IBIS Macromodel Task Group Meeting date: 31 March 2015 Members (asterisk for those attending): Altera: David Banas ANSYS: * Dan Dvorscak * Curtis Clark Avago (LSI) Xingdong Dai Cadence Design Systems: Ambrish Varma Brad Brim Kumar Keshavan Ken Willis eASIC David Banas Ericsson: Anders Ekholm IBM Steve Parker Intel: Michael Mirmak Keysight Technologies: * Fangyi Rao * Radek Biernacki Nicholas Tzou Maxim Integrated Products: Hassan Rafat Mentor Graphics: * John Angulo * Arpad Muranyi Micron Technology: * Randy Wolff Justin Butterfield QLogic Corp. James Zhou Andy Joy eASIC Marc Kowalski SiSoft: * Walter Katz Todd Westerhoff * Mike LaBonte Synopsys Rita Horner Teraspeed Consulting Group: Scott McMorrow Teraspeed Labs: * Bob Ross (Note: Agilent has changed to Keysight) The meeting was led by Arpad Muranyi. ------------------------------------------------------------------------ Opens: - Arpad: Would like to discuss a package modeling topic. -------------------------- Call for patent disclosure: - None ------------- Review of ARs: - Michael M update AMI Directionality BIRD - no report. - Arpad to review IBIS specification for min max issues. - In progress. ------------- New Discussion: PAM4 BIRD: - Walter: We have an updated PAM4 BIRD to send out. Package modeling: - Arpad: We have new package formats but the legacy formats might still be used. - Arpad showed a presentation. - slide 1: - Arpad: This is without [Pin Mapping]. - The inductors might not be present. - In case 1 there might be [Package] or [Pin] RLC only. - In case 2 there is a [Define Package Model] - In case 3 the [Define Package Model] does not define some pins. - In that case should the unmodeled pins use other mechanisms? - Or should it be a short? - There is no information on how power and ground pins are connected. - Signal models should come form the package or pin. - slide 2: - Arpad: This shows [Pin Mapping] shorting pads together. - Some inductors might get shunted together if the pins are also shorted together. - The question remains, how should unmodeled pins be handled? - Merging parasitics into one pin might make sense. - The specification does not say how to create this model. - slide 3: - Arpad: This shows power and ground pins not defined in [Define Package Model] - Maybe the [Pin] RLC or [Package]should be used. - Using a short is a possibility. - That would create some"packageless" pins. - The pins could be shorted together. - Without a combined model that will not work. - slide 4: - Arpad: Not all power and ground pads are shorted here. - The spec doesn't tell us what to do. - Those other pins can't be simulated anyway. - slide 5: - Arpad: This is from Randy's presentation. - Randy: You may get ports lumped into a single capacitance. - To make matrices the same size a lot of post-processing would be needed. - That might be error prone. - slide 6: - Randy: Without [Define Package Model] there is a question how to split up capacitance. - With just [Package] it usually represents only signal pins. - slide 7: - Arpad: Only a complete [Define Package Model] is correct for PI. - Merging parasitics would be desirable,but how will simulators deal with it? - This might be worthwhile because there probably will be more legacy models. - We would need to have this in the specification. - Bob: It would have to be clear to users that only 1 power and 1 ground would be used. - Arpad: Agree, the BIRD hopefully captures that. - Arpad showed a draft BIRD "Power Pin Package Modeling" - Arpad: The analysis part is similar to the slides. - No name yet for the new sub-parameter under [Pin Numbers]. - One paragraph becomes several paragraphs. - The "as many names as" clause has been clarified. - That needs to match the original intent of the spec. - Mike: The order of hierarchy might be more explicit here. - Arpad: For missing pins it states "may be merged", it's not required. - Radek: It should be more precise than "may" or "should". - Arpad: The sub-parameter tells the tool there is a merging pin. - The sub-parameter is followed by a list of pin names. - They have to be in [Pin] and have to be power or ground pins. - Without [Pin Mapping] PI can not be analyzed. - There are invalid combinations. - With [Number of Sections] the new sub-parameter is not allowed. - Randy said that is likely only with matrix formats. - The other sub-parameters are allowed only when the new one is absent. - There are other rules about names and buses. - Pin names may span lines, but each line starts with the new sub-parameter. - It may list only pin names not in [Pin Numbers]. - The EDA tool must short those pins to eliminate the possibility of disconnected pins. - Undefined pins must be disconnected. - Radek: Does it have to be exhaustive or a subset of [Pin Mapping]? - Arpad showed IBIS 6.0. - Arpad: You still can have a mixed case. - Bob: There is a question if we want that rule. - Radek: We might have unusual groupings. - What happens to pins not specified in [Pin Numbers]? - Arpad: We could have a requirement for that. - Randy: Agree, we could check for that. - Bob: I would prefer a new keyword like [Merged Pin Numbers]. - Arpad: The merged pin would list the pins it merges. - This compact. - A new keyword might make it harder to see what is merged. - Bob: It does not apply to the uncoupled case. - Walter: Presumably Micron has models that tools don;t know how to handle. - Randy: Yes. - Walter: It would help to see the IBIS file as an example. - Randy: OK. - Walter: Will those models be converted to new formats? - Randy: We have old models that will not change. - Curtis: This BIRD defines a way to address the ambiguity, but there will still be older models out there. - Do we need to clarify how we expect things to be handled in these situations with older models? - Randy: We might look at going back and updating our older models. - Walter: We introduce the BIRD, and for older models we assume user documentation or some other source explains what the model expects. - Arpad: There is a question which version update this might go into. - It could go to editorial to get it in sooner. - Bob: I'm not sure the sub-parameter is needed. - The buses might be merged by rules. - Arpad: That may be true, I need to consider that. AR: Randy provide example for new package BIRD draft. AR: Arpad send draft package BIRD to Mike for posting. ------------- Next meeting: 07 Apr 2015 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives