IBIS MINUTES 3/7/97

From: Bob Ross <bob@icx.com>
Date: Tue Mar 11 1997 - 08:54:00 PST

 DATE: March 11, 1997

 SUBJECT: 3/07/97 EIA IBIS Open Forum Minutes
  
 VOTING MEMBERS AND 1997 PARTICIPANTS LIST:
 AMP Jeff Walden
 Applied Simulation Technology Dileep Divekar*, Norio Matsui
 Cadence Design C. Kumar*, Don Telian
 Cypress Bruce Wenniger
 Digital Equipment Corp. Jeff Chu*
 Hewlett Packard, EEsof Karl Kachigan, Henry Wu
 HyperLynx Kellee Crisafulli
 INCASES Olaf Rethmeier*
 Intel Corporation Stephen Peters*, Arpad Muranyi*,
 Interconnectix Bob Ross*
 Mitsubushi Tam Cao
 Motorola Ahmed Omer
 National Semiconductor Syed Huq*, Cheng-Yang Kao, Mike Bristol,
                                Peter Laflamme, Kevin Smith,
 NCR Dave Moxley*, Richard Mellitz
 NEC (Hiroshi Matsumoto)
 Quad Design/Viewlogic Jon Powell, Chris Rokusek*
 Quantic Laboratories (Mike Ventham)
 Texas Instruments Thomas Fisher
 Thomson-CSF/SCTF (Jean LeBrun)
 UniCAD Canada Ltd. (Celso Faia)
 Veribest Ian Dodd
 VLSI Technology Harish Patel, D.C. Sessions,
 Zuken-Redac (John Berrie)
 
 OTHER PARTICIPANTS IN 1997:
 3M Fran Hart
 Actel Scott Schlachter
 Acuson & Free Model Foundation Richard Munden
 Alcatel John Fitzpatrick
 Ansoft Eric Bogatin
 Apteq Design Systems Dan FitzPatrick
 Compaq Weston Beal
 EIA Patti Rusher*
 EMC Fabrizio Zanella
 Micron Technology Brian Johnson
 Molex Gus Panella
 North Carolina State U. (Michael Steer)
 S3, Inc. Porsh Shih, Sarathy Sribhashyam
 Ultratest International Charles Im
 Zeelan Technology George Opsahl

 In the list above, attendees at the meeting are indicated by *. Principal
 members or other active members who have not attended are in parentheses.
 Participants who no longer are in the organization are in square brackets.

 Upcoming Meetings: The bridge numbers for future IBIS teleconferences are as
 follows:
     
      Date Bridge Number Reservation # Passcode
      3/28/97 (916) 356-9200 2-133309 2774522

 All meetings are 8:00 AM to 9:55 AM Pacific Time. 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, ask for the IBIS Open Forum hosted by Will
 Hobbs and give the reservation number and passcode.
 
 NOTE: "AR" = Action Required.
 
 -------------------------------- MINUTES -------------------------------------

 INTRODUCTIONS AND MEETING QUORUM
 Jeff Chu reported that Digital Equipment Corporation is a full EIA/IBIS
 member.

 MEMBERSHIP UPDATE AND TREASURER'S REPORT
 Patti Rusher shows $10,129 in the EIA/IBIS account, with about $6,500 in
 1997 membership dues pending. She listed the members still owing payment.

 REVIEW OF MINUTES AND AR'S
 No Changes.
 

 MISCELLANY/ANNOUNCEMENTS
 Bob Ross reported that the presentations from the EIA/IBIS Summit meeting of
 January 20, 1997 are now available on vhdl.org under:

   ftp://vhdl.org/pub/ibis/summits/summit97

 Also, the draft Version 3.0 of IBIS has been updated in the /pub/ibis/wip
 directory to ver3_0f.ibs to include the most recently passed BIRDs.

 PRESS AND WEB PAGE UPDATES
 Bob Ross reported that Applied Simulation Technology's IBIS products are
 mentioned on page 69 of Printed Circuit Design, March 1997 under PCB Design
 Conference West Product Preview.

 Syed Huq reported that he has completed converting his Electronic Design
 article to HTML, and he will be mailing it on to Patti Rusher to be put on
 the EIA/IBIS home page.

 He will establish a link to the s2ibis utility available through the VeriBest
 home page. He has received permission to copy an HTML version of the IBIS
 Cookbook from Andreas Thiel of Zurich, Switzerland.

 Finally, an HTML version of the IBIS roster will be available which will
 have direct e-mail address links. Syed reported that the text version will
 still be available on vhdl.org/pub/ibis/roster for easier downloading. Bob
 and Syed will keep in sync on roster changes.

 
 NEW MODELS AVAILABLE, LIBRARY UPDATE
 [Syed Huq reports after the meeting that National Semiconductor has released
 two IBIS models which are 3.3V LVDS line drivers and receivers: ds901v031.ibs
 and ds901v032tm.ibs. These are available on the National model site.]

 Chris Rokusek reported working with Texas Instruments on a question
 concerning the test load for timing for a differential pin. Ideally, the
 load should be differential. However, IBIS supports only single-ended
 Thevenin equivalents. Bob Ross suggested a possible "Rdiff" test load might
 be preferred. However, both Bob and Chris reported good results with the
 single-ended load currently in place.

 Bob reported that Cypress Semiconductor with plans to make IBIS models
 available in the future under a new page:

   http://www.cypress.com/tech_sup/web_tech/ibis_top.html

 A new mpc2605.ibis model is now available under the Motorola Power PC section
 of the Motorola IBIS models. Also Motorola now lists direct links (although
 not working yet) to their FAST SRAM IBIS models:

   http://www.mot.com/SPS/FastSRAM/models/IBIS.html

 
 OPENS FOR NEW ISSUES
 Bob Ross on ground/power bounce model extensions (not discussed)

 DESIGN AUTOMATION CONFERENCE 1997 IBIS MEETING PLANNING
 Patti Rusher asked if the IBIS Summit meeting will be held all day on
 Thursday, June 12, 1997 and received confirmation. The deadlines for poster
 information to be sent to the graphic artist is March 21. Bob Ross will ask
 Jon Powell if he can help with this. Patti will be sending by express mail
 the necessary information. We expect IBIS literature to be distributed at
 the Conference.

 INTERNATIONAL PROGRESS
 Patti Rusher stated that Greg Ledenbach of Sematech has provided this update:
 IBIS, Version 2.1 NP (New work item Proposal) (93/46/NP) passed with 6 out of
 6 voting members and prepared to participants. For reference, four out of
 the five NPs passed.

 EIA JAPAN I/O INTERFACE MODEL PROPOSAL
 Bob Ross reported that a committee meeting is planned, Wednesday, March 19,
 1997 at the Santa Clara Convention Center during the PCB West Design
 Conference. The purpose is for Electronic Industries Association of Japan
 (EIAJ) Committee Chair Hideki Fukuda from Hitachi to present "Standard for
 I/O Interface Model", Draft Version 0.0. Several IBIS Committee participants
 have expressed interest and have been mailed a copy. There will be room for
 a few more interested IBIS participants who are attending the PCB Conference
 but please let Bob know beforehand. Current invited participants are:

   Bob Ross, Interconnectix
   Syed Huq, National Semiconductor
   Jon Powell, Quad Design/Viewlogic
   Will Hobbs, Intel
   Patti Rusher, EIA
   Stephen Peters, Intel
   Dileep Divekar, Applied Simulation Technology (USA)
   C. Kumar, Cadence
   Shinichi Maeda, KAW/USA

 Patti Rusher has scheduled the Bayshore West Room from 8 A.M. to 2 P.M. for
 about 15 people at the Westin Hotel which adjoins the Convention Center. Bob
 plans the actual meeting to start at 9 AM. Refreshments will be provided.
 Bob and Fukuda-san will be developing the schedule.

 In a note to Bob, Fukuda-san outlined his general expectations for the EIAJ
 activity:

   Draft Version 0.0:
     Get CAE Vendor support and cooperation from the EIA/IBIS Open Forum.

   Draft Version 1.0:
     Get Semiconductor Company Commitment.

   Draft Version 2.0:
     Get Customer support and model providers and processors produce systems
     to support the Specification.

   Release:
     Release hopefully by the end of the year.

 Stephen Peters asked who was behind the EIAJ activity, and Bob responded that
 a group of 17 Japanese companies (including Japanese representative from some
 American companies). Patti asked if the draft should be forwarded to IEC,
 but Bob responded that the proposal was not ready yet. There is a potential
 problem with forwarding overlapping standards. Bob indicated that he felt
 the effort had a number of distinct elements to set it apart from IBIS, and
 he expected that IBIS will continue along its current course. C. Kumar noted
 that multiple standards are expensive for the industry and also noted that
 IBIS and EIAJ activities could have some overlapping package models.

 ELECTRONIC DESIGN REPRINTS
 As previously noted Syed Huq is posting an HTML electronic copy of his article
 "Ease System Simulation With IBIS Device Models" published in the December 2,
 1996 issue of Electronic Design on the EIA/IBIS Home page. Hardcopies are
 useful for handouts at shows, particularly at DAC97. Bob Ross indicated that
 the reprints cost:

     100 $2010
     500 $2215
    1000 $2715
 
 Bob proposed that the EIA/IBIS Open Forum purchase 1000. However, Patti
 Rusher and Stephen Peters felt that 500 would be adequate.

 By unanimous consent, the Open Forum voted to purchase 500 copies.

 AR - Syed Huq contact Patti Rusher with information on purchasing the reprints.

 Syed noted that he will send back about 150 copies of the older 1995 EDN
 article to Patti to also be used at shows along with the ED article reprints.

 ROSTER UPDATE
 Syed Huq and Bob Ross are coordinating roster updates. The new roster has
 five non-EIA/IBIS member companies deleted who have not responded.

 Syed volunteered to work on the contacts.txt list to update it and to produce
 an HTML version.
 

 IBIS COOKBOOK PROGRESS
 Stephen Peters states that an internal version is nearing completion, and he
 will be reviewing it shortly. He needs to sanitize it by taking out Intel-
 specific information before releasing it. He requested Chris Rokusek to
 provide him the test load information for waveforms for inclusion.

 Bob Ross indicated that this will be valuable tutorial material in response
 to some recent reflector discussion.

 
 BIRD41.1 - MODELING SERIES SWITCHABLE DEVICES
 Bob Ross had indicated that the author of BIRD41.1, John Fitzpatrick, was on
 vacation. So Bob went over BIRD41.1 in some detail, both for explanation
 and also to raise some technical concerns. Some of the points are presented
 here.

 BIRD41.1 Consists of two additions:

   Under [Component], [Series Pin] is defined to associate two pins with a
   series model.

   Under [Model], two new Model_types (Series, and Series_switch) are
   defined. Several Series keywords are introduced to provide the electrical
   description.

 BIRD41.1 handles both discrete series elements (series diodes, resistors,
 capacitors, etc. and also series switches such as represented by the Texas
 Instruments CBT technology and by the Quality Semiconductor Quick Switch
 technology. In fact, these devices should drive what features are included
 in BIRD41.1 for series switch elements.

 Regarding the [Series Pin], under [Component], Bob had these points:

   1. Is non-symmetrical operation really needed? It will lead to problems
      in other areas.

   2. The State table syntax needs to be standardized and restricted. "OFF",
      "ON" and "NA" may be restricted suggestions. However, a more general
      grouping method may be needed to relate to "Function Table" states.

   3. The data books give "Function Table" logic descriptions of the
      components. The Series_switch is a case where IBIS should embed some
      of the logic choices in the model to reduce, in most cases, the
      combined OFF-ON permutations. This is discussed later.

 Regarding the [Model] keyword additions:

   1. Just [Rseries], [Lseries], and [Cseries] are needed. The proposed
      [R_c_series], [L_c_series] and [R_l_series] additions are unnecessary.
      All series elements should be allowed to be connected in series, just
      as the "Terminator" model elements (such as [Rpower], [Rgnd], [Gnd
      Clamp], [Power Clamp and [Rac]&[Cac]) can all be used in one [Model]
      and connected in parallel.

   2. The [Series Impedance] table can also be connected in series.

   3. The [Series Impedance] table could be defined with a voltage range
      of -Vcc to Vcc for closed switch operation or low impedance, It could
      also have a range from -2Vcc to 2Vcc for open operation. So its range
      needs to be defined. The full ranges take care of Pin1-to-Pin2 and
      Pin2-to-Pin1 non-symmetrical operation.

   4. R_load is not needed.

   5. Also the Mode subparameter could be dropped. This could be handled
      another way for actual switches. One suggestion for Series_switch
      devices might be to introduce an [On] and an [Off] set of keywords
      under which the two states are described. The simplest description
      can be done by just using low and high values of [Rseries].

 Some general comments are:

   1. The [Pin] keyword is still required as part of the [Component] pinout
      description.

   2. There is no restriction on possible [Pin] model descriptions including
      GND, POWER, NC when the pin is also listed as a series pin. This is a
      potential advantage since the [Pin] model itself could still be used
      for describing any clamping structure for protection on the correct
      pin. Drivers could be put on one side. The package description methods
      and all possible extensions under the [Pin] structure remain unchanged.

 There are some questions:

   1. What does C_comp mean in a series model?

   2. Should a Series Pin be used to connect to a supply or ground, as
      proposed above?

   3. Should the Series Pins be used for Differential terminations?

   4. Can several different series pin models be connected to the same pin?

 The final discussion consisted of elaborating on the "Function Table" for
 series switches, as described in data books. Bob illustrated the need for
 such a Function Table to allow the user interface or simulator controlled
 operation to provide only the switch combinations that the device actually
 supports (rather than all possible switch setting permutations of the
 component). Several Components were used as examples:

   1. TI SN74CBT3125 Quadruple Bus Switch has 4 independent switches which
      can be turned ON or OFF - 16 combinations.

   2. TI SN74CBT3245 Octal Bus Switch has 8 dependent switches, all OFF or
      all ON - 2 combinations.

   3. TI SN74CBT3244 Octal Bus Switch has 8 switches of 2 groups of 4
      dependent configurations where each group can be OFF or ON - 4
      combinations.

   4. Quality QS3251 QuickSwitch 8:1 Mux/Demux has 8 switches which can be
      all OFF or configured with only one switch to be ON at a time to
      connect one of 8 lines to a second pin - 9 combinations.

   5. TI SN74CBT3383 10 Bit Bus-Exchange Switch describes 5 sets of exchange
      switches which ALL either connect-directly two lines, cross-over-
      connect two lines, or provide no connections - 3 combinations.

   6. (Not discussed was a case such as the TI SN74CBT6800 with 10 switches
      for direct connection and 10 switches for precharge bias, ALL of which
      are synchronized - 2 combinations.)

 Arpad Muranyi raised the issue whether BIRD41.1 should capture the dynamic
 turn-on and turn-off characteristics of a switch. Bob suggested that there
 was sufficient complication already without this addition. Also, Bob was
 not aware of a defined need. Furthermore Arpad questioned whether the group
 switching was too much of a "logic" issue to be in IBIS. Bob stated that a
 form of group change logic was already approved in Arpad's BIRD30.2 for
 Switching Buffer Models. Arpad wondered if IBIS should extend into Analog
 and Mixed Signal domains. However, someone stated that this would enter the
 domain where HDL languages exist. C. Kumar stated that the mission of IBIS
 should be where IBIS is best positioned - a higher level behavioral modeling
 format for analog interface analysis. Bob agreed.

 Steve Peters indicated that he had worked with designs using series switches.
 Dave Moxley also indicated that he has used series switches. He stressed the
 need for a good user interface since the current way he describes series
 switches is to manually change the values of series resistors.

 Arpad questioned whether some existing table keywords should be re-used, and
 Bob responded that, in his opinion, it was better to define a new keyword if
 there was any difference in usage.

 AR - Bob Ross write up the comments regarding BIRD41.1. Bob also work with
 John Fitzpatrick concerning any changes.

 PACKAGE COMMITTEE REPORT
 Stephen Peters stated that Jon Powell did not receive all of the Connector
 Spice model and matrix information, and that he will be delayed in doing
 the planned Spice and matrix model correlation study. He expects to produce
 a proposal which includes syntax for allowing a smaller model subsection to
 describe a full connector model, if possible. The next Package Committee
 meeting is scheduled for Friday, March 14, 1997.

 BIRD36.1 - ELECTRICAL DESCRIPTION OF BOARDS
 Stephen Peters received Bob Ross' comments. He generally agrees with the
 technical comments including that a Pin can exist only once as part of a
 path description in the file.

 AR - Stephen Peters issue BIRD36.2 with the suggested changes. [Done]

 Bob proposed with Stephen's concurrence that BIRD36.2 would be scheduled for
 a vote at the next meeting.

 VERSION 3.0 FUNCTIONALITY
 Stephen Peters questioned what will be in Version 3.0, and Bob Ross briefly
 stated that the planned functionality includes all approved BIRDs and these
 pending proposals: BIRD36.2, a connector model BIRD, and possibly BIRD41.x.
 Also last-minute possibilities are early clamps and some more ground/power
 bounce details. Bob hopes that we can target the DAC97 meeting in June
 (or the meeting before) for final closure of all issues and for Version 3.0
 ratification.

 C. Kumar raised the issue of correlation of IBIS simulations with Spice
 simulations. Syed Huq indicated that aside from ApsimIBIS from Applied
 Simulation Technology, there were no utilities that could produce Spice
 models from IBIS models. (Avanti! - Meta-software has withdrawn from the
 market.) Dileep Divekar indicated that ApsimIBIS produces Spice models
 compatible ApsimSpice only. Bob Ross deferred the correlation discussion
 and IBIS to Spice tool generation issue to the next meeting.

 MODELING EARLY CLAMPS IN IBIS
 Arpad Muranyi reported that at this time he has focused his initial research
 (with assistance from C. Kumar) on describing appropriate measurements for
 electrical description. The formatting details are not the difficult issue.
 He is planning a BIRD proposal for Version 3.0.

 PARSER STATUS
 Chris Rokusek plans to send his ibischk2 extensions to resolve BUGs 5-7 to
 Bob Ross for posting on vhdl.org. Only the executables will be posted since
 the source code is available only through purchase of a license. After some
 testing by the users, Chris may provide the enhanced code as the new base-
 line for future ibischk3 development for Version 3.0 of IBIS.
 
 
 BUG8 - RESERVED WORDS USED AS PINNUMBERS GENERATE ERRORS
 Bob Ross indicated that BUG8 has been in existence since the first ibis_chk
 parser. The IBIS specification could be interpreted to mean that this is
 not a bug. Bob did not recommend changing the parser at this time, but
 rather to classify it as an annoyance, with no fix. There are easy work-
 arounds. Stephen Peters proposed that BUG8 be fixed as part of the Version
 3.0 ibischk3 development project.
 
 After some brief discussion, everyone concurred by unanimous vote that BUG8
 will be fixed in ibischk3.

 Bob noted that the current release of s2ibis2_ver1.1 contains examples which
 has "gnd" as the IBIS file pin number, and this will generate an ibischk2
 error message.

 NEXT MEETING:
 The next meeting is on Friday, March 28, 1996, 8:00 A.M. to 9:55 A.M.
 BIRD36.2 is scheduled for a vote.
 ==============================================================================
                                       NOTES
 
 IBIS CHAIR: Bob Ross (503) 603-2523, Fax (503) 639-3469
             bob@icx.com
             Modeling Engineer, Interconnectix
             10220 SW Nimbus Ave, K4, Portland, OR 97223

 VICE CHAIR: Syed Huq (408) 721-4874, Fax: (408) 721-4785
             huq@rockie.nsc.com
             Staff Applications Engineer, National Semiconductor, M/S A-2595
             2900 Semiconductor Drive, Santa Clara, CA 95052
 
 LIBRARIAN: Jon Powell (805) 988-8250, Fax: (805) 988-8259
             jonp@qdt.com
             Transmission-Line Products Manager, Quad Design/Viewlogic
             1385 Del Norte Rd., Camarillo, CA 93010
 
 SECRETARY: Vacant
 

 This meeting was conducted in accordance with the EIA Legal Guides and EIA
 Manual of Organization and Procedure.
 
 The following e-mail addresses are used:

   ibis-request@vhdl.org
       To join, change, or drop from either the IBIS Open Forum Reflector
       (ibis@vhdl.org), the IBIS Users' Group Reflector (ibis-users@vhdl.org)
       or both. State your request.

   ibis-info@vhdl.org
       To obtain general information about IBIS, to ask specific questions
       for individual response, and to inquire about joining the EIA-IBIS
       Open Forum as a full Member.

   ibis@vhdl.org
       To send a message to the general IBIS Open Forum Reflector. This
       is used mostly for IBIS Standardization business and future IBIS
       technical enhancements.

   ibis-users@vhdl.org
       To send a message to the IBIS Users' Group Reflector. This is
       used mostly for IBIS clarification, current modeling issues, and
       general user concerns.

   ibischk-bug@vhdl.org
       To report ibischk2 parser bugs. The Bug Report Form Resides on
       vhdl.org in /pub/ibis/bugs/bugform.txt along with reported bugs.

 Information on IBIS technical contents, IBIS participants, and actual
 IBIS models are available on the IBIS Home page found by selecting the
 Electronic Information Group under:

   http://www.eia.org

 Check the pub/ibis directory on vhdl.org for more information on previous
 discussions and results. You can get on via ftp anonymous, or send an e-mail
 request to the automatic archive server, archive@vhdl.org.
 
 "IBIS Spoken Here" placards are available from Jon Powell (jonp@qdt.com) for
 use at trade shows.
 ==============================================================================

 
Received on Tue Mar 11 08:54:46 1997

This archive was generated by hypermail 2.1.8 : Fri Jun 03 2011 - 09:52:29 PDT