Re: [SI-LIST] : About HSPICE model

D. C. Sessions ([email protected])
Mon, 05 Apr 1999 17:28:39 -0700

Kim Helliwell wrote:
> D. C. Sessions wrote:
> >
> > chenlanbing wrote:
> > >
> > > HI,
> > > I'm working on SI.I like to simulate the transmiter line by HSPICE.Now I need the hspice I/O buffer model,such as AMCC 2064.Who can help me?
> >
> > The reason that IBIS was developed in the first place is that
> > semiconductor companies like (to pick one at random) VLSI
> > Technology keep their process models and to a lesser extent
> > their circuit architectures very confidential.
> >
> > IOW, don't get your hopes up.
> >
> On the other hand, I've had reasonable luck getting
> such models from vendors under an NDA agreement. The
> ability to do this probably depends on your company's
> relationship with the vendor; if you buy a lot of
> their parts, they'll be more willing to accommodate
> you.

One thing for sure: it's spotty. A contact at IBM gets
absolutely EVERYTHING and was amazed that we had a hard
time getting IBIS models; another contact at a small
company was impressed at our ability to get anything at all.

> My feeling is: a vendor should either give you an
> IBIS model or a SPICE model (subject to an NDA). If
> they are unwilling to do either, you have to wonder
> how they expect you to design their parts in.

Violent agreement.

> By the way: If you *DO* have an IBIS model, the latest
> version of HSPICE will read it directly, so you don't
> have to have an HSPICE model to use HSPICE in that
> event. You should see the manual for the new version
> for details. I make no representations as to accuracy,
> ease of use, etc. I've used it, and I think it's
> worth a look, anyway.

Assuming you have access to 98.4, it's reported to be
massively bug-infested. Many of us *don't* have access to
98.4, though -- thanks to Avant's brain-dead support
groutp, it's all we can do to get new license files for
the dozen or so copies we have before they expire.
Asking for license files for upgrades is beyond reason.

D. C. Sessions
[email protected]

**** To unsubscribe from si-list: send e-mail to [email protected] In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at ****