Attendence: joel@force10networks.com tomaz@force10networks.com GOleynick@fciconnect.com m_oltmanns@comcast.net john.dambrosia@tycoelectronics.com tpalkert@visi.com Steve.Anderson@xilinx.com Rob.James@idt.com popescu@mail.quaketech.com p-hanish@ti.com why@pmc-sierra.com william.r.peters@intel.com Doug.Day@taec.toshiba.com ahealey@agere.com charles_moore@agilent.com cathyl@lsil.com J.Mitchell@winchesterelectronics.com glen@vitesse.com brian.seemann@xilinx.com 22 april 2004 Thursday 10am PDT Conference call minutes - Peter Tomaszewski to record minutes. Meeting Agenda - Additions/corrections to minutes from last meeting due by Monday 04/26/2004. - Update on investigations - 2 to 4 port conversion - Graeme Boyd will put something together for next call - 5/5/2004 - We will go over 4 port to differential conversions - Materials - fr4 definition update - ANSI and UL have different definitions - adjust presentation to show current definitions and direction they are each taking. - We will stay with current Enhanced FR4 definition we are using until more information is available. Joel hopes to have a recommendation by 5/5/2004. Note - it was mentioned that the IEEE must look to international standards for definition of FR4 not local standards. - Update on calibration board scheme - Tyco and TTMT to work with Joel and John D'Ambrosia to create a round robin set of cards to be available to all to take measurements on and correlate. - Reiterate - We can only recommend not decide. Meeting - Test points - TP1 and TP4 - 2 directions - NEXT/FEXT/etc. or 4 port S-parameter descriptions - Joel would like NEXT/FEXT enhanced with 4port mag/phase as a starting direction. - Discussion of SNR/Stat eye/equation based models to define channel and avoid false positives/negatives - Missed group delay/NEXT/FEXT on XAUI? Specified TX/RX but not equalization. John D'Ambrosia - results are inversely proportional to length (UNH) - Joel OK with TX/RX/equalization - normative. Informative - show worst cases. - Stat Eye - discomfort with trusting it - Straw poll - Is there anyone in disagreement that we recommend to the group to use an SDD21 equation as an informative text for the worst case definition of a channel? Vote no if you disagree with this statement - We recommend to the group to use an equation based SDD21 as an informative - not taken at this time. - Channel should be code agnostic. Define channel then have the group go out and decide whether or not a TX/RX can be developed and a code scheme developed that can run on it and whether or not the channel needs to be redefined. - Launch and receive terminations would be included in the TX and RX definitions. - Straw poll - Pg. 3 of notes - Does anyone disagree with updated diagram TP1/TP4 as an informative point for us to move forward with? - No detractors. The TP1/TP4 points stand. - Terminations - The equation set will have to adjust as an informative to the TX/RX and code set. Assume that the terminations are ideal for now. - Straw Poll - Vote no if you disagree with this statement - We recommend to the group to use an equation based SDD21 as an informative point. No detractors. - Discussion of leaning towards a comprehensive model. Normative and completely defined on an informative. Table of margins that identify the channels' ISI, length, Dj, Rj, skin effect, impedance mismatch, etc. Mix/match TX/RX as needed to get it to work. Decide on a set of equations described in an informative text. May get included in a script to reference. Individual persons/companies could devise their own scripts as they deem necessary. History of IEEE shows inclusion of informative texts/scripts to clarify/aid users. Equation/equation set may be very complex, but there is consensus that it is the right direction. Request for detractors to speak forward - no detractors. - We will have another conf call Friday, April 30, 2004, 10AM PDT - Conf call with Brian/Graeme/Joel on Monday, April 26, 2004, 8AM PDT. Cover table of margins and defining the equations for each entry. All are welcome to listen (and participate as long as the meeting keeps moving forward). - Agreed to recommend 100MHZ to 20GHZ frequency range for s-parameter analysis. - Hold off on NEXT/FEXT until next call. 5/5/2004 - S-parameters - 2-port to 4-port conversion and on to differential s-parameters. Joel asks that if you disagree or want to add to the equations set on Pg. 16, please do so in an email to Joel before the meeting (call) on 5/5/2004. - Straw Poll - Do we need informative points on transmission line theory and measurement techniques? Reference other IEEE documents? Poll was placed on hold until further information can be obtained. Recap - 4/26/2004 conference call to define the table of margins. - 4/30/2004 conference call to discuss results of 4/26/2004 - 5/5/2004 conference call - next group call. - Return loss - Place text in justification column stating that these are numbers that have not been discussed or agreed upon by the group. Return loss numbers will be removed from the table. The following was never officially completed in the meeting: - Straw Poll - Does anyone disagree with: We recommend to the group that the transmitter and receiver will have the appropriate differential terminations?