Re(x) T_skew

Text item: Text_1

Hi IBIS folks,

Regarding 20-80%, as far as I remember, we (my group at Intel) started
to use 20-80% (and not 10-90%) in IBIS measurements simply because it is
not always easy to obtain a clean slope between 10-90% due to the
parasitics and noise.

Since the ramp in my original implementation of the behavioral HSPICE
model is a perfect trapezoidal representing an idealized internal node,
I felt that I should get the straightest part of the slope from a
measurement. This straight line ramp is then rounded by the packaging
and C_comp parasitics.

Arpad
Intel Corporation
Received on Wed Feb 2 08:58:16 1994