RE: [SI-LIST] : SI analysis: IBIS vs. HSpice

About this list Date view Thread view Subject view Author view

From: Dagostino, Tom ([email protected])
Date: Fri Jan 19 2001 - 10:19:12 PST


How are you extracting the transmission line models when you do the HSPICE
simulations?

Tom Dagostino
IBIS and Tau Modeling Manager
SDD
Mentor Graphics Corp.
503-685-1613
[email protected]

-----Original Message-----
From: Christian Trudeau [mailto:[email protected]]
Sent: Friday, January 19, 2001 8:03 AM
To: [email protected]
Subject: [SI-LIST] : SI analysis: IBIS vs. HSpice

Todd,

I am currently in the throws of performing some pre-routed simulations on a
design I am working on. In the past I have stuck to running my analysis
using HSPICE. Typically this analysis has been limited to reflection SI
and not crosstalk or simultaneous switching. We work mostly with serial
data link verification so following the net from output driver through the
interconnect to the receiver is the topology of choice. We then evaluate
the data eye by simulating and measuring data jitter. HSPICE has been
working fairly well, the W model for transmission lines has improved in the
recent future and therefore improved our hardware to simulation
correlation.

We are currently trying to work with our Cadence EDA tool SpecctraQuest to
take a different approach to simulation. The tool has advertised the
ability to per from Crosstalk and SS analysis as well as Reflection. This
sounds great and we are trying to use it. The problem is that Cadence
ultimately uses their own proprietary model type (DML) which they derive
from IBIS models. The trouble is that our IO folks for the TX and RCVs
create SPICE models. So I must translate the models from SPICE to IBIS.
This has been fairly painful, the models for our components are fairly
complex. I have yet to get an accurate IBIS model for a given SPICE model.
I am fairly sure that with enough time we could get close but we don't have
the time. So for now we stick with net by net simulation with HSPICE. It
sure would be nice if Cadence imported the SPICE model directly.

Conclusion is that SPICE to IBIS conversion is painful and time consuming,
not even considering model accuracy.

When we eventually need to attack other types of simulation (SS and
Crosstalk) we are looking at very complex multinet HSPICE decks .... Oh no
:)

Thanks,

Christian Trudeau
IBM Corporation
ASIC Core Development Circuit Board Design Team
E-Mail : [email protected]
Phone : 802-769-5333

**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected]. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****

**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected]. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Tue May 08 2001 - 14:30:39 PDT