=============================================================================== IBIS INTERCONNECT TASK GROUP Mailing list: ibis-interconnect@freelists.org ================================================================================ Attendees from October 17, 2024 Meeting (* means attended at least using audio) ANSYS Curtis Clark, Juliano Mologni Arista Networks Jim Antonellis* Broadcom James Church 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* ST Microelectronics Aurora Sanna Synopsys Ted Mido, Edna Moreno University of Illinois Jose Schutt-Aine Zuken USA Lance Wang Michael Mirmak called the meeting to order. No patents were declared. Jim Antonellis introduced himself. He is now with Arista Networks, formerly with Broadcom. He is part of a new team now working with SerDes differential simulation, up to 224 GHz down. The team uses a lot of Touchstone files. They are looking for ways to extract differential return loss, crosstalk, and the like, writing tools and scripts focused on ports and nodes. Michael welcomed Jim to the meeting. Michael noted that the October 2 meeting minutes were not ready for review. However, he did display the September 11 minutes for approval. Arpad Muranyi moved to approve the minutes; Walter Katz seconded. The minutes were approved without objection. During an informal AR review, Randy Wolff noted that he had closed his assignment to check how reference nodes are handled in EMD; he had provided an e-mail summarizing his findings to the reflector. Arpad noted that he had talked with Randy and Weston Beal for an hour on referencing, including a review of current documentation. He noted that there are no power-delivery-related examples using Touchstone available in IBIS. There is only one A_gnd instance of IBIS Interconnect and Touchstone in an example. Randy suggested, as an editorial action, to add a power-port-based IBIS example. There are no ground ports in Touchstone in IBIS to be added. Michael asked whether EMD was examined; Randy confirmed that only IBIS Interconnect was analyzed. Arpad added that the IBIS Interconnect models/examples examined connected the reference to bus_labels, pin_names, and signal_names which would be inconsistent with the rules we were going to suggest in the Touchstone 3.0 port mapping draft. Walter provided some background on the team's referencing disagreements for Jim's benefit. From his perspective, the reference for 4 ports on a SerDes differential channel, using local ground, is represented in all simulators as node 0 (gnd), or A_gnd in IBIS. Some advise measuring using a reference that is something other than A_gnd. Walter believes this to be incorrect and violating a basic assumption. Jim stated that he agreed with the local reference concept, adding that his previous employer separated references in 3D space. Arpad replied that referencing and return currents may be two separate things. Walter responded that he did not say that return currents flow across the reference node. Randy showed his proposed rule changes for EMD to link to Touchstone: among these is a suggestion to put S-ports in <> brackets for designators and pin_names. The same would be done for P ports. Michael asked about whether port ordering was always numerical, or whether port (node) ordering could be by location, as in the case of SPICE nodes, or C function arguments. Michael took the AR to clarify the language in the Touchstone proposal that there is NO support for order significance. Naming and port numbering are the only means to identify the port connections. [AR] Arpad suggested the proposal use language from IBIS to this effect, as it should be present there too. Randy agreed. Randy added that the text was not clear as to whether the new *. syntax is covered in current Touchstone definitions. Michael asked whether a table is needed showing the comparative labels between EMD, Touchstone, IBIS Interconnect, etc. Randy responded that there was no need for duplication, but that a quick summary or reference might be useful. There is also a need for new descriptive text. Walter added that noting the physical location of references is key. Arpad noted that he found Terminal_number entries for EMD, but did not yet have information about [Interconnect Model]. Michael accepted the AR to include examples from Arpad in the next port mapping proposal draft. [AR] Arpad showed examples for File_TS from IBIS 7.2's Interconnect Model keyword, where the reference for Touchstone is connected to "VSS" (Example 3). Examples 10 and 11 use pulldown reference (Pulldown_ref) as VSS. A_gnd is used in Example 15. The proposal was to connect all of these to A_gnd, but this may not be appropriate. For Example 2, Randy noted that "Pin_Rail" and "Buffer_Rail" were used for VSS in both cases. Michael noted that, in a conversation with Scott McMorrow, Scott pointed out that separate references may not always be possible or appropriate - measurement of the entire system would be much better. Michael asked whether we need to tighten the rules regarding A_gnd usage in the text. Walter suggested that the text could make everything referenced to ideal node zero, as a simple solution. Randy added that we would accordingly need to fix the examples. Arpad moved to adjourn. Randy seconded. The meeting adjourned without objection. The next meeting will be held October 23, 2024. ================================================================================ Bin List: 1) Complete port naming proposal (Katz et al) 2) Complete/revise Touchstone 3.0 draft outline (Mirmak) 3) Complete ISS-IRD 1 Draft - enable cascading of S-parameters through W-element (Mirmak) - TABLED Tabled ARs: - Arpad to give an example of the physical connectivity needed for EMD automation.