Re: Technology Models vs. full IBIS models

Bob Ross wrote:
>
> Hello:
>
> In my opinion your best approach is to develop/obtain the actual
> IBIS models. Real buffers may differ significantly from generic
> technology model buffers. In some cases the differences occur
> in parts or buffers from the same family of a manufacturer.
>
> Regarding benchmarking major tools, this is a topic for the
> Signal Integrity Reflector. The IBIS reflector is not used
> for Vendor Tool comparisons or business issues by EIA rules
> and by the fact that the EIA IBIS Open Forum is funded by
> many vendors.
>
> Best Regards,
> Bob Ross
> Mentor Graphics
>
> > eh wrote:
> >
> > Dear friends,
> >
> > We are evaluating a tool regarding signal integrity.
> > Our design is composed of modern devices ( DSP,DRAM,SRAM,FLASH,FCT etc).
> > Since we don't have the IBIS model for most of the parts, the supplier
> > suggested that I use technology models. I have a DSP, so I assigned it as
> > FCT, for it was the closest choise.
> > The PCB crosstalk simulation showed problems we NEVER faced with five
> > practical PCBs.
> > I assume that I have to get/order/develop IBIS models for all the parts, and
> > not
> > use "technology models".
> > Is that true ?
> > Has anyone went through these ?
> > Has anyone a benchmark, comparing the major tools , offered this field of
> > signal integrity (crosstalk...).
> >
> > Regards,
> > Simi Eshkol
> > Sn Elec Eng.
> > VideoCom , Israel
> >
> >
> >
> >
I'm not sure what is meant by the technology models. I presume this
means the SPICE transistor level models based on the process. If that
is the case then obtaining the IBIS models of the same may not help
since IBIS models are mostly made from the SPICE models anyway. However
obtaining the correct models for the part number from the vendor may
get you more realistic results, this is independent of whether the model
is in SPICE or IBIS format.

Best Regards,

-- 
Fred Balistreri
fred@apsimtech.com
http://www.apsimtech.com