Re: Correlating IBIS models with actual Silicon measurements

Amlan,

I have done a little work in this area and though I don't yet have a
definitive answer I can give you some data:

1) If you make you measurements in a packaged component and you know
what the parasitics of the packaged component are, you should be able to
back the parasitics out of the model. Unfortunately, this is probably
not easy.

2) I have made several models by doing the measurement and ignoring the
fact that I couldn't back the parasitics out. The QA to the actual
measurements was not perfect but it was pretty good and a lot better
that not having a model.

3) If you have SPICE models (or other internal part characterization
models), presumably these were made by correlation to bare die
measurements and as such would be good to use for creating models. Also,
if you have the raw die data it may contain the information you need for
an IBIS model.

regards,
Jon Powell
Senior Scientist, Viewlogic Consulting Services.
Received on Mon Apr 27 09:07:22 1998