================================================================================ 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 16, 2018 Meeting (* means attended at least using audio) ANSYS Curtis Clark Cadence Design Systems Bradley Brim Intel Corp. Michael Mirmak* Keysight Technologies Radek Biernacki* Mentor, A Siemens Business Arpad Muranyi* Micron Technology Justin Butterfield*, Randy Wolff* SiSoft Walter Katz*, Mike LaBonte* Teraspeed Labs Bob Ross* Michael Mirmak convened the meeting. No patents were declared. Justin Butterfield took minutes. Review of Minutes: - Michael called for review of the minutes from the March 14 meeting. Justin noted the date at the top of minutes should be March 14 rather than March 7. Walter Katz moved to approve the minutes with this correction. Randy Wolff seconded. The minutes were approved without objection. Review of ARs: - Mike LaBonte to post missing meeting minutes. - Mike reported this is done. - Michael to draft a BIRD189.5_draft18_v4 with the proposed changes to replace "ground" with "reference". - Michael reported this is done. Opens: Michael stated that rail rule relaxation is up for discussion. Walter asked if the rail rule relaxation was in the current draft. Bob replied he was requested to add these changes as a BIRD189.5_draft18_v3. Walter made a motion to exclude the rail rule relaxations from BIRD189, since these were not agreed to, and to not use BIRD189.5_draft18_v3 as the basis for the future drafts. Mike noted there are also some editorial fixes in BIRD189.5_draft18_v3, while the rail rule relaxations are noted in red. Walter clarified that he would like to modify the motion to remove the rail rule relaxations from BIRD189. There was no second. Michael declared the motion dead. Walter stated that he cannot support BIRD189 with the rail rule relaxation changes. Randy stated that we should discuss these changes further. Michael noted that the changes are not final and still up for discussion. BIRD189.5_draft18_v4 Discussion: Michael noted he replaced most of the instances of the word "ground" with "reference", but there were a few places where "ground" made sense. There are still a few instances of node 0 as well. Radek stated that node 0 is well known for SPICE simulators. He commented we need to be careful with the usage of the word "reference", as we need to be clear what the reference is local to. Bob asked if HSPICE supports global ground, and if the subcircuit needs to bring out the global ground node. Radek noted that the subcircuit instantiation should bring out the ground. Bob noted that global ground is widely used, but we can discourage it. Walter stated it is true that HSPICE has a global reference, and you can measure single-ended voltages with respect to node 0 or measure differentially. A single-ended voltage is only valid when measured to the appropriate local reference. Michael stated we need to discuss [Component] versus [Model] reference as stated in the A_gnd text in the Background Information/History section on page 5. There was some controversy over whether the A_gnd is local to the [Component]. Walter noted that node 0 and local references are not compatible. Arpad stated that we are saying A_gnd may be connected to node 0. Michael agreed. Walter asked if VSS in a component is 1V, does this say that we connect it to node 0. Michael replied that you may connect A_gnd to VSS, and it is possible to tie A_gnd to node 0. Bob suggested to remove [Component] from A_gnd definition sentence, as we are discussing Interconnect Models not [Component]s. Arpad commented that if we are saying A_gnd is local, we need to say where it is local to. Bob stated we originally intended A_gnd to be global node 0. Radek noted we are changing the meaning of A_gnd in BIRD189. Bob commented we are changing the IBIS specification, but in practice, people will still connect A_gnd to node 0. Radek stated for different IBIS [Component]s, A_gnd may represent different nodes. Walter suggested to delete the [Component] and "node A_gnd" from this sentence. Arpad asked if we have [Model]s and they each have a reference are those references independent. Walter replied, yes, these could be connected as independent local references. This would be up to the EDA tool. We can use Pin Mapping to find the reference nodes to determine the local reference. Michael stated we need to make a decision on how to wordsmith the A_gnd definition. He thought that stating [Component] does not really resolve the locality question. Walter commented that local is in the eye of the beholder. This can be left to the EDA tool and user to decide, or the model maker can force the reference to a terminal other than A_gnd. Michael suggested to change the wording to an "EDA-tool-defined local reference node". Radek disagreed with this. He commented that A_gnd is the reference node for that specific IBIS model is used for the entire [Component]. Radek thought we need to define the level of hierarchy where the reference is local to. Arpad noted A_gnd is only used in the [Model] for BIRD158, but we have decided that these types of models are not for power aware simulations. It does not make sense to have separate references in this case. He prefers to use [Component], as it is not necessary to define the reference at the [Model] level. Michael asked if we are gated by BIRD158. Walter stated we are not. We are not doing power aware simulations in BIRD158. A_gnd can be whatever we want for BIRD158, since there is no current flowing through the reference terminal. Arpad asked if the [Model] keyword does not use A_gnd, why we need to define it that level. Michael asked Walter to create an example with a local reference using A_gnd at the [Model] level [AR]. Michael asked Arpad to create an example with a local reference using A_gnd at the [Component] level [AR]. Next Meeting: The next meeting will be March 21. Mike moved to adjourn. Arpad seconded. The meeting adjourned without objection. Task List BIRD189.5 editorial additions/changes to be completed: