[SI-LIST] Re: IBIS modeling
- From: "Jon Powell" <jpowell@xxxxxxxxxxxx>
- To: "'Scott McMorrow'" <scott@xxxxxxxxxxxxxxxx>
- Date: Fri, 12 Oct 2001 09:20:56 -0700
Just to add a little more food to the frenzy.
In complete agreement with Scott + this data:
LVDS is defined as being implemented with an "Ideal" current source.
Since this is "a little hard" what I have seen used is essentially just
a big capacitor to VCC and VSS. This gives a pretty good transient current
source
but ONLY at a given operating frequency. Which means (as stated by Scot)
that you
can only make good models for a given operating frequency and you have to be
careful
to try and get one that will give accurate AC and DC results.
jon
> -----Original Message-----
> From: Scott McMorrow [mailto:scott@xxxxxxxxxxxxxxxx]
> Sent: Thursday, October 11, 2001 4:45 PM
> To: jpowell@xxxxxxxxxxxx
> Cc: chris@xxxxxxxxxx; khalida@xxxxxxxxxxx; si-list@xxxxxxxxxxxxx
> Subject: Re: [SI-LIST] Re: IBIS modeling
>
>
> It turns out that it is possible to create IBIS models for
> differential
> drivers with quite good correlated results to SPICE differential
> simulations. We've created quite a few using the following
> techniques:
>
> 1) The SPICE circuit used for extraction of the buffer I/V and V/T
> curves must operate in a balanced mode at the correct operating
> point.
>
> 2) Instantaneous common mode offset in the SPICE simulations must
> be eliminated from the simulations prior to extraction. There are a
> few methods available to accomplish this.
>
> 3) V/T extraction must be performed at the normal operating conditions
> for the device, and be consistent with the operating conditions
> of the I/V curve extractions. If this is not possible, then
> the I/V curves must be
> extracted from a transient simulation at the nominal operating point,
> rather than a DC simulation.
>
> 4) The IBIS model is valid only over a small set of operating
> conditions
> near the common mode operating point, because the IBIS model
> is inherently
> single ended by construction. For multiple operating points,
> such as might
> be seen in BLVDS applications or with non-100ohm system
> termination, multiple
> IBIS model extractions would be necessary.
>
> LVDS drivers are rather complex creatures and are generally extremely
> "twitchy" to work with in SPICE. This tends to complicate
> the single ended
> IBIS model extraction. In general, there is no one way to
> create an IBIS
> model from a differential SPICE model. However, there are
> quite a few tricks
> in the bag that can be used to create useful and highly
> accurate models.
>
> best regards,
>
> scott
>
>
> --
> Scott McMorrow
> Principal Engineer
> SiQual, Signal Quality Engineering
> 18735 SW Boones Ferry Road
> Tualatin, OR 97062-3090
> (503) 885-1231
> http://www.siqual.com
>
>
>
> Jon Powell wrote:
>
> > I have done some LVDS models and have NOT had any (good) luck
> > in simulating them as single ended buffers.
> > Let me re-phrase that. Most of the SPICE models that I have run do
> > not give correct (or self-correlating) results if you don't put the
> > series current return resistor into the LVDS circuit. If you try to
> > simulate this resistor with a resistor to some constant
> node you also
> > get bogus results. In order to get correct SPICE results you have to
> > simulate
> > both drivers at the same time with the correct termination.
> >
> > I have seen this result in 3 independent LVDS drivers from different
> > vendors.
> >
> > For an example LVDS IBIS model, you might check out the
> National or Altera
> > web pages.
> >
> > jon
> >
> > > -----Original Message-----
> > > From: si-list-bounce@xxxxxxxxxxxxx
> > > [mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of Chris Brewster
> > > Sent: Thursday, October 11, 2001 10:32 AM
> > > To: khalida@xxxxxxxxxxx
> > > Cc: si-list@xxxxxxxxxxxxx
> > > Subject: [SI-LIST] Re: IBIS modeling
> > >
> > >
> > > Khalid,
> > >
> > > Most of the information describes the process for extraction
> > > of a Single
> > > Ended buffer. With LVDS, you have to treat each output as if
> > > it were a
> > > single ended buffer and then combine them to create a
> > > differential model.
> > >
> > > The literature from the IBIS summit by Hegazy is a good place
> > > to start.
> > >
> > > http://www.eda.org/pub/ibis/summits/jan01/hegazy.zip
> > >
> > > regards,
> > >
> > > -Chris
> > >
> > >
> > > At 09:01 AM 10/11/2001 -0700, Khalid Ansari wrote:
> > >
> > > >Hi all,
> > > >
> > > >I am looking to create IBIS models for my LVDS I/Os. I already
> > > >have Hspice models created. Is there any good literature on how
> > > >to create IBIS models out of Hspice models. I haven't
> had a chance
> > > >to work with IBIS in the past so any introductory
> literature will be
> > > >helpful. I guess the IBIS version I will be using is 3.2
> > > with scripts
> > > >created by North Carolina State University.
> > > >
> > > >Thanks in advance,
> > > >Khalid
> > > >
> > >
> >------------------------------------------------------------------
> > > >To unsubscribe from si-list:
> > > >si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the
> Subject field
> > > >
> > > >or to administer your membership from a web page, go to:
> > > >//www.freelists.org/webpage/si-list
> > > >
> > > >For help:
> > > >si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
> > > >
> > > >List archives are viewable at:
> > > > //www.freelists.org/archives/si-list
> > > >or at our remote archives:
> > > > http://groups.yahoo.com/group/si-list/messages
> > > >Old (prior to June 6, 2001) list archives are viewable at:
> > > > http://www.qsl.net/wb6tpu
> > > >
> > >
> > >
> > > ------------------------------------------------------------------
> > > To unsubscribe from si-list:
> > > si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the
> Subject field
> > >
> > > or to administer your membership from a web page, go to:
> > > //www.freelists.org/webpage/si-list
> > >
> > > For help:
> > > si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
> > >
> > > List archives are viewable at:
> > > //www.freelists.org/archives/si-list
> > > or at our remote archives:
> > > http://groups.yahoo.com/group/si-list/messages
> > > Old (prior to June 6, 2001) list archives are viewable at:
> > > http://www.qsl.net/wb6tpu
> > >
> > >
> > >
> >
> > ------------------------------------------------------------------
> > To unsubscribe from si-list:
> > si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the
> Subject field
> >
> > or to administer your membership from a web page, go to:
> > //www.freelists.org/webpage/si-list
> >
> > For help:
> > si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
> >
> > List archives are viewable at:
> > //www.freelists.org/archives/si-list
> > or at our remote archives:
> > http://groups.yahoo.com/group/si-list/messages
> > Old (prior to June 6, 2001) list archives are viewable at:
> > http://www.qsl.net/wb6tpu
> >
>
>
>
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
