IBIS Macromodel Task Group Meeting date: 10 May 2016 Members (asterisk for those attending): ANSYS: * Dan Dvorscak Curtis Clark Broadcom (Avago): Xingdong Dai Bob Miller Cadence Design Systems: * Ambrish Varma Brad Brim Kumar Keshavan Ken Willis Cisco: Seungyong (Brian) Baek eASIC: David Banas Marc Kowalski Ericsson: Anders Ekholm GlobalFoundries: Steve Parker Intel: * Michael Mirmak Keysight Technologies: Fangyi Rao * Radek Biernacki * Ming Yan Maxim Integrated Products: Hassan Rafat Mentor Graphics: John Angulo * Arpad Muranyi Micron Technology: * Randy Wolff Justin Butterfield QLogic Corp.: James Zhou Andy Joy SiSoft: Walter Katz Todd Westerhoff * Mike LaBonte Synopsys: Rita Horner Teraspeed Consulting Group: Scott McMorrow Teraspeed Labs: * Bob Ross TI: Alfred Chong The meeting was led by Arpad Muranyi. -------------------------------------------------------------------------------- Opens: - None. ------------- Review of ARs: - Ambrish to check for a collaborator's feedback on his nearly ready new version of the Backchannel proposal. - This is in progress. - Ambrish will be unable to get to this for a few weeks. - Ambrish will send the latest BIRD 147 version to Bob and Walter, as requested [AR]. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: Radek motioned to approve. Dan seconded. No one objected. ------------- New Discussion: [Pin Reference] BIRD draft: - Michael M: Will this BIRD be addressed and a recommendation made by this group? - Arpad: That has not been discussed, but it probably will be. - Radek: It might not be a separate BIRD, it could be included in the ground cleanup BIRD. - Bob: This should be a separate BIRD. - Arpad: It would have a dependency and would have to be approved at the same time as the other BIRD. - Mike L: We would vote on all BIRDs considered for 6.2 together. - The edited 6.2 docx file can pre-incorporate the anticipated BIRDs. - Radek: [Pin Reference] was an alternative to extending [Pin Mapping]. - That would have been another BIRD. - Michael M: [Pin Reference] should not be a large editorial effort. - Any changes should be easily accommodated. - Arpad: We should try make some progress on it today. - Mike L showed a rev 1.3 of the BIRD draft. - Bob: We may be crossing paths, I have my own edited version 1.3. - There is a note paragraph in Mike L's version that should not be there. - Mike L: An example has IBIS syntax commented out. - We should not show commented out keywords in the IBIS specification. - Radek: That could be omitted and the result would effectively be the same. - Bob showed his rev 1.3 BIRD draft. - Bob: This has to assume [Pin Mapping], and it is stated as such here. - For RS232 there may be no 0V reference voltage available. - Mike L: It needs to identify a reference terminal, not a voltage. - Does a bus_label map to a single node? - Bob: It may be several nodes that map to one bus_label. - Mike L: The voltages such as Vinl/Vinh are specified for a [Model], but we are giving the reference terminal at the component level? - Bob: For PECL the threshold shifts up and down relative to 5V or 0V rails, for example. - Different voltages can have the dominant effect on this shift. - Arpad: Should there be a different reference for voltages that are divided down to produce a threshold reference? - We should do this right, make sure it covers all possibilities. - Radek: It could be an [External Reference]. - Bob: This BIRD as written is intended to shift between the chosen voltage and the embedded reference. - Arpad: We used to use 0.8 and 2.0 for TTL, but actual values could vary. - Bob: Those voltages were based on having one diode tied to ground. - It was very insensitive to VCC changes. - Arpad: The question is if Vinl/Vinh are specification based or behavior based. - Bob: They are specified levels. - Arpad: If 30% and 70% of actual are specified, do we have a way to express that? - Bob: Only for typ/min/max, using [Model Spec]. - Arpad: That may not be sufficient if we need to model the effects of the supply noise modulating the divided or "proportional" thresholds. - Bob: It is not determined which voltage rail is the correct one for a given technology. - Mike L: It's whatever the model maker knows is correct. - Radek: Agree. - Bob: The model maker may not know. - Arpad: Should we assume everything on a [Model] uses the same reference? - For example, we might need to associate Vinl with a GND rail and Vinh with a POWER rail. Could we add something to [Pin Reference] to indicate that? - Bob: There probably is no demand for that. - Bob: The model maker needs to understand what is usually put in for common technologies. - Radek: The examples help, but we can't impose anything. - Bob: It needs to be clear enough to support the needs of future technologies. - Arpad: As long as we provide the means for model makers to describe what they need to describe instead of writing a spec with assumptions or predetermined rules, we won't have to worry about unforeseen needs in the future. - Mike L motioned to adjourn. - Radek seconded. ------------- Next meeting: 24 May 2016 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives