waveforms can be shared to be able to simulation corroborate results.
As far as the budget and how much it allows, that is still ahead of us. The idea is to gather data on what is and what folks want, and trade it off vs. all the other items in the budget (such as 220 vs. 300 meters, % coverage, Rx sensitivity, Tx range, etc.). We're trying not to pre-suppose the outcome at this point. We need to get the data first and then determine what it tells us.
All this being said, using the present mask as a reference point is sensible, but understand that when scaled per OFSTP-4A, the mask can allow a lot of variation well beyond the simple concept of rise and fall times. For some linear examples, see from John Ewen.
----- Original Message -----
Sent: Friday, October 29, 2004 2:48 AM
Subject: Re: [10GMMF] Notes for TP2 call 10/28




We plan to run simulations using various Tx realistic characteristics (rate equations+ Gaussian etc) for support of the objective as listed below.  The question is what are ‘realistic’ characteristics?  A starting point could be that they barely pass the eye mask (this leads to ~47 ps rise/fall time). Typical 10G Txs have, however, lower rise-fall time than the ~47ps since a large margin is desired. One could argue that a Tx/Rx with ~47ps rise/fall time is already a ‘relaxed’ transmitter (which the budget may not allow for).  


Any discussion/direction on what ‘realistic’ characteristics are in order to focus the simulations?


Any plans on distributing the various Tx waveforms that have been generated/captured by Clarify and others?




From: owner-stds-802-3-10gmmf@IEEE.ORG [mailto:owner-stds-802-3-10gmmf@IEEE.ORG] On Behalf Of Tom Lindsay
Sent: 29. oktober 2004 01:16
Subject: [10GMMF] Notes for TP2 call 10/28


All - here are my notes for today's (10/28) call. Feedback would be appreciated.



Greg Lecheminant, Agilent T&M

Pavel Zivny, Tektronix

Tom Lindsay, ClariPhy

Andre Van Schyndel, Bookham

Piers Dawe, Agilent

Nick Weiner, Phyworks

John Ewen, JDSU

Tom Whitehead, Emcore

Others? Please let me know.



Approved, per below.



  • Andre invited feedback on the list of laser non-idealities sent to the reflector this week.
  • Spent most of the time discussing the work plan/outline I sent to the reflector 10/26 for testing numerous waveforms for penalties, etc.
    • Nick said that work should also look for direct relationships between TP2 properties and penalties, such as rise/fall times, overshoot, asymmetries, etc. Piers added that eye diagrams should be captured for the same purposes.
    • Discussed if relationship between Gaussian pulse shape and more realistic waveshapes had been established - no. Tom's thoughts are that a relationship could be established based on an agreed reference Rx, but this has not been established. Piers said that it is important to know if ranking (fiber vs. penalty) retained the same order over the Cambridge set with various waveforms and EDC architectures (I probably did not record this right - Piers?).
    • It will be more clear what E/O waveshape should be used for TP3 testing after these tests/sims are completed.
    • General agreement that 4 samples per bit was not adequate, that 8-16 was the right range.
    • General agreement that PRBS10 is better than PRBS7 for this work, due to Tx/channel/Rx memories, but some may be limited by capture record length (8 samples per UI x 1023 UI = 8184 points, but some scopes are limited to ~4000).
    • Do both 220 and 300 meters sims.
    • Run multiple EDC configurations to learn more about penalties and sensitivities.
  • Nov presentations
    • Work required is clear. Objective is to do as much of the plan as possible and present any conclusions in Nov. TBD.
    • I also plan on presenting a modified version of TP2/TP3/budget slides discussed previously on the calls.
    • Others should begin thinking about TP2 presentations.

Next call

  • Date: Thurs, 11/06/04 (regular day/time)
  • Time: 9:00 AM Pacific
  • Duration: 1:30 max
  • Number: 401-694-1515
  • Access code: 421721#


Tom Lindsay
ClariPhy Communications
phone: (425) 775-7013
cell: (206) 790-3240


  • Proposed agenda (for 10/28)
    • Attendance
    • Agenda
    • Review previous notes (above)
    • Technical discussion
      • Outline for TP2 waveform analysis (sent 10/26 8:57 PM)
        • Obtaining waveforms
        • EDC model to use?
      • "Graphical view" (mask) progress?
      • More on budget and metrics?
    • Plan next call

Any comments or changes, please let me know.