Ramp Times (kellee's two cents)

From: Kellee Crisafulli, HyperLynx
Re: Ramp Notes:

I agree with Jay Diepenbrock completely. The ramp time should not be
converted directly to a linear slew except for the region from 20% to 80%.

If you need a method of creating the output curves, a suggested spice model
for IBIS created by Arpad at Intel is useful for a starting point on a new
simulator. Perhaps we should consider adding this to the specification as
a reference point (An appendix perhaps)?

As to why 20% to 80% was chosen instead of 10% and 90%, I don't remember
the specifics, however I believe that only 2 or 3 semiconductor house people
were available during the initial release to contribute to this effort.
They were a major source for these percentages. Apparently some of the
semiconductor people do use 20% to 80%. Hopefully this won't be a major
problem.

If it turns out to be a problem we could consider extending the measurement
points linearly to 10% and 90%. (Are you interested in another BIRD on this Bob?)

Have a great day IBISians!
Kellee
Received on Wed Feb 2 08:01:07 1994