(attaching a text version of the minutes for ease of archiving) ====================================================================== 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 April 20 Meeting (* means attended at least using audio) ANSYS Curtis Clark Cadence Design Systems Bradley Brim Cisco David Siadat Intel Corp. Michael Mirmak Keysight Technologies Radek Biernacki*, Ming Yan Mentor Graphics Arpad Muranyi* Micron Technology Justin Butterfield*, Randy Wolff* SAE ITC Maureen Lemankiewicz, Logen Johnson Signal Integrity Software Walter Katz*, Mike LaBonte* Teraspeed Labs Bob Ross* University of Aveiro in Portugal Wael Dghais Arpad Muranyi convened the meeting in Michael Mirmak's absence. Randy Wolff took minutes. No patents were declared. Arpad called for comments on the minutes of the April 6 meeting. Radek Biernacki moved to approve the minutes. Bob Ross seconded the motion. No objections were raised and the motion carried. No opens were raised. Mike LaBonte shared a presentation "IBIS 6.2 Editorial Resolutions". He added date codes to individual pages. Page 5 was added to mention that IBIS may describe a Device In Action. Radek mentioned that IBIS should be a combination of DIA and DUT data. Also, AMI describes a DIA. Walter Katz noted he came to the conclusion that a person writing a package model can specify the reference node for W-line and Touchstone files in a correct way, because package models can have a reference go to any pin and any buffer. But model makers can be sloppy and have node 0 in subcircuits. When using a shortcut, the reference node should reference a pin that is power or ground. A model maker could choose to reference a node that is a ground pin, and a simulator can assign node 0 to that pin. Bob noted the issue is if we should provide guidance or not about how to hook up a model. Walter said that a normative description for the shortcut is that the reference is a power or ground. An informative description is saying you should hook up to a power distribution, but in a ground referenced system, the EDA tool could hook up grounds to node 0. Should informative rules be in a separate document? Bob noted it could be informative to say what node n+1 is hooked up to. Walter thought this should be a normative rule that the n+1 node hooks to a power or ground pin. Slide 7 was updated to show a new statement to be added in the IBIS specification to clarify the reference node. This would be a new rule 12. Bob noted that this statement needed to be expanded. Mike noted that the "except where otherwise stated" means we have to make sure we state the exceptions clearly. Radek stated that having one reference node is fine. When simulating, you can choose to hook that reference to node 0 or to something else. The only problem is if you have an IBIS-ISS subcircuit that includes node 0. Then you have no choice but to have the node hook to node 0. Bob thought we should make a statement that use of node 0 is not recommended. Bob thought we may need to state that all voltages in IBIS should be considered 0 potential unless otherwise stated. The group discussed draft 32 of the Interconnect BIRD. Uppercase "m" was used in the name "Model". The Terminal_type names were updated. Randy Wolff noted that the table title shows "Terminal_Type" but should be "Terminal_type". Radek noted that the "Terminal Line rules" section needs some editorial work. Bob stated that sections need some titles for each of the Terminal_type columns. We should also refer to the entries as columns. Bob noted that, in the table, Aggressor is not a Terminal_type_qualifier, so the separator line might need to be extended to indicate that. Bob requested that the word "field" be replaced with something else. Radek suggested using "column entry". Mike noted he would finish some editorial work and post an official draft 32 (without making a new draft 33). Bob noted that all examples need some updating of new Terminal_name spellings. Mike noted that we need to address comments added in the draft. Bob noted that some of the text on definitions in the analysis section may need to be included in the BIRD text and would need careful review. An example is that the mention of "I/O" would need to be very specific in applying to all [Model] types. Mike asked about the comment mentioning the need to describe the 1:1 pin/pad or pin/buffer assumption. Randy replied that this is covered in comment #7. Radek thought we should get rid of comment #6 so we are not mentioning EBD and EMD. Bob noted the diagram on page 2 does not look complete. Bob moved to adjourn. Mike seconded the motion. The meeting adjourned.