From: Hassan Ali (firstname.lastname@example.org)
Date: Fri May 04 2001 - 08:20:12 PDT
Mark Alexander wrote:
> Many of us are familiar with the fundamental differences
> between SPICE and
> -- one is difficult and highly accurate, the other is simple
> and moderately
> accurate. What might be more useful in this discussion is
> information from
> personal design experience about when you would use one
> versus the other.
Mark, I'm afraid your statements are pretty general. You can indeed obtain
from device vendors *inaccurate* SPICE and IBIS models. For example they can
give you an "old" SPICE model that doesn't match the features of their final
design. This is particularly true these days when most of us wish to have in
our designs the latest and greatest devices which are still in development.
You can also obtain inaccurate SPICE models from unscrupulous modeling guys
from some device vendors. I have seen some SPICE models with inaccurate core
and pad power supply voltages. I have seen SPICE models with GLOBAL
statements that affect the rest of the circuit (while they shouldn't). I
have seen SPICE models that do not match datasheet numbers e.g. for Vil,
Vih. In short, do not ever think that you always get *highly accurate* SPICE
models from device vendors. You always have to do your homework and make
sure the model gives *reasonable* results. And the word *reasonable* is the
key here because the hardware performance will not necessarily be the same
as that provided by the SPICE model.
The same can be said for IBIS models as in 99.9% of the time these days,
IBIS models are generated from SPICE models. Garbage in, garbage out!
However, it is not generally true that IBIS models will give you results
*inferior* to the results obtained using SPICE models. But it is certainly
true that IBIS models are more prone to inaccuracies than SPICE models. You
have to do an *even harder* homework before you decide to use a particular
IBIS model. In one IBIS-related web pages (which I don't remember its URL -
p'se somebody help!) I once saw a checklist to go through to ascertain the
accuracy of an IBIS model. The first thing I do with an IBIS model is to
check if it matches the data sheet numbers.
But there is one more *scary* thing about IBIS models which I learnt
recently: Implementation of IBIS models is *different* in different circuit
simulation tools! I recently had an IBIS model which I created from a SPICE
model, and when I used it in HSPICE and Cadence's SigXplorer I obtain very
different results! One of them matched the expected results. But I must
admit that this was just a single case. Most IBIS models I have used gave
same results in both HSPICE and SigXplorer.
The bottomline is that you should *never* use device models blindly. Do your
**** To unsubscribe from si-list or si-list-digest: send e-mail to
email@example.com. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
This archive was generated by hypermail 2b29 : Thu Jun 21 2001 - 10:11:50 PDT