================================================================================ 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 February 28, 2018 Meeting (* means attended at least using audio) ANSYS Curtis Clark Cadence Design Systems Bradley Brim Intel Corp. Michael Mirmak* Keysight Technologies Radek Biernacki* Mentor, A Siemens Business Arpad Muranyi* Micron Technology Justin Butterfield*, Randy Wolff* SiSoft Walter Katz*, Mike LaBonte* Teraspeed Labs Bob Ross* Michael Mirmak convened the meeting. No patents were declared. Justin Butterfield took minutes. Review of Minutes: - Michael called for review of the minutes from the February 23 meeting. Mike LaBonte moved to approve the minutes. Randy Wolff seconded. The minutes were approved without objection. Review of ARs: - Mike to look for missing meeting minutes. - Mike noted that he is still working on this, and it might take some time. Michael stated he will also look to see if he can find any of these old minutes [AR]. - Radek Biernacki stated that he sent out some text modifications in regards to suggested changes for A_gnd. Opens: - Michael stated that he set up a meeting with the IEEE DASC P2401 chair to coordinate activities. He noted if anyone has any items to discuss he can bring them up. - Michael noted he will be on travel for the next few meetings, but he should still be able to attend the meetings. A_gnd Modifications: Radek stated that the changes are to clarify the A_gnd generally for different EDA tools. The second change is that if A_gnd is used it should be used as the reference for the signals. Radek added some changes for page 94 in the Multi-Lingual section of the IBIS specification. He noted that the A_gnd is similar to SPICE node 0 and global ground is not prohibited, but it is loosely defined. Arpad Muranyi asked, in the second sentence of the A_gnd definition, if A_gnd can also be the reference for the rails in addition to the I/O signals. Radek thought that this could be extended to rails as well, but the section on page 94 of IBIS Version 6.1 relates to signals. If the rails are defined as ports, then it would apply. Michael suggested to add the phrase "supply rails". Bob Ross asked about using the word "node", since the word "port" is used in the Multi-Lingual chapter. Michael asked if we should change the sentence to use "port". Bob noted that "port" is used throughout the Multi-Lingual chapter, while here it is okay to use "node". He suggested to not change the rest of the Multi-Lingual chapter. Mike noted on page 5 there is an extra column in the table. Michael noted Radek added the phrase "local ground" to the definition of A_gnd on page 6. Radek added a comment asking what is the meaning of the word "reference" in the General description of rail terminals section. He thought the word "reference" is unclear in the last bullet item. Randy thought the word "reference" only makes sense when talking about S-parameter models. Bob stated that this is for both Touchstone and IBIS-ISS models. For example, a simple transmission line SPICE model would also have a reference. Radek suggested we need to be thorough in this regard. Arpad suggested for this sentence to say "to provide a reference node". Michael stated, on page 28, Radek changed the Terminal_type A_gnd description relating it to a [Component]. Arpad asked if each [Component] would have its own local ground. Radek thought we should leave it up to the EDA tool and the user how to connect the local grounds. Michael noted on page 29 a new sentence was added regarding the A_gnd as the reference. He asked if the supply rails also need to be added to this section as well. Arpad agreed supply rails should be added. Arpad asked about the second sentence in this paragraph, and if we need to keep it. Radek thought we could delete the second sentence. Arpad noted that it is an incorrect statement. What we are trying to avoid is having node 0 inside the subcircuit. The second paragraph states how a node 0 inside the subcircuit should be treated. Radek noted the second sentence was the original text, and he wanted to keep it for the record. Arpad thought we should delete it, but we can make a note of it and decide later. Michael commented that in Example 13, node 0 is changed to the local ground node. This is only a change in the comments of the examples, and the example itself did not change. Bob noted that we need to fix the line wrap issue. Randy commented that there are some instances of "vss" that should be capitalized. Bob thought the subcircut terminal might be lower case. Michael stated we should make note of it and double check this is consistent. Michael noted Example 14 has similar changes to Example 13. Michael asked if the A_gnd changes should be rolled into the official BIRD189.5 draft. We would need to cross-check to see if we want to accept these changes. Arpad liked the idea of calling the A_gnd a local ground. But we have not really resolved the N+1 connection issue. Radek agreed that his intent was not to solve the N+1 terminal concerns. Bob stated he tends to favor using the local ground as the reference always. Arpad commented that there is text on page 27 which states what the N+1 terminal can be connected to. It only states that the N+1 terminal must connect to a rail terminal. Bob noted this includes local and global grounds. Arpad asked if we want to restrict this to only A_gnd. Radek thought it is okay as it is. Bob noted that A_gnd might not be defined as rail. Bob asked about the order of the document and if A_gnd has been introduced. Michael noted that A_gnd is previously defined in page 7. Michael asked if Arpad's concerns are taken care of with these changes. Arpad said he still has some concern if we do not restrict the reference connection to A_gnd. Radek noted that people need to understand how the models are generated. Michael asked if having a statement warning model makers that connecting the reference to anything other than A_gnd may result in non-physical results would be sufficient. Arpad thought this would be okay. Arpad will draft this warning text about not connecting the reference to A_gnd [AR]. Next Meeting: The next meeting will be March 2. Radek moved to adjourn. Arpad seconded. The meeting adjourned without objection. Task List BIRD189.5 editorial additions/changes to be completed: