Re[2]: EGG #2 -- controlled rise-time devices

---------------------------- Forwarded with Changes ---------------------------
From: Arpad Muranyi at JFCCM3
Date: 4/8/94 3:40PM
*To: ibis@vhdl.org at Internet_Gateway
*To: cpk@Cadence.COM at Internet_Gateway
Subject: Re[2]: EGG #2 -- controlled rise-time devices
-------------------------------------------------------------------------------
I am not sure if you got this, so I am trying again. (It came back to me)

Kumar,

I agree with what you are saying, but that actually means that we want to
describe a surface point by point. This is the best way, but would require a
huge data table for each structure. For example, if you allow 100 data points
for each independent axis (V, t) you end up with 100^2 data points, which you
have to multiply by three, since we have three numbers describing each point. I
don't think we can handle that yet. The proposed method would at least give us
a good start.

Arpad

The important issue for the user is how many v-t curves to generate
and for how
many loads. In principal the number of v-t curves should be unrestricted.

The v-t data format should
specify whether the data is for rising, falling and
what load it was generated for.

I feel uncomfortable with the modelling strategy which says that the behavior
has to be scaled from dc which it may not have any relation to. In
a ideal case
I would have preferred i-v data at each time point.

-kumar
Received on Fri Apr 8 17:20:59 1994