IBIS Macromodel Task Group Meeting date: 16 January 2018 Members (asterisk for those attending): ANSYS: * Dan Dvorscak * Curtis Clark Broadcom (Avago): Xingdong Dai Bob Miller Cadence Design Systems: * Ambrish Varma Brad Brim Kumar Keshavan Ken Willis eASIC: David Banas Marc Kowalski Ericsson: Anders Ekholm GlobalFoundries: Steve Parker IBM Luis Armenta Trevor Timpane Intel: Michael Mirmak Keysight Technologies: Fangyi Rao Radek Biernacki Ming Yan Maxim Integrated Products: Hassan Rafat 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 TI: Alfred Chong The meeting was led by Arpad Muranyi. Curtis Clark took the minutes. -------------------------------------------------------------------------------- Opens: - Arpad noted that the meeting scheduled for January 30th will not be held because most attendees will be at DesignCon. - Walter noted that BIRD189.5_draft15_v6 had been created and mailed to the Interconnect list. ------------- Review of ARs: - Arpad to submit his Figure 29 cleanup proposal to the Open Forum as a BIRD. - Done. Submitted as BIRD193. - Mike L. to add a legend to his Aggressor_Only graphic. - Done. Mike noted that he added some descriptive text, not a legend per se. - Arpad to take Mike's new version and clone it to create an example with a pin that appears multiple times as Aggressor_only but never as a victim. - Done. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: - Arpad: Does anyone have any comments or corrections? [none] - Mike L.: Motion to approve the minutes. - Bob: Second. - Arpad: Anyone opposed? [none] ------------- New Discussion: BIRD189.5_draft15_v6 - Walter: [sharing his draft15_v6 sent in an email to ATM on Jan 12th] - Discussion: Walter reviewed and accepted (change tracking) the simple editorial corrections, none of which were controversial. For technical changes, Walter noted that "Pin_Rail terminal" had been used in the text but not strictly defined. He had introduced a new paragraph describing Pin_Rail terminals and their relation to pin_name, signal_name, and bus_label entries. Bob agreed in principle with what Walter had added, but said the actual text was self-contradictory and needed to be fixed. Walter said he would send out this version (with the editorial changes noted in the meeting) as v7, and Bob and Walter agreed discussion should continue in the Interconnect Task Group meeting. BIRD189: Aggressor_Only graphic - Mike L.: [sharing the example as updated after last week's meeting] - Discussion: Mike noted that he had added the text "can be victim" next to the rows of pins in order to define the meaning of the check (yes) or X (no). - Arpad: [sharing the example he'd prepared for the "problem" case in which a pin appeared in multiple models as an aggressor, but did not appear in any as a victim] - The original example used blue to indicate pins included in the model. - Here I've used: - green for the model containing pins 2, 3, 4. - red for the model containing pins 4, 5, 6. - red + green = brown (pin 4 appears in both, shown as brown). - Green model would be missing coupling info from 1 to 2 and from 4 to 5. - Red model would be missing coupling info from 3 to 4 and 6 to 7. - Only pin 3 (green model) and pin 5 (red model) can be used as victims. - Do we need to say something about the fact that pin 4 could only be used as an aggressor? How would we select the model if the user wants to simulate pin 4 as a victim? Walter had suggested that we could add a rule stating that the first model that included pin 4 would be used. That would be the green model in this example. - Discussion: Walter noted that he thought Arpad's example was worth adding. He noted that he had proposed the "use the first model that contains that pin" rule so there would be no ambiguity in a case like this (pin 4), but he said he wasn't sure he liked the rule. Arpad noted that he was asking the question because we had decided that the user should only have to choose which Model Group they wanted to use, and that we didn't want the user to have to make any further sub-selections. Walter noted that we could also avoid the ambiguity by disallowing the configuration in this example. Instead, the model maker could place the two models in separate Groups (thereby avoiding the issue of having pin 4 appear in two models in the same Group as an Aggressor_only but never as a victim, and thus avoiding the ambiguity about which model to choose if the user wants to simulate pin 4 alone without coupling). Arpad said we could add language to that effect. Bob supported this idea because he preferred to be more precise and avoid the ambiguity altogether. Bob/Arpad/Walter agreed that we could add language to specify this rule. BIRD189: A_gnd examples. - Bob: [sharing his A_gnd examples (13 and 14) sent to ATM on Jan 15th] - Note one editorial issue in BIRD189: We should make sure we use lower case "n" when we use the word "node". - Example 13: - Shows A_gnd used to short terminals to ground. - An s8p is reduced to an s2p by stubbing out all the other buffers to node 0. - Demonstrates that use of A_gnd is allowed on any number of terminals in an IBIS-ISS subcircuit. - Example 14: - Example used A_gnd on a terminal line to create an ideal short connection to the IBIS-ISS node 0 used internally. - I agree with Arpad's comments (reply to the original email) that this example, as it stands, is unnecessary. - Discussion: Arpad noted that he agreed with Example 13, but as a minor point he would also like the example to demonstrate A_gnd being used to provide a reference into the subcircuit, perhaps by adding a 9th terminal line to the example. Bob agreed and said he had also considered adding an example with Vdd and Vss instead of just I/Os. Arpad confirmed that he thought Example 14 was unnecessary. Instead he proposed that we convert the example to a Touchstone example and show the N+1st (reference) terminal as the A_gnd. Bob agreed, but noted that he had not created the Touchstone example because he felt the rule (that A_gnd could only appear as the N+1st (reference) terminal of an N port Touchstone file) was clear. Forward Error Correction: - Mike L. noted that he would have a DesignCon IBIS Summit presentation on this topic and that it could be removed from the ATM agenda. - Mike L.: Motion to adjourn. - Curtis: Second. - Arpad: Thank you all for joining. AR: Walter to send BIRD189.5_draft15_v7 to the Interconnect list. ------------- Next meeting: 23 January 2018 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives