==============================================================================

IBIS INTERCONNECT TASK GROUP
Mailing list: ibis-interconnect@freelists.org 

==============================================================================

Attendees from May 27, 2026 Meeting (* means attended at least using audio)

Chipletz			                 Stephen Newberry
Intel Corp.                          Michael Mirmak*
MathWorks                            Walter Katz*
Siemens EDA                          Weston Beal, Arpad Muranyi*, Randy Wolff*
Synopsys                             Ted Mido, Edna Moreno

==============================================================================

Michael Mirmak called the meeting to order.  No patents were declared.

Arpad Muranyi previously noted a typographical error in the April 22 minutes:
 "woudl" appears at least once.  

Regarding the May 20 minutes, Arpad sent the following to the reflector:

	... “supproted” -> “supported”

	Also, I think you should mention that after the discussion following my 
	concern “about the Bus_label:designator:name syntax” I withdrew my concern 
	and the document was actually correct as written.  You might also want to 
	note that Randy suggested to explain/clarify the text in this area to 
	eliminate confusions or misunderstandings in the future.

Arpad moved to approve the May 20 minutes with the corrections noted.  Randy 
Wolff and Walter Katz both seconded.  The minutes were approved without objection.

Michael reviewed the April 22 minutes.  Arpad moved to approve the minutes with
the correction noted; Walter seconded.  The minutes were approved without objection.

Michael recapitulated the concept of having cross-references between the Port 
Mapping proposal examples and IBIS 8.0 examples, without having Touchstone 3.0
include the IBIS text.  Arpad voiced concerns about "pointing to a moving target" 
when trying to ensure examples stay current.  

Michael asked whether chapter numbers should be noted as part of the example 
number format, so that the general order and associations can be maintained 
when updates are made to the documents.

Walter stated that this approach would be fraught with problems.  He suggested
just keeping the references to a specific document and version number.  Randy
added that IBIS version 8.0 would be specific enough.

Michael reviewed new examples in a separate presentation.  He noted that there
were both EMD and IBIS Interconnect ambiguities that needed resolution.

Randy stated that reserved names exist for Physical connections, including
 pin.pin_name and buf.pin_name formats.

Walter objected to example 2b in IBIS 8.0, as VSS is used as a port on an 
S-parameter block, but it should be the reference.

Michael wondered whether a less controversial example exists.  Arpad noted
that IBIS would need N+x terminal referencing to allow deletion of this example.

Walter replied that one would need to enable 2N definitions in IBIS as well.
Randy added that this is the ultimate solution.  Walter stated that, if you 
use N+1, then you have the same reference at pin and pad side.  Why have port 15?  
The definition in IBIS was likely done when we didn't understand referencing 
but only SPICE syntax.  Every signal port has a reference terminal, which should
be a local terminal.

Arpad agreed that Example 2b ideally would remove terminals 7 and 14, with 
renumbering of the remaining terminals, and use of terminal 15 as VSS.

Arpad added that he receives models with VDD and signals but no mention of VSS; 
models end up with dangling VSS connections.  Randy noted that Arpad earlier 
showed a way to create this in simulation; we could use example 15 with only 
signals and A_gnd.  

Michael noted that he saw a lot more IBIS-ISS examples than Touchstone examples 
in IBIS.  

Randy suggested reordering the Port Mapping examples to show the fully legal 
ones first; the EMD- and IBIS-Interconnect-like examples may be handled later.
The pin.x, pad.x, and buff.x format shows the Physical format.

Arpad asked whether, for IBIS-Interconnect, we have bus_labels.  We do for EMD.
Randy highlighed Example 5, where pin:Bus_label: etc. syntax is mentioned but 
not defined.

Arpad stated that a pin name is individual; if you use bus_label, how can you 
use one pin name to identify an entire bus?

Michael asked whether the usage and definition of Net are still acceptable.  
Randy and Arpad replied that the treatment of Net is still OK.

Randy noted that there was was a Touchstone example for EMD in IBIS, in  
Section 13.7, with the following definition:

	[EMD Model] DQ1_Victim
	File_TS DQ1_Victim.ts

Michael took an AR to use this format for the EMD example in Port Mapping.  [AR] 

Arpad moved to adjourn; Randy seconded.  The meeting adjourned without objection.