RE: Questions

From: Ingraham, Andrew <Andrew.Ingraham@compaq.com>
Date: Fri Sep 08 2000 - 08:08:42 PDT

>Are you developing a version2.1 TO version 1.1 ?
>If not, how I have to proceed to make the converter work with a Ibis model
>version 2.1?

I don't think you need to do that. It should work with version 2.1 models
just fine.

>-Could you give me the instructions on how to properly convert IBIS models
>for accurate signal integrity simulation.

Most IBIS models came from SPICE, so you are better off getting the SPICE
models if you can negotiate it. When you go through a SPICE -> IBIS ->
SPICE process, you will lose accuracy. Best to stay as close to the source
as possible.

If an IBIS model is all you can get, your results will depend on what
features they put in the IBIS model and on how accurately they made it, and
what your simulator does with all parts of the model. For example, the
IBIS2SPICE converter ignores V-T curves, and uses only the simplistic R-L-C
package model, the results of which are questionable at higher ramp rates.
You might consider it a "challenge", to augment the IBIS model with (say)
other measurements you have done on real components, or on your experience.

Regards,
Andy

 
Received on Fri Sep 8 08:11:34 2000

This archive was generated by hypermail 2.1.8 : Fri Jun 03 2011 - 09:53:47 PDT