Minutes, IBIS Quality Teleconference August 27, 2002 11-12 AM EST (8-9 AM PST) Bob Haller chaired the meeting. - Roll Call Adam Tambone Adam.Tambone@fairchildsemi.com Barry Katz bkatz@sisoft.com Benjamin P Silva benjamin.p.silva@intel.com *Bob Ross bob_ross@mentor.org *Eckhard Lenski Eckhard.Lenski@icn.siemens.de Eric Brock ebrock@sisoft.com *Gregory R Edlund gedlund@us.ibm.com Hazem Hegazy hazem_hegazy@mentorg.com John Figueroa jfigueroa@apple.com Katja Koller (Siemens) *Kevin Fisher kfisher@sisoft.com Kim Helliwell kimgh@apple.com *Lynne Green lgreen@cadence.com Mike Labonte milabont@cisco.com [mlabonte@hhnetwrk.com] Peter LaFlamme plaflamm@amcc.com *Robert Haller rhaller@sisoft.com *Roy Leventhal Roy_Leventhal@3com.com Sherif Hammad sherif_hammad@mentorg.com Todd Westerhoff twesterh@cisco.com Tom Dagostino tom_dagostino@mentorg.com Everyone in attendance marked by * Interested, but have not yet attended: Ralf Bruening Zuken ralf.bruening@zuken.de Yabu Akira Elpida yabu-akira@elpida.com Jun Kitano Elpida jun.kitano@us.elpida.com Kazuyoshi Shoji kshohji@hitachi-ul.co.jp Sadahiro Nonoyama nonoyama@hitachi-ul.co.jp - Review of minutes for August 13 Kevin Fisher will be at the Open Forum instead of Bob Haller. - AR Review * Parser bugs: BIRD 71 distributed, 2 more in process, one not started. BIRD 71 was assigned "Enhancement, Low Priority". We requested Bob Ross to re-assign it from enhancement to annoyance; final assignment will be made Friday at Open Forum). The BIRD's intention is to remove spurious warning messages that cause engineers to waste time reviewing curves that are monotonic for combined curves. Kevin Fisher will present BIRD 71 at the Open Forum on Friday, and=20 recommend the higher priority. Bob Ross noted that historically tables have been checked individually, since summing tables with voltage shifts etc. was considered challenging for the parser. Also, some tools require all tables to be monotonic, even before summing. In an example distributed on the ibis reflector last week, the parser reported the first non-monotonic point, which was OK after summing, but missed the second, where a "real" non-monotonic problem=20 occurred. Tradeoff between keeping all existing warnings and adding new ones, and keeping only summed tables for Pullup and Pulldown operation. Greg Edlund's BIRD on standard loads has been tentatively named BIRD 74. Roy Leventhal's BIRD ready except for test case. He will email to=20 the ibis-quality list. Bob Ross noted that testcase IBIS files should have no other=20 warnings or errors except the one reported in the BIRD. * Open: consistency checking between I-V and V-t curves. - Announcements * EIA Call for Patent Disclosure No companies made a disclosure. * IBIS Summit date (Tuesday, October 15th) First Summit announcement with Call for Papers has gone out. - Opens for new items None. - October release planning: Barry will present the checklist and roadmap. Several people will not be able to attend, due to travel budgets. Summit may have a teleconference. =20 * Documentation Need a brief description (1-2 lines) for each item on the checklist. Checklist is posted on the IBIS web site. AR: Lynne Green will collect descriptions and update the checklist. AR: Bob Haller will put together a list of sections, so people can volunteer for a section. * Roadmap for levels Brainstorming session. Too many levels would be too confusing. =20 Level 0 is pretty clear. Some stuff has already been pushed to Level 1. Maybe add things like visual checks at Level 1. Level 2 some sort of correlation to SPICE or bench? Level 0 is correct syntax, could run a simulation. Level 1 checks against datasheet or SPICE model. Level 2 verified against some bench results. IBIS Accuracy handbook describes a way to compare curves. Statistical spread and envelope comparisons might be needed, since ICs vary so much, which would be the highest level of quality. =20 Where would differential pair be checked. Are levels cumulative (probably not?). Does each model have a=20 quality value, or the entire file (maybe model-by-model, with summary at top of file?). We discussed adding Model Spec and IBIS 4.0 parameters to Level 1. Is Golden waveforms and simulation compared to golden waveforms Level 1 or Level 2? Could be Level 1 since all data is inside the IBIS file. Level 0: syntax parser checks, can be simulated. Level 1: add visual checks, contains data sheet information, pinlist is complete. EDA simulation compared to Golden Waveforms if they exist in the file. Level 2: SPICE correlation. Golden Waveforms required in model. Level 3: Bench checks. Possibly have both Spice and Bench at level 2 - Comparison of IBIS data to known good data - per the IBIS accuracy handbook. - Plans for next meeting Continue October release planning. - Schedule for next meeting September 10, 2002 11-12 AM EST (8-9 AM PST) Phone: 408-974-8478 or 800-249-5618 Meeting access: CONFERENCE SERVER 2, Conference ID: 4203 Meeting adjourned at 9 AM PST.