From: Mirmak, Michael
Sent: Monday, January 29, 2007 18:19
To: 'Bob Ross'; 'Fabrizio.Zanella@cst.com'; 'john_angulo@mentor.com';
'Lynne D. Green'; 'samchitwood@sigrity.com';
'sanjeev_gupta@agilent.com'; 'rrwolff'
Subject: IBIS Interconnect Ad Hoc Meeting Jan. 24 Minutes and Feb. 7
Agenda

======================================================================
IBIS INTERCONNECT MODELING AD HOC TASK GROUP MEETING MINUTES
Meeting Date: January 24, 2007

http://www.eda.org/ibis/adhoc/interconnect/

Attendees:
----------
(* denotes present)
Agilent                    - John Moore*
Intel                      - Michael Mirmak*
Mentor Graphics Corp.      - John Angulo*
Micron Technology          - Randy Wolff*
Sigrity                    - Sam Chitwood* 
Teraspeed Consulting Group - Bob Ross* 

======================================================================

Next Meeting:
Wednesday, February 7, 2007

               Telephone     Bridge   Passcode
              916-356-2663      2     745-3142

LiveMeeting Details:

  Subject: IBIS Ad Hoc Interconnect Meeting
  Meeting URL: https://www.livemeeting.com/cc/intel/join
  Meeting ID: 93K26W
  Meeting Key: 8jC<qh7Km


Agenda:
- Final review of IIRD9
- Review of updated Touchstone(R) draft 

========================================================================

The meeting opened with a summary of the Agilent Touchstone(R) reference
proposal by John Moore.  He suggested, with agreement from Sam Chitwood,
that supporting complex reference impedances would lead to either
ambiguous interpretation by tools or resistance from tool vendors to
changing established algorithms.  Additionally, the references themselves
would need to be frequency-dependent.  The team agreed that real-valued
reference impedances would be suitable, as calibration in the lab handles
adjustment for frequency-dependence, and mathematical manipulations are
straightforward.

In terms of changes, the team agreed that an advanced Touchstone(R) format
should support:
- unlimited ports
- real-valued per-port reference impedance values
- an unambiguous version control marker, to prevent evaluation of the data
  using "old" Touchstone(R) rules.

John Moore noted that CITI also has a problem with supporting multiple
reference impedances.  he suggested that the header line of Touchstone(R)
not be used for version or reference impedance data, as the header is 
already overloaded and may cause parsing issues.

For making reference nodes explicit, the team also agreed that ICM is
the best place for this data; Touchstone(R) should not include it.

The team reviewed Michael's proposed IIRD9.  In general, the response was
positive.  Sam wanted to ensure that (a) floating or unconnected nodes
were acceptable and that (b) nodes could be grouped to correspond to
a single port.  For example, a 1000 pin device may only have four
unique rails or signal paths -- Vcc, Vss, signal1 and signal2.  In this
case, large numbers of physical nodes would have to use the same name,
and connect to a single port.  Michael suggested that the current 
syntax proposal supports this.

Sam also raised the issue of reference nodes that are local, not shared
across all measurement nodes and not physically connected to pins.  This
"virtual reference" case may not be handled by IIRD9.  Sam took the AR
to draw a diagram of this case.

Michael will release an updated IIRD9 to address the concerns of the team.


ARs
-----
Sam     - provide a simple drawing of "virtual reference" case
Michael - update IIRD9 draft with changes for final review next time


Team Objectives:
1) complete ICM-IBIS linking BIRD and any associated changes to the 
   ICM specification
2) update the ICM specification, if needed, to clarify the mapping of 
   ICM nodes to S-parameter ports
3) complete a specification for "Touchstone Plus" or similar 
   industry-standard definition for Touchstone-like files, to include 
   complex impedance references, removal of limits on the maximum number 
   of ports and per-port impedance references