====================================================================== IBIS EDITORIAL TASK GROUP http://www.ibis.org/editorial_wip/ Mailing list: ibis-editorial@freelists.org Archives at http://www.freelists.org/archive/ibis-editorial/ ====================================================================== Attendees from May 27 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 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. Mike LaBonte moved to approve the minutes of the May 5 meeting. Randy Wolff seconded. No objections were raised and the minutes were approved. Walter Katz raised the issue of Pin Reference as discussed in ATM. He noted that Bob Ross added notes and comments regarding ECL/PECL and other non-CMOS technologies as part of the discussion. Walter proposed that a “technology” keyword be introduced and a BIRD written to add it. He suggested the team should simply assume one of these will be approved, and that buffers will be referenced to some other rail voltage. Arpad Muranyi objected to the assumption of approval. Bob added that it was not necessary to assume approvals at all, as no dependency on those BIRDs exists for Editorial work. Bob prefers getting rid of Pin Reference – he stated that it is incomplete and limited, and doesn’t deal with supply rail scaling issues. We can proceed with editing references anyway, as we have to support them. Michael asked whether ATM is taking the Pin Reference proposal up as an ongoing topic. Radek Biernacki stated that he thought Pin Reference is needed for proper reference treatment. It is academic for DUTs, but does apply to several sections where DUT does not apply (e.g., [Receiver Thresholds] measurement thresholds in Vinh, Vinl, [Model Spec], etc.). Bob agreed. Regardless of whether we are calling it Pin Reference or scaling, it describes the same concept: re-mapping to declared rail voltages. He does not agree with Radek that a Pin Reference declaration is required. Walter added that something *like* Pin Reference is required. He proposed proceeding assuming it’s needed. We should always assume scaling even for models without explicit scaling. Bob replied that what we declare within IBIS should be the proper way. IBIS is limited to the degree to which one can modulate the thresholds. We can release IBIS 6.2 with or without this treatment. This is a separate issue from C_comp. Michael asked whether we can separate this issue from the rest of the reference revisions. Arpad noted that the original question was what our references were; we addressed the voltage rails, but then the question of thresholds came up. Bob stated his position that they are referenced to node 0 reference in simulators, and used for rails. Arpad disagreed, asking about ground bounce and voltage shifts. Bob replied that scaling solves this. Michael reiterated his question about separating the issues. Bob suggested that they could not be separated. Mike replied that one should avoid using a voltage or a number as reference, and added that the reference is not really node zero. Walter added that, to move forward, we simply have to stay there’s a methodology for thresholds. Michael responded that the discussion strongly suggests the issue is a gating item to completing the referencing revisions to IBIS. Bob replied that we should fully vet the BIRD in ATM. He added that the work is not gated, as Pin Reference is a totally new capability. Arpad asked what the team should do with the current thresholds language. Bob noted his disagreement with Mike regarding lack of node 0. Michael suggested that the team eliminate the model subparameter language in proposed introduction, and revise it later based on the Reference BIRD resolution. Arpad replied that we can’t take out the statement. We still have to rewrite the model for every different reference for thresholds. Some live editing was conducted on the current referencing language. Arpad asked whether package referencing is addressed in the document. Mike asked, just in case it’s vague, whether we are excluding some voltages in terms of referencing here. Michael suggested we use the proposed introductory text to compare against Mike’s issue list to say whether each issue is resolved or not. Arpad stated that for packages, there is no G, no parallel R, only RLC even in matrix format. If only C effects apply to packages, we should be OK no matter what the voltage is. Bob replied that, if it is not defined otherwise, the self-C matrix for a coupled matrix goes to ground. This is similar to S-parameters in that it smuggles in the idea of an ideal reference. Arpad asked whether it is possible that the C reference is a moving, noisy reference. Does the statement refer to a DC source? Michael responded that the concept that a ground reference “moves” is incorrect. A reference cannot move, as it would need another reference against which to move; any reference can be used, as long as the math is consistent. Conversions can be done between different nodes as references. Arpad asked whether this implies a rock-solid reference for thresholds. The problem is that the I-V curve reference for DIAs is a moving target. Michael noted that we must distinguish between DIA vs. DUT. Correct math transformations between the two must be used and stated explicily. Bob added that one can remap DIA in simulation to an output terminal using DUT data. Michael will send out the current text with revisions plus transformation language separately. Next time, the team will review Mike’s list against this statement. Mike moved to adjourn. Arpad seconded the motion. Walter Katz noted that objections should be called for before deciding upon adjournment. No objections were raised. The meeting adjourned.