Re[2]: connector models

     Greetings All...
     
     Just for clarification...
     
     The connector companies main issue with IBIS is the lack of coupled
     line modeling.
     
     The connector companies suggested the use of a Maxwell matrix for the
     connector models a couple of years ago. I think a partial option has
     been installed in the V.3.0 IBIS specification.
     
     As far as "first pass".... I beleive that IBIS in V.2.0 can support
     series L, parallel C, and series R. This is the simplest solution
     with the most probablility of inaccurate results.
     
     
     ~~~~~~~
     Now I think the other significant problem is going to become
     apparent... The fact the IBIS simulators have different methods of
     "circuit solving". When a connector model is provided in matrix
     format, the results will depend on the simulators solution algorithms
     (not an issue with Berkeley SPICE).
     
     I am concerned that connector models will perform differently on
     different simulators. Is this a valid concern??
     ~~~~~~~~
     
     
     A part of the problem might be... (besides the above concern)
     
     * Semiconductor manufactures need a way to provide models that do not
     provide insighth to proprietary fabrication methods. Connector
     companies already have this... SPICE.
     
     * Some SPICE simulators can accept some versions of IBIS models some
     can not support IBIS at all. Are there any IBIS simulators that
     support SPICE? In this case, it all depends on were the "energy" is
     assigned. Most connector companies do not sell simulation software.
     As such connector companies might not be the best source for driving
     this issue.
     
     
     
     Gus Panella
     Molex, Incorporated.

______________________________ Reply Separator _________________________________
Subject: Re: connector models
Author: Kellee Crisafulli <kellee@hyperlynx.com> at INTERNET
Date: 97/12/12 11:08 AM

Hi John,
     
At 09:28 AM 12/12/97 +0100, you wrote:
>Are the following statements reasonably truthful?
> - Many attempts have been made to adapt IBIS to
> the creation of connector and cable models. No
> agreed synthax has been found, due mainly to the
> complex nature of coupling.
The problem is just the lack of energy.
We were very close to agreement at last years
Design Super Con. The major problem I saw is that the participants
from the connector companies have not provide enough energy to
get resolution. If one or more connector companies put in even a
small fraction of the energy that an Intel or National Semi puts into
IBIS this would have been completed long ago.
     
> - Limited connector and cable models are possible with
> IBIS 3.0, using the expanded package model and the new
> series element.
yes
     
> - There is no medium-term prospect of a better IBIS synthax
I would agree there is no short term prospect. Medium term is
always possible just needs energy from a few good people.
Long term is highly probable
     
> - Most simulator companies have developed their own
> connector model synthax.
Or they use one of the formats developed by a connector company. However
there is alot to be desired in any format I am currently aware of. This is
one thing that has made it so difficult to accomplish an IBIS standard for
connector
modeling. It is easy to agreement on a simple standard for connector models.
It was
difficult to get acceptance of that being good enough for a first pass. So
rather than
go forward with something basic we got nothing!
     
Summary:
We let those that must have everything or some special interest prevent the
creation
of simple starting point. This could have been overcome with a small push from
one or two people after last years Design Super Con. Anyone with some energy to
provide could make the difference on this one. It will take a few months to
revive it
though.
     
     
     
     
-------------------------------------------------------------------------
Have a great day...
Kellee Crisafulli at HyperLynx
kellee@hyperlynx.com http://www.hyperlynx.com
-------------------------------------------------------------------------
 
Received on Mon Dec 15 04:49:39 1997