IBIS Open Forum 1/28/94 Minutes

From: Derrick Duehren <Derrick_Duehren@ccm.hf.intel.com>
Date: Wed Feb 02 1994 - 22:59:01 PST

Text item: Text_1

 Date: February 2, 1994

 From: Derrick Duehren (503) 696-4299, fax (503) 696-4904
          dduehren@ichips.intel.com
          Intel Program Manager
          (Will Hobb's IBIS Assistant)
          Intel Corporation
          5200 NE Elam Young Pkwy, Hillsboro, Oregon 97124 USA
          and
          Will Hobbs (503) 696-4369, fax (503) 696-4210
          Intel Modeling Manager

 Subject: Minutes from IBIS Open Forum 1/28/94

 To:
 Anacad Petra Osterman
 AnSoft Henri Maramis
 Atmel Corporation Dan Terry
 Cadence Design Sandeep Khanna, Chris Reed,
                               Pawel Chadzynski, Kumar*
 Contec Maah Sango*, Dermott Lynch,
                               Clark Cochran, Mike Venthon*
 Digital Equipment Corp. Berry Catz*
 High Design Technology Michael Smith
 HyperLynx Steve Kaufer, Kellee Crisafulli*
 IBM Jay Diepenbrock
 IBM-Motorola alliance Lynn Warner*, Soa Quoc*
 Integrity Engineering Greg Doyle, Wayne Olhoft
 Intel Corporation Stephen Peters*, Don Telian, Will Hobbs*
                               Arpad Muranyi*, Derrick Duehren*
 Interconnectix, Inc. Bob Ross*
 Intergraph Ian Dodd*, David Wiens*
 IntuSoft Charles Hymowitz
 Logic Modeling Corp. Randy Harr
 Mentor Graphics Greg Seltzer, Ravender Goyal
 Meta-Software Mei Wong
 MicroSim Arthur Wong*, Jerry Brown, Graham Bell
 National Semiconductor Syed Huq*
 North Carolina State U. Paul Franzon, Michael Steer, Steve Lipa
 Performance Signal Integrity Vivek Raghawan, Eric Bracken*
 Quad Design Jon Powell
 Quantic Labs Mike Ventham, Zhen Mu
 Siemens Nixdorf Werner Rissiek, Olaf Rethmeier*
 Texas Instruments Bob Ward*
 Thomson-CSF/SCTF Jean Lebrun
 Zeelan Technology Hiro Moriyasu*, George Opsahl*,
                               Samie Samaan, Tay Wu

 CC:
 Intel Corporation Randy Wilhelm, Jerry Budelman,
                               Intel IBIS team

 In the list above, attendees at the 1/28/94 meeting are indicated by *

 Upcoming Meetings: Date: Bridge: Res:
                     2/18/94 (415) 904-8800 661436
                     3/11/94 (415) 904-8944 661901
                     4/1/94 (415) 904-8944 661905

 All meetings are 8:00 AM to 10:00 AM PST. We try to have agendas out 7
 days before each open forum and meeting minutes out within 7 days after.
 When you call into the meeting, as always, ask for the IBIS Open Forum
 and give the bridge operator the reservation number.

 Please note: If you know of someone new who wants to join the e-mail
 reflector (ibis@vhdl.org), or have updates to your e-mail address, send
 e-mail to ibis-request@vhdl.org.

 ------------------------------------------------------------------------

 Meeting Agenda
 --------------

 In keeping with the wishes of various participants, Will arranges the
 order to keep technical discussions later.

 8:15 Check-in
 8:20 Introductions of new IBIS members/participants Hobbs
 8:25 1/7 minutes review Duehren
       Opens for new issues All
       Treasurers Report Hobbs
 8:35 Press updates Hobbs/All
       Progress toward enlisting new IC vendors All
       IBIS 2.0 Hobbs
       Forms of EDA support of IBIS Duehren
       IBIS Cookbook Hobbs
       Dues? For what? Hobbs
 8:50 BIRD 6, Differential Pin Spec. Ross
       BIRD 7, Open Specification Completion Ross
       Egg 1, mutual pin coupling Bracken
       Spice to IBIS Converter Lipa
 9:15 Canright paper Ward
       Formal BNF notation Reed, Harr
       BIRD 2, VIH, VIL Thresholds for Inputs Powell
 9:30 3D Modeling (U of NC) Bracken, et. al.
       High freq. $ EMI Goyal, et. al.
 9:45 Phased turn-on/off of multiple devices Powell
       Foils of IBIS Summit Hobbs
 9:55 Wrap-up, Next Meeting Plans Hobbs

 Minutes
 -------

 1. New members, 11/12 Minutes review, Open time for new issues:
 There were no new members to announce and no corrections were made to
 last month's minutes.

 2. Opens for new issues
 Derrick will post all the BIRDs to the forum this week. He has cleaned
 up formatting inconsistencies and problems (such as converting tabs
 to spaces). The files are named with p (new/pending), a (approved), or
 d (dead) prefixes (such as abird3_0.txt, for approved BIRD 3.0).

 Arpad Muranyi: We need agreement on when to clamp the low state of the
 pull-down curve of tri-stateable components.

 AR Arpad M.: Post your ideas to reflector.

 In regard to monotonic current tracing, one axis must be monotonic. It
 is not specified well and needs a BIRD. Discussion to be continued on
 the reflector.

 AR Kellee C.: Please be the owner for this BIRD.

 Intel has Pentium(TM) processor chip set IBIS models on vhdl.org. No
 other company directories exist.

 3. Press updates
 HyperLynx press release in Test & Measurement World, Jan. '94.

 4. Progress toward enlisting new IC vendors
 NEC and another IC vendor have shown IBIS interest.

 5. Treasurers Report
 Contractor work on the parser has been paid. We now have a $2,195.21
 balance.

 6. IBIS 2.0
 Clarification of last meeting's decision. Will suggests a full-day
 meeting to ratify 2.0, perhaps during the DAC (design automation
 conference) conference (June 6- 10? in San Diego).

 7. Forms of EDA support of IBIS
 Derrick has compiled a database of all forum members.
 AR Derrick D.: Send out existing database with instructions for
 participants to fill out/correct their information.
 2/2/94 Update: Done. Numerous responses already received.

 8. IBIS Cookbook
 Intel is producing an internal cookbook that describes how to generate
 IBIS models from simulated and actual data. It needs to be more robust.
 Intel will make non-proprietary parts available to the forum.

 Lynn Warner expressed need for basic instructions now. He was directed
 to the Overview book and Steven Peters volunteered to help.

 9. Dues? For what?
 Will asked the forum to consider what dues (if collected) would be used
 for. The following ideas were voiced:
 o Cookbook contractor
 o DAC booth
 o IBIS checker program (parser) maintenance

 10. BIRD 6, Differential Pin Spec.
 Bob Ross summarized BIRD 6. IBIS needs to support differential devices.

 Bob Ward asked for clarification on definition of the differential
 devices; signal/signal differential around a common reference. Lively
 discussion ensued with Kumar, Syed H., Steven P., and Kellee C. Bob Ross
 would like to have this referenced to the data book.

 Kumar asked how you do the V/I differentially? Answer: No, two single-
 ended V/I measurements. You would model each pin individually.

 Syed: NSC is interested in developing such models, the reference is not
 necessarily around zero.

 HyperLynx: We need some type of skew specification. Bob: min, typ or max
 or just max? Kellee: all three allowed, typ is required. If not
 specified, = 0? Is that min or typ? How about low-to-high skew versus
 high-to-low? Min will always be zero.

 AR Bob Ross: Revise Bird 6 to address skew and post for discussion.

 11. BIRD 7, Open Specification Completion
 Bob Ross, summarized and gave the background of the BIRD. Bob wants to
 expand the BIRD (spread its wings).

 The fundamental question is: should we explicitly specify all the model
 types, even though the data may make it implicit (no pull-up vi=> open
 pull-up, e.g.,). There is human readability value. Bob favors the
 latter, and expanding to include ECL and pull-up resistors, etc.

 For IBIS version 2, new model types will be required to be specified,
 whereas in 1.1, open side is inferable and will remain so.

 Kellee: The recommendation we have in version 1.1 for derivation of ramp
 rates only really works well for CMOS. It needs revision.

 Kumar proposed terminal mapping. Need to ensure backward compatibility.

 AR Bob Ross to update the BIRD to 7.1 and post for discussion.

 Terminator issue are to be left to another BIRD.

 12. Egg 1, mutual pin coupling
 Eric B. summarized the Egg.

 Sparse matrix versus banded matrix. DEC feels mutuals are very
 important. IBM has requirement for this capability in a full diagonal
 matrix (banded or sparse will not be sufficient)or else IBIS won't work
 for IBM. Mutuals account for 90% of the capacitance of the package,
 particularly with ball grid array. Capacitive and inductive parameters
 are only ones required. NSC and TI also feel this is important, too.
 Bob Ward also emphasized that the power and ground pins need to be
 included, too. On a 200 pin package, that would be a 200x200 matrix,
 40,000 entries!

 Lynn of IBM: Recommended breaking it into several triangular matrices,
 one for capacitance, one for inductances, etc.

 An include file looks like a win here, so we can change packages. Kellee
 points out we have side issues with include files: file naming
 conventions, directory conventions. And if we have a part that comes in
 multiple packages, could the IBIS file have all the pointers for each of
 the package types?

 Lynn: Are we focusing too much on file size and not enough on
 functionality?

 AR Eric: Generate Egg 1.1 for more discussion.

 13. Spice to IBIS Converter
 A Ph.D. student has a grant to write a first-level converter. We need to
 discuss this issue on reflector. Lingering question: what flavor of
 PUBLIC Spice model as source?

 14. Canright paper
 No discussion, out of time.

 15. Formal BNF notation
 No discussion, out of time.

 16. BIRD 2, VIH, VIL Thresholds for Inputs
 No discussion, out of time.

 17. 3D Modeling (U of NC)
 No discussion, out of time.

 18. High freq. $ EMI
 No discussion, out of time.

 19. Phased turn-on/off of multiple devices
 No discussion, out of time.

 20. Foils of IBIS Summit
 5 copies of the IBIS foils of the November summit are available from
 Derrick Duehren (address information above).

 21. Wrap-up, Next Meeting Plans
 We decided to start the meeting at 8:00 in the future so we'll have a
 full two hours.
Received on Wed Feb 2 22:58:14 1994

This archive was generated by hypermail 2.1.8 : Fri Jun 03 2011 - 09:52:28 PDT