============================================================================== IBIS INTERCONNECT TASK GROUP Mailing list: ibis-interconnect@freelists.org ============================================================================== Attendees from January 7, 2026 Meeting (* means attended at least using audio) ANSYS Curtis Clark, Wei-hsing Huang, Juliano Mologni Arista Networks Jim Antonellis Broadcom James Church Chipletz Stephen Newberry Intel Corp. Michael Mirmak*, Xiaoning Ye Keysight Technologies Ming Yan Marvell Steve Parker MathWorks Walter Katz* Micron Technology Justin Butterfield Siemens EDA Weston Beal*, Arpad Muranyi*, Randy Wolff* Simberian Yuriy Shlepnev ST Microelectronics Aurora Sanna Synopsys Ted Mido, Edna Moreno University of Illinois Jose Schutt-Aine ============================================================================== Michael Mirmak convened the meeting. No patents were declared. Michael reviewed the minutes of the December 17, 2025 meeting. Arpad Muranyi noted that the minutes incorrectly stated that: Arpad requested that one cannot get to second case without assuming the "hidden" reference of 0. He observed that this was not a request, and suggested that "requested" be changed to "clarified". Arpad moved to approve the minutes with this change; Randy Wolff seconded. The minutes were approved without objection. Michael reviewed draft 24 of the Port Mapping proposal. Arpad asked about whether the Units parameter should be singular or plural. Weston Beal observed that IPC 2581 uses "unit". He asked about what the IEEE style guide used. Michael took the AR to send out links to the relevant IEEE units documents. Walter Katz observed that ODB++ uses "units" as do many other tools. Weston noted that the current trend is to use singular terms in standards. The International System of Units states that the term should be singular. NIST does not pluralize. The IEEE Editorial Style guide promotes singular. Arpad stated that EBD has "Len" but no subparameter for the associated units. Michael changed the text to use a singular spelling. In a discussion regarding spacing around the colon character for Units, Arpad and Weston both expressed a preference for extra spaces, for readability. Walter added that, in IBIS 8.0, both Unit and Units are used, suggesting the team should maintain "Units" in Port Mapping, for seconds, farads, etc. Randy raised a concern regarding the use of Group for Type P ports vs. Type S ports. Group does not exist in IBIS. If we define physical location information in Touchstone, how does that relate to IBIS? Walter answered that PIM uses this concept; don't we also have something like that in IBIS Interconnect? Arpad replied that we don't call it "group" (a named group); it can be a bus label name. Arpad asked what the issue was. Randy replied that using Port Mapping to create an EMD for a Touchstone file is possible, but if you have Group in Touchstone but not in IBIS, the automation would seemingly be broken. Arpad answered that the group expands the port terminal to multiple IBIS designator pins, similar to bus labels or designator names. Randy suggested that maybe it just needs a limitation, such as only pin names instead of allowing xyz coordinates. Weston: it's the same thing if it's a pin or physical location. Multiple connection points as a short or single connection points. Arpad replied that it's the same as Randy's merged pins in IBIS. Randy asked how one creates a bus label. If you are listing some physical XYZ coordinates vs. pin numbers, you can find pin numbers in an IBIS file, but what do XYZ coordinates mean? Arpad replied that XYZ coordinates are on the board side of EMD. Weston asked Randy whether he expected the coordinates to match up against an EMD pin. Randy confirmed he did, suggesting it was required. Weston replied that it's all informational, and potentially useful to the user. E.g., a probe could hit in the middle of a trace or area fill. The user can decide how close this is to a component IBIS model. Arpad stated that it's really about the automation. Randy agreed, asking whether it be limited to what's needed for useful IBIS Interconnect data. IBIS Interconnect and Package should be pin numbers, not coordinates. Walter added that the elephant in the room is what is the reference of a port. The same question applies for rail nets and where you are measuring things. It's not a group of pins but instead one point for use in mating to a board or making a measurement. Randy observed that Port Mapping's Example 3a, 3b, etc. are correct but don't involve IBIS. Arpad stated that Group is for use in generating EMD; Walter suggested it can *help* but is not required. Michael suggested a rule that one could either ignore Group or, when EMD/IBIS Interconnect is present, Group could be prohibited from using coordinates. Randy agreed. Arpad asked whether the EMD pin list make use of XYZ coordinates on the motherboard. Michael noted that IEEE 2401 LPB has an explicit map of physical to model to geometry, including IBIS and Touchstone. Arpad replied that EMD has a pin list, connecting EMD pin to RefDes Pin; the same applies to Interconnect in IBIS but without XYZ. This should support Buffer, Pin, and Rail only. It would be useful to prohibit Physical XYZ connections when Data_usage is IBIS_Interconnect, EMD, etc. Michael added that Ts2file and PIM are likely to be needed in the near future. Walter noted that one can have a measurement at a test point on a via. Randy suggested that two fields be supported for Physical. Walter suggested a new name-value pair called Probe. Randy alternately suggested Probe Location, noting that the probe must use XYZ coordinates. Walter added that any Port can have Physical, Probe, both or none. Michael took the AR to add Probe pair definitions to the document. [AR] Arpad noted that Examples 19 and 20 return to transistor examples; we could get rid of these. Michael answered that they were there to demonstrate UD. Arpad suggested moving these examples earlier in the document, using subletters for the new examples. Randy pointed out an extra space in the . usage for Group on page 9. Arpad also mentioned that Refdes U1 should be U3 in some cases. He advised checking the Refdes numbers in the examples. Arpad moved to adjourn; Randy seconded. The meeting adjourned.