Re: package and connector model questions

From: Stephen Peters <sjpeters@ichips.intel.com>
Date: Fri Sep 24 1999 - 13:21:52 PDT

Hello Kellie, Aurbey, all:

> At 03:59 PM 9/23/99 -0500, Aubrey_Sparkman@Dell.com wrote:
> >The first question has to do with Kellee's outstanding item below. I.E.,
> >why not use the RLC matrix description in section 7 on package modeling? I
> >am not sure why Connector modeling isn't (or couldn't be) a subset of the
> >package modeling spec. After all, a package is a connector that connects
> >the silicon to a PCB....or socket, which is again another connector. And
> >before someone comments "un-mated connector" I also wonder why I need the
> >model for an unmated connector or a package out of the socket, etc.
>

  As one of the authors of the .pkg extensions, I'd like to add a comment. It is tempting to think of an IC package as 'just another type of connector', but in reality an IC package is much closer to a circuit board than a connector. In fact, for many of the high performance products I've seen, the package IS a circuit board, complete with edge fingers or pins to connect it to a mother board. The problem is that when trying to create an electrical description of this type of package you're not dealing with a group of parallel pins of equal length. You've got *traces* that have different lengths, diverge, switch layers, who knows what else. To model the coupling requires having many small matrixes that are named or somehow interrelated. After we played with it for a while we realized that 1) Even if we could agree on a description no one is going to create such a data base by hand and 2) it's a problem the CAE companies have already solved by integrating PCB extraction with!
!
 their SI tools. Thus, we settled on what we could achieve -- a simpler uncoupled description.

  As Kellie mentioned, I do expect that some of the extensions the connector group has come up with will find there way into the .pkg section. But I suspect other approach will be needed before IBIS will support full coupled
package models.

   Regards,
   Stephen Peters
   Intel Corp.
Received on Fri Sep 24 13:22:03 1999

This archive was generated by hypermail 2.1.8 : Fri Jun 03 2011 - 09:53:46 PDT