====================================================================== 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 March 11 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. Radek Biernacki commented on the minutes. He asked whether the word “fiat” was correctly used, and additionally asked that the statement attributed to him – “Global node zero is OK to use as a reference” – have the phrase “but only under certain circumstances” added to the end. Radek moved to approve the minutes with the changes noted above. Mike LaBonte seconded. No objections were raised and the minutes were approved. No opens were raised. Mike shared one of Walter Katz’s documents, entitled “GND BIRD”. Mike will highlight what needs to be done. Some slides in the document are for background; Mike will review it, looking for context statements and explicit changes. Walter suggested reviewing the presentation in this meeting. Bob Ross suggested that we need to fix the “must not be used” list in the beginning of Section 3 of IBIS. Instead, we need to define what we mean by Terminal names, Node names, etc. GND usually refers to a single pin name in a pinlist. He suggested reviewing the usage of GND as a signal name in a pin list (e.g., VCC, VDD, etc.) Walter suggested fixing Figure 16 in the specification, captioned “External Reference – used only for non-driver modes”. He would like to see this diagram broken up into perhaps four parts, with clear labeling of reference points/connections, rather than [External Reference], [POWER Clamp Reference] (which are voltage values). Walter added that Figure 16 is using “test fixture ground” for purposes of measurements. Two cases must be covered with diagrams such as these: the first is a simulated situation, while the second would be the measurement situation. Radek noted that Walter may be trying to suggest showing supply connections individually (meaning, situations with different references), plus hidden supplies. Mike suggested that the object of phrases such as “connected to” should always be a node. Radek stated that we need to talk about terminals, references, and nodes separately. We have Pulldown_Ref, we have name of column in Pin Mapping, etc. but we need to clarify each of those contexts/usages. Bob agreed. Pullup Reference is just a reference voltage with respect to some reference (as if a test fixture were connected to it). We could technically separate this from pins. Reference nodes do not need to be explicitly stated as pins. Arpad Muranyi noted that, when this text was written, the fixture that was assumed as a “GND” that would not need to be a pin. Mike suggested a table of definitions be generated. Bob volunteered to generate this table. Walter noted that a key issue is values vs. nodes or rails: for instance, [Pulldown Reference] is a number with respect to a particular pin. Arpad asked whether, if we define references for all terminals (e.g., [Pulldown Reference]) to the GND symbol, we would need to define Vinh, Vinl referencing? Radek suggested that the team needs to address each of these issues sequentially. Arpad noted that Fig. 16 does not need to have “GND” as a separate pin but is a measurement reference. Radek made reference to the input ports and output ports concept; the figure doesn’t have to refer to a device-under-test. Walter described several cases: 1) Legacy case: when Pulldown_Ref and GND Clamp_Ref are both zero but hooked up to a terminal on the device. 2) When the test fixture connects to *another* ground on the device 3) No 0 V or GND pin at all on a device (e.g., RS-232). With no reference for the component, how do you apply -2 V to Vcc and +2 V to Vee without a ground pin? Radek stated that the specification needs to have statement of C_comp and related structures and how it is connected to external reference. Arpad added that C_comp is physically connected between pad and substrate (though substrates could be floating) Walter noted that, for a DUT, C_comp could be connected anywhere, as the rails are stable; if voltages float, then the reference for C_comp floats too. Radek added that V-t tables and referencing should be addressed. Walter replied that the reference is “device-in-action” as opposed to a “device-under-test”. Arpad noted that split C_comp covers some of this. Walter noted that page 72 of the specification also states that, “It can also serve as a reference for C_comp, unless C_comp is optionally split into component attached to the other reference voltages.” We need to get the legacy cases correct first. Bob noted that a value of zero is mentioned without a physical reference point. Also, as a separate task, the specification must include PECL model coverage, with Pulldown Model type reference used as Pullup Reference (see note on top of page of 73). Arpad asked whether there is separate referencing for ECL, PECL and their thresholds. Bob replied that, whether we show it or not, there is a ground (reference) used for extraction (though these may not be devices-in-action). Arpad added that the group should seriously consider expressions for Vinh, Vinl to address referencing. Bob replied that these are hardcoded values, based on datasheet rules. Arpad suggested that these concepts may require specification extensions or additional capabilities. For next time, two items will be discussed 1) Bob’s table of definitions 2) Mike’s task list for the team Arpad moved to adjourn. Radek seconded the motion. The meeting adjourned.