====================================================================== 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 September 7 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 Michael Mirmak convened the meeting. No patents were declared. Review of ARs: Review of Minutes: Michael called for review of the minutes from the August 31 meeting. Walter moved to approve the minutes. Bob seconded. The minutes were approved without objection. Opens: Michael said that the buffer rail terminal naming is the main discussion topic. Walter commented that the IBIS Editorial and IBIS ATM groups have been working on prerequisite BIRDs. The only impact on those BIRDs is to our existing package solutions where the reference is ideal ground. He said the work on those BIRDs should not effect the Interconnect BIRD. Bob noted that there is an issue with confusion on reference node 0. The confusion is when we describe an interconnect model for the power/ground rails only, where ground is a terminal but cannot serve as a reference as well. The question is what to use as the reference in these cases. Walter suggested to schedule some discussion on the agenda for this topic. His position is that this decision should be up to the model maker to set up the reference. Michael asked whether we can use a signal as the reference. Bob agreed we can save this for a future discussion. Walter said there is a distinction between normative and informative rules, and these are informative rules. Michael asked about how the new BIRDs that had Radek and Bob are working on will be tracked. Walter commented that they new BIRDs on referencing will be tracked in the ATM group. Buffer Rail Terminal Naming: Michael said that Walter had sent out an email to the reflector detailing the issue. Walter's motion is to change the model rail terminal names to those described in his email. Bob seconded the motion. Arpad asked how those names match up to our existing naming conventions. Walter said that the same names are used in the C_comp BIRD. In the case of the C_comp BIRD, these are the terminals that the [C_comp Model] connects to. In [Receiver Thresholds], these nodes are the references. For [Pin Mapping], these names are the subparameters that serve as the column headers. In the Interconnect BIRD, we want to denote the terminals that connect the buffer to the interconnect model. The names refer to a specific rail terminal for a specific buffer. [External Model] is an exception, but it has a different naming convention based on analog and digital signals. Bob commented that he agrees Walter's proposal is more consistent, but his concern relates to the typing of the terminals. The terminal type stresses the location, but for the buffer it may be redundant to have "Buf" in the name. He commented that it might be nice to have a different syntax for the interconnect rail names since they have different rules. He noted that it is a choice between consistency and clarity. Arpad mentioned that he would favor to use fewer names in order to reduce confusion. His concern is on [Receiver Thresholds], as they are referred to as supply voltages. But, now we are using them as terminals. Walter commented that it is a difference between terminals and voltages, but the meaning is the same. He added that the terminal names are also used in the new BIRD181 for I/V Table Clarifications. They are used as voltages in this BIRD, but the meaning is the same. Michael suggested we put together a BIRD to fix the IBIS 6.1 specification page 43 issue on how the voltages are defined for [Receiver Thresholds]. He will take an AR to draft a fix for this section. He also asked if we are confident that we have identified all of the places where voltage referencing is an issue in the IBIS specification. Bob commented that Walter's proposal applies only to Interconnect BIRD. But, there are other referencing issues in the IBIS 6.1 specification that need to be taken up by the Editorial Task Group or by new BIRDs as these are separate issues. Michael agreed. Bob also stated a concern, noting that the Interconnect BIRD is using Buf_I/O, while the I/V Clarification BIRD is using Buffer_I/O. That naming should be looked at for consistency as well. Walter agreed and thought Mike and Bob should discuss this issue and make a recommendation. Mike and Bob took an AR to resolve the Buffer_I/O naming. Mike asked whether the first character of the terminal names should be upper or lower case. Bob answered that for all [Component] subparameters, which are column headings, we adopted lower case throughout for that section. For [Receiver Thresholds], the subparameter names begin the line and are upper case. Walter asked about how the [Pin Mapping] column headings are used. He proposed to add the upper case subparameter names as legal for these [Pin Mapping] column headings. This way both upper and lower case will work, and we will avoid confusion. Mike commented that he tested the parser, and it does not complain about the case for any of these subparameters. Michael called for any objections to Walter's motion to change the rail terminal names. No objections were stated. Arpad noticed that the Vcross_low and Vcross_high are specified with 0V on page 43 of the IBIS 6.1 specification. Mike took an AR to look into this for a possible change. Walter noted that this is 0V as a value of voltage and not ideal ground, since these are differential signals. Michael commented that it still needs to be clear how this voltage is measured. Mike moved to adjourn. Arpad seconded. The meeting adjourned without objection.