Microstrip is easy compared to multiple-pin connectors or
device packages. Can anyone demonstrate useful modelling
and simulation of crosstalk and ground bounce in a connector
with 40 single-ended switching lines and user-defined victim
line and signal/ground pin assignment, in which the connector
model is calculated by a solver rather than derived from
measurement? Let's say that "useful" means that the predicted
xtalk/gbounce are accurate to within 20% of measured values,
and that the model solver and the simulation together run in
less than 24 hours on the platform of choice.
Regards,
Richard Schumacher
I do not speak in any capacity for the Hewlett Packard Company.