As a potential future user of IBIS models, my greatest concern is
that commercial models will not have been confirmed by measurement.
What ever modelling procedure you use, please confirm the accuracy
of the model by building and measuring the behaviour of actual
circuits using representative devices mounted on representative PCBs.
This is especially important for devices with multiple pins
(connectors and large packages) where the crosstalk and ground
bounce caused by multiple simultaneously-switching lines is often
quite different than one would predict from single-line results.
The signal-ground pin pattern used in connectors is of course
particularly significant. Even all-differential applications may
have surprises for unwary modellers and users.
------------------------------------------------------------------
Richard Schumacher
I do not speak in any capacity for the Convex Division
of the Hewlett Packard Company.