====================================================================== IBIS INTERCONNECT TASK GROUP http://www.ibis.org/interconnect_wip/ Mailing list: ibis-interconnect@freelists.org Archives at http://www.freelists.org/archive/ibis-interconn/ ====================================================================== Attendees from September 8 Meeting (* means attended at least using audio) ANSYS Curtis Clark Cadence Design Systems Bradley Brim Cisco David Siadat Intel Corp. Michael Mirmak Keysight Technologies Radek Biernacki, Ming Yan Mentor, A Siemens Business Arpad Muranyi* Micron Technology Justin Butterfield, Randy Wolff* SAE ITC Maureen Lemankiewicz, Logen Johnson Signal Integrity Software Walter Katz*, Mike LaBonte* Teraspeed Labs Bob Ross* University of Aveiro in Portugal Wael Dghais Mike LaBonte convened the meeting. No patents were declared. Randy Wolff took minutes. Review of ARs: - Walter to send out the draft with the proposed change to groups for everyone to review. - This was done. - Arpad mentioned he took an AR to write up some text on referencing, even though it wasn't noted as an AR in the minutes. Opens: Bob, Arpad, and Walter began discussing a recent email exchange and potential changes related to referencing of Touchstone files. Walter noted the exchange was about having three types of files, one where we can have File_IBIS-ISS where all terminals are defined, a second with File_TS where there are N+1 terminals, and a new third option called File_TS0 where there is no N+1 reference terminal defined. This would create a requirement to specify where the terminal N+1 is Node 0. Review of BIRD189.5_draft6_groups_2.docx document: Walter noted that one decision to make related to groups is if we call the new keyword [Interconnect Model Set Group] or [Interconnect Model Group]. This is TBD. The concept is you have groups, and each group has a name. It is academic if there is one group, but if more than one the EDA tool has to choose which to use. Interconnect Model Sets are in a group, and all are used within that group. Mike wondered if there should be a default in the list of groups such as with [Model Selector]. We didn't specify the default with Model Selector originally, but every EDA tool started using the first item in the list as the default. Walter noted he agreed with Mike. Bob shared his document draft6_groups_comments.docx. One problem he saw was the rule stated that "If an Interconnect Model Set name appears in different Interconnect Model Groups, then the file containing the two Interconnect Model Sets must be identical". Walter said it should say "file name". It is the combination of the file name and the set name in the file that is important. Walter thought the entire sentence should be deleted. He though the previous line should be deleted too, since two Interconnect Model Sets could have the same name and come from different files. This may not be good practice, but it should be allowed. Bob noted an exact repeat of an Interconnect Model Set name and file name within a group should be an error. Walter added the EDA tool should only look at one group and not across groups. All the sets within the group are put within one bucket and used. Bob noted his comment that the [Interconnect Model Set Group] should be scoped by [Component]. Walter recommended we accept Bob's changes and vote on use of [Interconnect Model Set Group] or [Interconnect Model Group]. Bob noted there is an implication with [Interconnect Model Group] that we are grouping Interconnect Models instead of Interconnect Model Sets. He prefers [Interconnect Model Set Group]. Walter also added the possibility of [Interconnect Set Group]. Arpad said he prefers the full name related to the keywords of [Interconnect Model Set Group]. Randy agreed. Walter moved to make the name [Interconnect Model Set Group]. Arpad seconded the motion. There were no objections. Walter suggested an AR for Bob to add his changes to Walter's document for a draft 7. Bob requested we review the additional changes before a draft 7 is prepared. Arpad commented that the new draft should become the official draft 6. Mike thought draft 6 was already posted, so this should become draft 7. Walter noted that the examples could be improved, but those could be done in a new draft. Arpad commented that the Example 4 was expanded during real time editing in a previous phone conference, but it needs to be updated to fit with existing examples. Mike noted that Bob missed an instance of [End Interconnect Model Set Group] keyword, spelled as [End Interconnect Model Group]. Referencing Topic: Walter asked if we have established a requirement for a TS file for either the ability to assign a pin by signal_name with a model_name of POWER or GND to terminal N+1 or to assign A_gnd or Node 0 to the N+1 terminal. What are the three ways we can implement Node 0 on terminal N+1? First, if we don't include it then terminate it as Node 0. Second, Bob recommends we require File_TS with terminal N+1 or File_TS0 where it is implied to connect to Node 0. Third, Arpad proposed we could have a new type of terminal that is Node 0. Arpad stated there could be a fourth type, already handled in his software, where if we have a buffer with no POWER or GND pins in the Pin List, then the tool looks at the [Voltage Range] or other keywords and attaches voltages to power the buffer. The requirement would be to connect the reference terminal to any of the model terminals because they would be powered no matter what. Walter noted if you have buffers with different power rails, that complicates things. If we had a parser for it, then if you don't have terminal N+1, then make it Node 0. That is the simplest thing to do. Arpad thought the subparameter suggestion from Bob is the next simplest thing to do. Walter noted he is ok with either idea. Arpad noted if we use the File_TS0 idea, then we need to change the wording of the number of terminals rules to adjust to having two types of Touchstone file entries. Randy asked when one would have a Touchstone file with a reference of Node 0. Arpad noted this could happen for a buffer-only model with no POWER or GND pins. Walter noted that once draft 7 is posted, people can add some ideas for referencing and make proposals. Mike asked if we should cancel the meeting on 9/13, as Mike, Justin and Randy will not be in attendance. Everyone agreed to cancel, so the next meeting will be 9/20. Mike noted that good practice for change proposals would be to name the document with the original draft name and "plus" and a word describing the change. Walter moved to adjourn. Arpad seconded. The meeting adjourned without objection.