Programmable buffers

From: Arpad Muranyi <Arpad_Muranyi@ccm.fm.intel.com>
Date: Fri Jul 14 1995 - 14:38:00 PDT

Hi IBIS folks,

I am writing this EMAIL to generate a discussion and a possible future
EGG and BIRD on how to deal with programmable buffers in IBIS.

Problem:

There are buffers out there which have programmable strength, in other
words, buffers which have not just a single set of I-V and edge rate
description per pullup or pulldown structure. Currently there are no
provisions in the IBIS specification for such buffers.

The work around for such buffers is to create independent buffer models
for the various buffer modes and include them as separate models in the
IBIS file. In the notes section, one can make a comment about the use
of these additional models for the programmable buffers.

This method is possible in IBIS, since it is not illegal to have more
buffers in an IBIS file than what the pin list section calls for. The
area where IBIS is lacking is that there are no nice ways defined to
help the user to switch between the various modes of a buffer.

Suggestions:

One thing that came to my mind was to allow the listing of several
buffers for one pin name, something like this (forgive me the bogus pin
and buffer names):

1 reset 74ALS244
1 reset 74F244
1 reset 74HCT244 | or without repeating the pin and signal names:
2 AD0 74ALS241
              74F241
              74HCT241
3 AD1 74ALS241 ..... and so on.

The simulation tool could then give the user the option to select
between the buffers when simulating with that pin.

One problem with this method is that the pin list can grow fairly large
if there are lot of programmable pins in it.

Another idea would be to allow buffer models to contain more than just
one set of I-V and edge rate data section. If this method is used, we
might need to invent another keyword, or some mechanism to be able to
select between these sections. This method would be more compact (at
least in terms of the pin list) but might be more complicated on the
buffer model level.

This is what I came up with so far. Does anyone care to add comments,
ideas to this conversation starter?

Sincerely,

Arpad Muranyi
Intel Corporation
Received on Fri Jul 14 14:51:15 1995

This archive was generated by hypermail 2.1.8 : Fri Jun 03 2011 - 09:52:28 PDT