Re: BIRD24 COMMENT

Text item:

Bob, and others,

On reflection, I tend to agree with Bob. In a pure simulation environment,
removing C_Comp is not much of a problem, but for those doing IBIS models by
measurement, it it exceedingly difficult. I also agree with Bob's comment about
the sanctity of the reference wave form. Maybe what is needed is an additional
parameter that says whether or not C_Comp was used. E.g. a boolean:

 C_Comp_in = 1 or 0 (or T or F, etc.)

so the simulator correctly interprets the data. If this parameter were not
present, we could have a default vale, tbd by the forum.

Will

To Committee:

I only agree that there needs to be clarity in IBIS, but I do not support
that C_comp = 0 is the mandated test condition for RAMP. I could accept
that, but I definitely do not support such a mandate for Waveform Data.
My concerns are these:

(1) Removing the "effective" C_comp which is USUALLY not obvious cannot
and should not be done to any Spice die model. That is a cumbersome
process which can introduce unanticipated artifacts and other error. It
is also one additional burden on top of a number of complicated processes
such as those necessary for encoding Spice models for s2ibis conversions.

(2) Responses based on measurements will always contain the effective
"C_comp", and IBIS processes should accomodate that situation. So the
problem is simple - either data based with C_comp included has to be
"transformed" to no-C_comp data, or visa versa. Since C_comp cannot
be removed for most cases, inclusion should be the default situation.

(3) While I am not able to be specific at this time, I am convinced
that the existence of C_comp in the extraction poses no insurmountable
technical problem. I believe proper de-embedding techniques can be
used, if they are necessary.

(4) The Waveform Specification contains the purest form of untampered
characterization data which is at the Pin if necessary, or at the Die if
the packaging effects can be removed through simulation or die measurement.
Under absolutely no condition should this (set of) reference waveforms
be corrupted with tampering assumptions such as setting the C_comp to 0.
While there may perhaps be inconveniences now, it is the wrong decision
for future integrity of IBIS to conclude that such extractions or measurements
which include C_comp can not be accurately dealt with in the future.

Bob Ross,
Interconnectix, Inc.

Text item: External Message Header

The following mail header is for administrative use
and may be ignored unless there are problems.

***IF THERE ARE PROBLEMS SAVE THESE HEADERS***.

Subject: BIRD24 COMMENT
To: ibis@vhdl.org
From: bob@icx.com ( Bob Ross)
Date: Tue, 6 Dec 94 19:20 PST
Message-Id: <m0rFCvG-000FV3C@icx.com>
Received: by icx.com (/\==/\ Smail3.1.28.1 #28.14)
        id <m0rFCvG-000FV3C@icx.com>; Tue, 6 Dec 94 19:20 PST
Received: from icx.com (icx.rain.com) by vhdl.vhdl.org (4.1/SMI-4.1/BARRNet)
        id AA07628; Tue, 6 Dec 94 19:38:04 PST
Received: from vhdl.vhdl.org by hermes.intel.com (5.65/10.0i); Tue, 6 Dec 94 19:
Received: from hermes.intel.com by ichips.intel.com (5.64+/10.0i); Tue, 6 Dec 94
Received: from ichips.intel.com by relay.jf.intel.com with smtp
        (Smail3.1.28.1 #2) id m0rFDYG-000tx9C; Tue, 6 Dec 94 20:00 PST
Received on Wed Dec 7 09:51:07 1994