(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 March 2 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. No minutes were available from the previous meeting for approval. Walter Katz presented “Currants in S-Parameters” examples, including a two-port network shown in Keysight Technologies on-line material. The “-“ terminal is the return terminal, which is usually ignored in schematics. Each port has two terminals. The return terminal is on the “ground” pad. Currents are equal and opposite. Bob Ross noted that the reference can be on the power rail. Walter replied that “reference” as a term is deliberately not being used in his discussion. Bob suggested that the term “return terminals” is appropriate. Brad Brim asked whether every terminal has a voltage. Voltage is measured between the terminals. This is a subtle point, but a necessary one. Walter agreed, adding that a voltage can always be measured with respect to the core of the earth, but that is not what is used in lab measurements. The total current in and out of a terminal block is zero. Brad clarified that the “universe” here is simply one signal node and one Vss (etc.) node. Arpad Muranyi responded that he agreed with everything discussed so far except the last slide. This assumes that the terminals on the die side are using the same terminal (due to N ports + 1) for a reference as the pin side. Walter disagreed, stating that this is a convenience for linking the return terminals to the reference node of the simulator. N+1 convention does NOT mean that the N+1th terminal is the reference; the simulator is responsible for figuring out how to connect the node to the reference. Brad and Radek Biernacki disagreed with this directly. Brad summarized three ways of connecting this kind of network: 1) terminal 1, terminal 2 – reference is zero 2) terminal 1, terminal 2, another node – no node zero usage, and the other node is used as reference 3) terminal 1, reference 1, terminal 2, reference 2 – no node zero usage Walter agreed. Using an S-element with node zero at the end of the node list means that you had better be sure that everything in that path uses reference node zero. Arpad stated that there is an implication about an ideal short between references at the pin side, and another at the pad side. You cannot simulate ground bounce in that situation. Walter replied that he will correct the title of the slide to, “So What Does the S-parameter Shortcut Mean?” Radek commented about the distinction between terminals and ports. Walter emphasized the 2N vs. N+1 distinction. Radek replied that one cannot infer the reference if you are not told what it is. Walter responded that the near terminal is used as reference. Radek repeated an example case described in the last meeting: imagine a traditional IBIS package RLC with only R and L. This is two terminals, or 1-port data in Touchstone. Reference would be a shunt element that does not exist. Brad replied that this situation could not be handled through the shortcut approach. Radek disagrees, stating that he wanted to have 100% equivalent treatments for all circuit topologies. Arpad added that the shortcut is based on an assumption. Radek noted that the original solution had a certain number of terminals available for connection, with N+1 terminals for TS data with N port. Brad suggested that 99% of models that Walter was concerned about would be covered by the shortcut. If the shortcut is not desirable, then perhaps requiring a wrapper in all cases was appropriate. Brad presented one slide, due to time constraints. The objective is to provide a shortcut so that you don’t have to provide an IBIS-ISS wrapper for many cases. A key requirement is that all simulator tools would use a common format for netlisting the shortcut in a universal, identical fashion. Brad and Walter agreed that “node” and “terminal” are not synonymous. Next time, Brad will complete his presentation and the team will resolve the shortcut approach. Arpad moved to adjourn. Brad seconded the motion. The meeting adjourned.