Minutes, IBIS Quality Teleconference
August 26, 2003


TIME AND PHONE FOR NEXT MEETING
September 9, 2003
9-10 AM EST (6-7 AM PST)
Phone: 408-974-8478 or 800-249-5618
Meeting access: CONFERENCE SERVER 1, Conference ID: 54644

ROLL CALL
Adam Tambone
Barry Katz, SiSsoft
Benjamin Silva, Intel
*Bob Ross, Teraspeed
Brian Arsenault, EMC
*Eckhard Lenski, Siemens
Eric Brock, SiSoft
Gregory Edlund, IBM
Hazem Hegazy, Mentor Graphics
John Figueroa, Apple
*John Angulo, Mentor Graphics
Katja Koller, Siemens
Kevin Fisher, SiSoft
*Kim Helliwell, Apple
Lance Wang, Cadence
*Lynne Green, Cadence
*Mike Labonte Cisco
Peter LaFlamme, AMCC
*Robert Haller, SiSoft
Roy Leventhal, 3com
Sherif Hammad, Mentor Graphics
Todd Westerhoff, Cisco
Tom Dagostino, Teraspeed
*Kazuyoshi Shoji, Hitachi
*Sadahiro Nonoyama, Hitachi

Everyone in attendance marked by  *

Interested, but have not yet attended:
    Ralf Bruening, Zuken

NOTE: "AR" =3D Action Required.

-----------------------MINUTES ---------------------------

- Review of minutes for August 12, 2003
   Minutes approved as written.


- Announcements
   EIA Call for Patent Disclosure.  No company announced a patent.

   IBIS Summit at PCB East.  First Call for Papers has been
sent out.  Barry and Bob plan to present IBIS Quality paper.


- AR Review
   Files VT_curves.txt renamed to VT_tables.txt.


- Opens for new items
   Email addresses will be removed from all future minutes.

   IBIS Model Review Committee.  Add Bob Haller.  Some reviewers
are already using the draft IBIS Quality checklist.

   Documentation plan is to have spreadsheet, items linked to
each detailed document (HTML and/or PDF format).


- Detailed Documentation review, VT_tables.txt
   Item 2: checking I-V against V-t curves.  Was reworded
to clarify how tables are summed prior to checking the
load line.  Syntax needs to be consistent with IBIS use
(V_fixture etc need underbar).  This is a Level 0 check.

   Item 5: Replace 50 Ohms with Rload.

   Item 8: dV typically largest at Max corner,
dt typically smallest at Max corner.  Model should be
documented if model does not follow this.

   Break Item 9 into 2 items?  One on V-t points,
the other on time coverage.

   Visual inspection: for same termination, final value
of the falling curve matches the starting value of the
rising curve and vice-versa.   R_fixture should match
the expected board load.  (R_fixture is not required to
be the same as Rload.)
  =20
   Recommend at least 10 points in transition region.
No issue if extra points are left in the flat regions.
A linear slope still needs points, since it is being
used with a non-linear I-V curves.

   V-t tables should cover 1/2 of fastest expected
clock period.  All of V-t curves must reach steady-state
by the end of the V-t tables.  Note that different
simulators handle different leading and trailing flat
zones differently when the buffer is toggled faster
than the length of the V-t table (over-clocking).

   IBIS does not deal well with buffers that do not
reach steady-state.  Arpad Muranyi presented an
algorithm for handling over clocking, at an IBIS summit
in the past year.


- Plans for next meeting
   Finish V-t table review.
   Begin review of Eckhard's document on error checks.


Next meeting:
September 9, 2003
9-10 AM EST (6-7 AM PST)
Phone: 408-974-8478 or 800-249-5618
Meeting access: CONFERENCE SERVER 1, Conference ID: 54644



Meeting ended at 10:00 AM Eastern Time.


Links:
IBIS Home Page, http://www.eigroup.org/ibis/ibis.htm
IBIS Quality Checklist, http://www.sisoft.com/ibis-quality/checklist/
IBIS Quality Minutes, http://www.sisoft.com/ibis-quality/minutes/
Documentation drafts, http://www.sisoft.com/ibis-quality/docs/
---------------------------------------------------------------------
IBIS Quality website:  http://www.sisoft.com/ibis-quality
IBIS Quality archives: http://www.freelists.org/archives/ibis-quality
To unsubscribe send an email:
  To: ibis-quality-request@freelists.org
  Subject: unsubscribe

