Minutes, IBIS Quality Committee 08 September 2009 11-12 AM EST (8-9 AM PST) ROLL CALL Adam Tambone * Anders Ekholm, Ericsson Barry Katz, SiSoft Benny Lazer Benjamin P Silva Bob Cox, Micron * Bob Ross, Teraspeed Consulting Group Brian Arsenault David Banas, Xilinx * Eckhard Lenski, Nokia Siemens Networks Eric Brock Guan Tao, Huawei Technologies Gregory R Edlund Hazem Hegazy Huang Chunxing, Huawei Technologies John Figueroa John Angulo, Mentor Graphics Katja Koller, Nokia Siemens Networks Kevin Fisher Kim Helliwell, LSI Logic * Lance Wang, IOMethodology * Lynne Green, Green Streak Programs * Mike LaBonte, Cisco Systems Mike Mayer, SiSoft * Moshiul Haque, Micron Technology Muniswarareddy Vorugu, ARM Ltd Pavani Jella, TI Peter LaFlamme Randy Wolff, Micron Technology Radovan Vuletic, Qimonda Robert Haller, Enterasys Roy Leventhal, Leventhal Design & Communications Sherif Hammad, Mentor Graphics Todd Westerhoff, SiSoft Tom Dagostino, Teraspeed Consulting Group Kazuyoshi Shoji, Hitachi Sadahiro Nonoyama Liqun, Huawei Everyone in attendance marked by * NOTE: "AR" = Action Required. -----------------------MINUTES --------------------------- Mike LaBonte conducted the meeting. Call for opens and patent disclosures: - No one declared a patent. - Lynne would like to discuss her feedback, emailed to Mike AR Review: - Mike try to reduce TOC to 2 pages - Dropping from 12 pt to 10 pt did not do it - We will keep the 3 page format - Mike link to new IQ 2.0 document on our web page - Done - Mike asked what date the Open Forum might vote to accept - Bob: Oct 30 will be OK - Mike change web site to avoid using frames - Ongoing New items: Mike showed Lynne's file Touchstone_quality.txt: - Lynne: These are suggestions I am throwing into the hat - Mike: The TSCHK checker is in development now - Lynne: One of the suggested checks is causality, which TSCHK will not check - Bob: Most EDA tools do this check - It would be nice to have it in TSCHK though - Passivity is not always important - It might be a warning check, not error - Mike: Maybe someone would contribute code - Lynne: It's fairly easy to write - Mike: One of the suggestions is about noise limits - Why is this a problem? - Lynne: Converting matrices creates a lot of magnitudes close to zero - Mike: Is this avoidable? - Lynne: More digits would help - The fuzz can cause passivity failures - Bob: Not sure what tools do about this - Mike: It would be a problem if TSCHK passed a file and an EDA tool failed it - Lynne: It can be a manual check too - Bob: We should check for reciprocity, where S12 must equal S21 - Lynne: That is not required - Bob: Passive networks should inherently be reciprocal - Lynne: That would be an easy TSCHK check anyway - Bob: It may fail due to noise - There may be helpful methods using Eigenvalues or averaging - Lynne: It is best to have people actually looking at the data - Lynne's email: "Port mapping to physical comopnent or structure" is a datasheet check - Bob: Walter (Katz) is working on Electrical Module Description, which is related - Mike: What is the port mapping issue? - Lynne: There are no port names in a Touchstone file - Some automatic tools put matrix elements in random order - Lynne's email: "Special checks for 2-port, mixed mode, and differential" - The file has to say mixed-mode or differential - Bob: TS2 has requirements for these - Lynne's email: "Simulation with AC source and resistive terminations" - Lynne: Hot button is DC operating point - "Overall correctness in frequency domain" looks for resonances - "Correct step response and ramp response in time domain" - Mike: The end user can't make these checks - Lynne: They can see if it simulates - Lynne's email: "Repeat with a reactive load?" - Lynne: This is a diagnostic if the others don't pass - Lynne's email: "Comparison of both frequency and time domain simulation to test data" - Lynne: This is against bench test data - Lynne: Should test data be included, like Golden waveforms? - Bob: That might belong in a separate file - Mike: There has already been concern about the potential number and size Mike showed Lynne's file Multilingual.txt - Lynne's email: "IBIS file call calls model file in same directory" - Lynne: There should be no slashes - Lynne's email: "IBIS language call matches model file's language" - Lynne: This might have to be a manual check - Lynne's email: "Parameters passed to model file are accepted by (used in) model file" - Mike: The opposite check might be good too - Bob: We currently don't look inside the external language files - Lynne's email: "Parsers available for SPICE, HSPICE, *-AMS?" - Mike: Do people make IBIS files declaring HSPICE? - Lynne: Yes - Mike: That is non-compliant - Bob: Berkeley SPICE is not truly useful here - Each company supports it to the extent they can - Lynne: No company has implemented the entire AMS spec - Bob: Multilingual did not really take off - They took it as a starting point and added their own parameters - The files will fail IBISCHK - Lynne: The SPICE files should be checked somehow - Lynne's email: "Port/Interconnect mapping to physical comopnent or structure" - Mike: Same as with TSCHK - Lynne's email: "AMI information used, if provided" - Bob: This is completely separate from Multilingual - AMI is having it's own troubles right now - Lynne's email: "Parameters passed with correct names and resonable values" - Mike: This is a datasheet check - Lynne: Tools might recognize the wrong language spec and still do the right thing - Lynne's email: "Differential checks (such as, polarity same for [model], [component] and [diff pin])" - Mike: Does IBIS Polarity matter? - It seems no EDA tool needs to use it - Lynne: Polarity has to be right for diff pins - It matters for external models - Lynne's email: "Simulation with any driver and/or resistive load, as appropriate" - Lynne: Trying to get a DC operating point can reveal many problems - Also correct response in time domain, if active - Correct response in frequency domain, if passive - Lynne: We can at least do a sanity check - Lynne's email: "Extend comparison if model includes additional effects" - For example, thermal effects - Lynne: This is just to get the conversation started - Mike: We have to decide what to do next - One idea is to update the Accuracy Handbook - Unfortunately Roy and David are no longer able to help - We could also file parser bugs to implement IQ checks Moshiul: Should we update our checklist spreadsheet? - Mike: We can look at that in the next meeting Moshiul: Are there any order requirements for M, S, X, G? - Mike: None Meeting ended at 12:08 PM Eastern Time.