Re: Question - BIRD 36.2 - Electrical Board Description

From: <MLeonard@bangate.compaq.com>
Date: Sat Mar 22 1997 - 11:46:16 PST

Dileep,

We have a pressing need for board-level descriptions of memory modules and
AGP cards. For the most part, board-level crosstalk isn't too much of a
concern there. Accordingly I'm very willing to forgo coupling information
if we can get this BIRD approved and out in the hands of these module and
card vendors. This BIRD, as well as a couple others in work, are very
important. Unfortunately the approval process has become a long, drawn out
affair. The modeling capabilities of IBIS are falling further and further
behind our simulation needs.

I would, however, welcome coupling information on Intel's processor
daughter cards. Their board descriptions are currently being provided in a
non-IBIS format. The descriptions do not include coupling information.
Would we get it with an expanded BIRD 36.2? Probably not.

My preference is to get this BIRD approved, then worry about board
descriptions with coupling information later.

Mark Leonard
Compaq Computer Corporation
-------------
Original Text
From: dileep@contec.contec.COM (Dileep Divekar), on 3/21/97 1:27 PM:
To: <ibis-users@vhdl.org>
Cc: <dileep@contec.contec.COM>

I am trying to get an understanding of how many users and
designers consider doing only delay analysis adequate
enough for their designs. If they have to use some other
interface for crosstalk analysis, then they might as well
use that interface for delay analysis also, instead of
switching back and forth between two different interfaces.
Therefore, they will use BIRD 36.2, if they think
that only delay analysis is sufficeint and there is no
need to perform crosstalk analysis.
Dileep Divekar

> From owner-ibis@vhdl.vhdl.org Fri Mar 21 09:11:03 1997
> To: ibis-users@vhdl.org
> Subject: Question - BIRD 36.2 - Electrical Board Description
> Content-Length: 762
> X-Lines: 23
>
>
> Hello Dileep:
>
> Yes, the present description -- by design-- does not
> allow for describing coupling bettween traces. If a user
> does wish to give enough information about a board in
> order to do coupling the best medium in a physical description
> (PCB data base). This takes advantage of the vendors already
> existing path into there products, and there are now hooks
> into EDIF 4.0 that will call out IBIS files.
>
> Regards,
> Stephen
>
>
>
> Hello,
> The present version of BIRD 36.2 allows the description
> of only un-coupled lines. I would like to know if the
> users and designers think that this description is adequate
> for them. It will not be possible to use the IBIS models
> to analyze crosstalk due to interconnects.
> Dileep Divekar
>
 
Received on Sat Mar 22 12:35:20 1997

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