SPICE problems

Hi Andrew,

>Tom Warneke wrote:
>> I am perplexed why one would want to convert IBIS to Spice models.
>
>Andrew wrote:
>I disagree that IBIS is intended to serve digital designers who want
>to move away from SPICE. IBIS has everything to do with analog
>simulation, the same stuff we use SPICE for. IBIS models are not
>inherently different from SPICE models in terms of signal integrity,
>crosstalk, EMC, etc.

 I think you are missing Tom's point. Yes I agree IBIS simulators
and SPICE do the same thing functionally and yes I agree that
all SPICE packages should be able to run IBIS natively and certainly
could be made to.

 I believe Tom's point about digital engineers was related to ease of use.
IBIS defines many things automatically that SPICE models of the same
component do not. This is due mostly to lack of standards. There
is no 'SPICE' model to define the standard method of pin-outs, power
supplies, package parameters etc in SPICE. IBIS is the method of doing
that and much more. Whether it is used with SPICE as the engine or a
signal integrity tool which has an equal or better engine for SI analysis.

  There are other problems with SPICE that make it difficult for digital
engineers including DC and AC stability problems that are a part of most
if not ALL SPICE packages. As an avid SPICE user I bet you have run into
more than one IC models that you had a problem getting to stabilize or
you had trouble figuring out which polarity on the enable to use or had
trouble figuring out how to connect the package modeling to
the die model, or had trouble figuring out how to convert it from
HSPICE to PSPICE or PSPICE to Berkley SPICE etc.

  I separate digital engineers from analog engineers only in that they
will not use signal integrity tools if they have to put up with the amount
of work required to get a board full of SPICE models hooked up and running.

  An analog engineer often only models a small part of a board
associated with a one or two pins. He can afford to invest the time to
use SPICE without the high level gain of using IBIS models and the other
high level interface gains any good signal integrity tool offers.

  Don't get me wrong SPICE is a great tool for more constrained problem
sets. I have spent many hundreds of hours using it.

 Bottom line. I felt Tom's point was ease of use differences.

 
Received on Sun Oct 26 09:56:09 1997