David, and All,
Regarding your proposed definition:
“A so called “ground based” simulation is one in which no component resides in
between any two nodes considered to be voltage reference nodes.”
while this is true, I don’t think it fully captures the point. It is possible
to make a
model that has nothing between reference nodes (A_gnd in the drawing below)
while not following the “Ground Based” modeling concepts in Sam’s presentation:
Anyway, I feel we are going on a tangent with this. I would like to summarize
what brought us here.
In the work on BIRD189 we stumbled on the question on what syntax for the N+1th
terminal line in
the [Interconnect Model] keyword should be. I don’t remember the chronology,
but at some point
we only had the option to use pin names, or the corresponding pad or buffer
terminal names, or any
of the busses defined by the [Pin Mapping] keyword or the corresponding new
names I can’t remember now. The question was raised what if someone would like
to use Node 0
for it. That’s how “A_gnd” came into the picture. Then the question was raised
whether this A_gnd
is the same thing as those little triangle symbols in BIRD158’s Tx and Rx
drawings, and whether these
two BIRDs were “compatible” with each other. Related to that, the question was
A_gnd should be global or local, and if local whether it should be local to a
[Component] or [Model].
Along this conversation the concept of “Ground Based” modeling was also
Since the original problem was to figure out the syntax of the N+1th terminal
line for BIRD189, I wonder
what can we conclude from all this. Should we take Walter’s recommendation and
1. IBIS 189 should state that A_gnd is the Simulator Reference Node (Node 0
2. We add the following Note:
* Using A_gnd as Interconnect Model Terminals or Node 0 in IBIS-ISS
subckts may not account for all currents going to “ground” pins, and therefore
potentially introduce simulation results when doing power aware simulations
that are not ground based.
From: David Banas [mailto:capn.freako@xxxxxxxxx]
Sent: Thursday, March 22, 2018 12:49 PM
To: Walter Katz <wkatz@xxxxxxxxxx>
Cc: Muranyi, Arpad <Arpad_Muranyi@xxxxxxxxxx>; IBIS-Interconnect
Subject: Re: [ibis-macro] Re: [ibis-interconn] Re: Putting to Bed A_gnd
Thanks for your thoughtful comments in reply.
They helped me better understand the disparity between our two perspectives.
Regarding your presentation, I read it when you first sent it out.
And, when I got to slide 15, I followed up by rereading my copy of Brian’s book
(well, I mean, not the whole thing, but the section that pertains).
It all makes sense to me.
However, you don’t actually define the term, “Ground Based,” formally on slide
Rather, you allude to what it might mean, by referencing Scott’s comment, re:
“ground referenced” simulations.
Sorry to be so persnickety about this, but I really think a formal definition
of the term, “Ground Based,” would help.
A so called “ground based” simulation is one in which no component resides in
between any two nodes considered to be voltage reference nodes.
In suggesting this definition, I’m looking at Brian’s Fig. 5.21 and am noting
that I’m suggesting something more restrictive than what he outlines. For
instance, he allows an inductor between two reference nodes to be “correct”, as
long as the two reference nodes are different. (See his Fig. 5.21a.) And
because of this extra strictness, I think it’s important that we all be on the
same page, regarding the formal meaning of the term, “ground based.”