Minutes for IBIS Futures Group Meeting

Next meeting:   Thursday, Oct 11, 2001
Dial-In Number: 1-408-894-3000 (Meeting Place)
Meeting ID:             42479 (IBISX)


Minutes
Date: Thursday Oct. 4, 2001
Meeting chaired by Stephen Peters
Minutes recorded by Lynne Green
-------
Attendance:
Stephen Peters, Intel
Lynne Green, Cadence
Bob Ross, Mentor Graphics
John Angulo, Innoveda
Arpad Muranyi, Intel


* Minutes and AR review:

Minutes of Sept 20 approved.

AR - John Angulo will upload minutes to web page for 7/31, 8/23, 9/6.
Minutes of 9/20 have been distributed.

After discussion, Lynne will take minutes for this meeting, and
continue until John wants the job back.  John is currently overloaded
with Majorodomo work.

AR - Lynne sent out IBIS ML Spec, Rev 0.3b.  Documents for internal
review will be distributed as WORD docs, and Stephen will convert
them to PDF and arrange to have them posted to the IBIS web page.
Lynne will update the spec based on this meeting and send it out.


* New Business:

Al Davis withdrew from group.


* Old Business:

tline format was discussed.
Format for ideal, lossless transmission line agreed on:
tline instance_name (node1 node1ref node2 node2ref)
         len=<len_val> Z0=val TD=<sec/unit length>
tline instance_name (node1 node1ref node2 node2ref)
         len=<len_val> R=<ohms/unit length> C=<Fd/unit length>
         L=<henries/unit length> G=<Siemens per unit length>
The <len_val> and <per unit length> must use the same units for
length (meters, inches, mils, etc), but any valid length unit
can be used.
Format for multi-conductor, coupled and/or lossy lines was
discussed.  Reference node (one per side, ideal ground ref,
or one per node) needs to be finalized.
Format for matrix and frequency dependence are still open.  There
appears to be a preference to use the same formats as in the
connector spec.
Suggested formats include:
tline 'name' (n1, ref1, n2, ref2, n3, ref3, ...) matrix_name= matrix name
tline 'name' (in1, in2, in3...inN, out1, out2, out3...outN) rlgcfile='name'
It was pointed out that in either case, the number of nodes must be an even 
number,
since each transmission line has exactly two ends.

There has been no significant progress on 4.0 bird descriptions
John has looked at BIRD62 (receiver threshold enhancement), but
did not have anything to present at the meeting.  Lynne will
look at the enhanced golden waveform load bird.


* IBIS-ML Spec Review

Section 2.1, item #6: Note that node names may be alphanumeric,
or just numbers.  Item #8 was removed.

Section 2.2, it was agree to updated the reserved word list,
and to make words case-insensitive where there was no conflict
with IBIS 3.2.  IBIS keywords and subparameters defined in the
IBIS spec are reserved within their normal scope.

Section 3: all references to "SPICE" should be removed.
Stephen pointed out that there were nodes named "output" in
the receiver example; the example and its explanatory text need
to be updated.  The concept of "IBIS Library" needs to be
better explained.

Section 4: Table 3 and associated paragraphs removed, since
this is not part of the required implementation.  All
controlled sources must depend only on terminals available to
the controlled source.  We still need to decide how to spec
current control for these sources.  Concept of "whitespace"
and allowed separators needs to be improved.

The Section covering the "Define" statement is missing.  It
should either be a stand-alone section early in the document,
or it should be incorporated into the receiver example and
by reference to the IBIS spec.

Respectfully submitted,
Lynne Green