Re: low budget IBIS tools

Hi,

I understand the need for SPICE and the value it can provide in offering
a more robust environment and control over modeling characteristics than
IBIS models and simulators support.

It seems that something will be lost in going from SPICE Model -> IBIS
-> SPICE model. My understanding is that (behavioral) IBIS models are
generally derived from (functional) SPICE models. At best, it seems,
that you will get a behavioral -- not functional -- SPICE model.

I suppose it makes sense if only an IBIS model is available. It seems,
however, that the SPICE run will be limited by the behavioral data that
can be extracted from the IBIS model.

IBIS models and simulators are fine tuned to be faster and easier to use
for their intended purpose. This imposes limitations which are not
realized in the SPICE world (i.e. support for voltage controlled
multi-stage drivers and the examples you provided). It seems to me that
using a SPICE simulator with models derived from IBIS is taking on the
worst of both worlds: SPICE complexity and run times and IBIS model
limitations.

Would it make sense to convert your SPICE model library to IBIS -- it
takes effort and care to translate models accurately -- to use in the
IBIS domain for those cases when you must use IBIS models because SPICE
models are not available? Then you get the relatively easy and fine
tuned IBIS environment which is better suited to the required IBIS
models, and you can use SPICE on those circuits for which you have all
the SPICE models available.

Further feedback is welcome...
Regards,
Tom

Dan Aleksandrowicz wrote:
>
> Hi Tom.
>
> You are correct in your assumption that IBIS users, as much as IC vendors,
> wanted to move from spice simulation/models to IBIS simulations/models.
>
> Yet, there will always be that digital circuit that IBIS does not apply.
> This might be in today's in serial IC's or even an active termination.
> Another, most common possibility would be in the initial stage of a project
> were technologies and concepts are tested for signal integrity.
>
> So we try to find a combined solution which enable us to
> use IBIS models and to some spice simulation for a wide range
> of scenarios.
>
> And a pure routed PCB simulation is good as a verification procedure.
> Solving problem at that stage is usually very problematic.
>
> Dan Aleksandrowicz.

-- 
Tom Warneke
CAD Application Engineer
Cisco Systems
250 Apollo Drive
Chelmsford, MA 01824
email: twarneke@cisco.com
Phone: 508-244-8214
  FAX: 508-244-8347