IBIS Macromodel Task Group Meeting date: 24 April 2018 Members (asterisk for those attending): ANSYS: Dan Dvorscak * Curtis Clark Cadence Design Systems: Ambrish Varma Brad Brim Kumar Keshavan Ken Willis eASIC: David Banas GlobalFoundries: Steve Parker IBM Luis Armenta Trevor Timpane Intel: * Michael Mirmak Keysight Technologies: Fangyi Rao * Radek Biernacki Ming Yan Mentor, A Siemens Business: John Angulo * Arpad Muranyi Micron Technology: * Randy Wolff * Justin Butterfield SiSoft: * Walter Katz Todd Westerhoff * Mike LaBonte SPISim: Wei-hsing Huang Synopsys: Rita Horner Kevin Li Teraspeed Consulting Group: Scott McMorrow Teraspeed Labs: * Bob Ross The meeting was led by Arpad Muranyi. Curtis Clark took the minutes. -------------------------------------------------------------------------------- Opens: - Arpad noted that Michael M. will not be available to run the next Interconnect meeting. Arpad said he would like the ATM group to come up with a recommendation on the warning messages issue in BIRD189. Michael M. agreed that this would help streamline the next Interconnect meeting and allow them to focus on submitting the next revision to the Open Forum. ------------- Review of ARs: - Arpad to contact Fangyi about whether he wants to be listed as an author on the new BIRD to supersede BIRD158.7. - Done, Fangyi replied that he would like to be listed as an author. - Arpad to send edited node 0 note text to the list. - Done. Issue discussed in subsequent Interconnect meetings. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: - Arpad: Does anyone have any comments or corrections? [none] - Randy: Motion to approve the minutes. - Bob: Second. - Arpad: Anyone opposed? [none] ------------- New Discussion: BIRD189 and BIRD158 related issues: - Discussion: Arpad summarized what had happened in the previous week's Interconnect meeting. In version 7 there had been three "Important Notes", the first two by Radek and the last one by Arpad. In last week's Interconnect meeting there had been a motion to remove the third of these warnings, and the motion had passed. Arpad noted that during the discussion he thought some people didn't realize the Interconnect group had reviewed and adopted his paragraph. He referred to Interconnect meeting minutes from April 6th to confirm the group had voted to include his change. Arpad said he was okay with removing his paragraph if the group wanted to, but he noted that the rationale for doing so was not wanting to give educational or cookbook material in the spec. Arpad said he felt the other two "Important Notes" fell into the same category and could also be removed. Randy noted that he had reviewed the original proposal and history for Arpad's text. He noted that it was contained in just two emails. Randy noted that Arpad's proposal originally contained two paragraphs. In the course of the group distilling it down to the item in version 7, Randy thought it ended up losing most of its value. Randy said the remaining stand-alone sentence, stripped of the history of discussions that produced it, no longer made much sense. Randy said he thought Arpad's original two-paragraph proposal had been more meaningful, and he was okay with having such warnings in the spec. Michael M. noted that the vote to remove it was likely based on editorial fatigue, since the group had spent the previous 40 minutes hashing out the language for the preceding warning. Arpad noted that his removed paragraph was intended to warn model makers that if they used BIRD158 models they should be careful about their package models because of the BIRD158 buffer's implicit use of node 0 (A_gnd). Randy said the original text had been closer to that than the final text that had been removed from version 7. Radek noted that his only issue with the text that had been removed from version 7 was the use of "reference node for the package" because we hadn't defined that. Mike L. and Michael M. suggested that we could add Arpad's original text back in if it weren't controversial. Bob noted that he wasn't sure what the paragraph was getting at, and that he didn't like any reference to Ts4file in BIRD189. Radek noted that Arpad had preferred not to discuss packaging in BIRD158 and to consider BIRD158 a buffer modeling BIRD. He said we should have something about how generic package models would be used with BIRD158. Ideally, a package model used with BIRD158 would only contain the I/0 terminals and A_gnd. If a package model also contains the rails, then we should say something about how to use it with BIRD158. A reasonable solution would be to state that all the rail terminals will be connected to A_gnd. To Bob's point, Arpad clarified that the point of the removed sentence was that a package model used with BIRD158 really had to use A_gnd as its reference. Michael M. pointed out that the sentence did not contain A_gnd. Mike M. said that based on Arpad's description he finally understood what the sentence was trying to say. Randy pointed out that Arpad's original version had contained A_gnd and was more informative and helpful. Randy asked Arpad to share the original email (April 6th email - "Revised text for my AR"). The group reviewed Arpad's original text and compared it to the final (removed) text. Mike L. and Randy said the first paragraph of Arpad's original text had been helpful. Radek said that the paragraph was actually talking about Touchstone models needing to use the same reference terminal. So, we were mixing topics and not focusing on the Ts4file BIRD158 issue. Walter noted that the fundamental problem was confusing the concept of a reference for simulators measuring voltages vs. a reference for return current flows. He noted that BIRD158 is differential, so it doesn't matter what the reference node is for voltage measurements. The only thing that matters is where the currents flow, and there's no current flowing to the reference terminal of a BIRD158 s4p. Bob agreed, and noted that there is nothing to raise an alarm about unless you throw in an interconnect model package with a reference node. Arpad shared his "Referencing Problems" slides again (ATM archives February 27th, see slide 2). He noted that ports are connected from block to block, and you need a loop or your simulator can't calculate the voltage. Bob said he would still prefer to remove any mention of Ts4file from BIRD189, and to add it to the BIRD158.7 replacement instead if necessary. Arpad said it was in BIRD189 because we might be using BIRD189 package models with BIRD158. Arpad asked to review the "Note:" on ground 0 that had replaced the warnings. He wanted to clean up a grammatical issue with respect to what "may not account for all the currents". Radek asked that we review his original statements (Radek's April 11th reply to the "Mystery Component" email). He noted that he had been unable to attend the Interconnect meeting on that date, but he had heard from Michael M. that his original statements were reviewed favorably. However, the final text that appeared in version 8 was quite different. Michael M. then noted that he hoped to avoid re-litigating all of these various warning/note messages, and that he was coming around to Arpad's suggestion to simply remove them all. Arpad agreed and noted that we could always add more text later if necessary. Walter noted that the warnings might be better addressed with white papers discussing the issues. Arpad moved to remove all warning statement text from BIRD189.6 draft 18 version 8. Michael M. seconded. No one opposed. Michael M. took the AR to remove the warning/note messages from version 8 to produce a version 9. Michael M. noted that the only remaining topic for BIRD189 was the rail discussion. Bob noted that he would prefer to officially submit the version 9 document as BIRD189.6 and deal with rail issues in a subsequent dot revision. - Mike L.: Motion to adjourn. - Randy: Second. - Arpad: Thank you all for joining. AR: Michael M. to produce BIRD189.6 draft18 version 9 and email it to the list. ------------- Next meeting: 01 May 2018 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives