FPGA Central

All,

On die S-Parameter is just one part of the problem. Models requiring on die
S-Parameter are incompatible with the rest of IBIS IV and VT data, including
things liked Vdiff, Cdiff.

The problem with [External Circuit] it that it is explicitly outside of a
model definition and therefore not model based. So if a chip had 300
instances of the same SerDes differential model, then using [External
Circuit] would require 300 instances of the same [External Circuit], this
combined with IBIS inability to properly handle true differentials makes
putting this capability in IBIS problematic.

As soon as the current revision of the AMI specification is completed,
SiSoft will be introducing new AMI parameters that satisfy the Analog
modeling requirements to support the AMI algorithmic modeling. One
additional compelling reason to put this into the .ami file is that we have
found that configuring the ami algorithmic model also configures the analog
parasitics.

Walter

Walter Katz
303.449-2308
Mobile 720.333-1107
wkatz (AT) sisoft (DOT) com
www.sisoft.com

-----Original Message-----
From: si-list-bounce (AT) freelists (DOT) org [mailto:si-list-bounce (AT) freelists (DOT) org]On
Behalf Of Muranyi, Arpad
Sent: Tuesday, February 09, 2010 10:33 AM
To: si-list (AT) freelists (DOT) org
Subject: [SI-LIST] Re: IBIS-AMI Vendor Support Help

Ken,

[External Model] may not work because it is specifically designed to
replace the [Model] with a better modeling language content while
preserving the same connectivity of the [Model] keyword. This means
that the all of the nodes of [Model] or [External Model] are supposed
to be connected the same exact way, which would prevent us to put die
parasitics between the buffer model and the pads.

However, [External Circuit] does not have this restriction, so we
could potentially make use of that keyword for this purpose, perhaps
referencing the new (proposed) IBIS-ISS capabilities which allow
us to use S-parameters. If we went with this route, the question
is, can we have a [Model] behind the [External Circuit], which is
something I don't remember for sure now. I need to research the
spec on that a little more...

Arpad
================================================== ======================

-----Original Message-----
From: Ken Willis [mailto:kwillis (AT) sigrity (DOT) com]
Sent: Tuesday, February 09, 2010 9:19 AM
To: 'Todd Westerhoff'; Muranyi, Arpad
Cc: si-list (AT) freelists (DOT) org
Subject: RE: [SI-LIST] Re: IBIS-AMI Vendor Support Help

Hi Todd,

On #1, it makes sense to incorporate S-parameters into IO models,
especially
since some major IP suppliers have this data today. I think the item
that
brings some uncertainty is why this would be pushed into the algorithmic
(AMI) portion of the model, especially if new syntax/API has to be
invented
and supported. The intent of AMI was to capture the adaptive filtering
behavior of these advanced Serdes devices, which you couldn't really
model
well in a circuit model. If the S-parameters are largely parasitic data,
it
seems more straightforward to include them in the analog circuit part of
the
model.

I understand that there is no standard "IBIS" syntax for S-params right
now,
but I would agree with Arpad that Touchstone is widely supported. Maybe
we
should be looking at [External Model] > Touchstone or something like
that.
But we can continue the discussion in the IBIS forum.

Thanks,

Ken Willis
Sigrity, Inc.
860-871-7070
kwillis (AT) sigrity (DOT) com
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request (AT) freelists (DOT) org with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request (AT) freelists (DOT) org with 'help' in the Subject field


List technical documents are available at:
http://www.si-list.net

List archives are viewable at:
http://www.freelists.org/archives/si-list

Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu


------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request (AT) freelists (DOT) org with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request (AT) freelists (DOT) org with 'help' in the Subject field


List technical documents are available at:
http://www.si-list.net

List archives are viewable at:
http://www.freelists.org/archives/si-list

Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu

See below:

Walter Katz wrote:
> All,
>
> On die S-Parameter is just one part of the problem. Models requiring on die
> S-Parameter are incompatible with the rest of IBIS IV and VT data, including
> things liked Vdiff, Cdiff.
>
I really hate statements without proof. Certainly "standard" IBIS
models with IV and VT tables could benefit from a frequency dependent
parasitic model.
> The problem with [External Circuit] it that it is explicitly outside of a
> model definition and therefore not model based. So if a chip had 300
> instances of the same SerDes differential model, then using [External
> Circuit] would require 300 instances of the same [External Circuit], this
> combined with IBIS inability to properly handle true differentials makes
> putting this capability in IBIS problematic.
>
There are several issues with this statement. Certainly an External
Circuit is model based. It is just not encapsulated in the same model
you are discussing. The use of 300 instances of the same External
Circuit is a red herring. This is just a matter of simulation model
management. By the same definition, 300 instances of the same SerDes
model are required. Finally, IBIS-AMI does not handle true
differential. All AMI algorithmic modeling has been decomposed to
differential mode only. This excludes any technique that might operate
on the two paths independently.
> As soon as the current revision of the AMI specification is completed,
> SiSoft will be introducing new AMI parameters that satisfy the Analog
> modeling requirements to support the AMI algorithmic modeling. One
> additional compelling reason to put this into the .ami file is that we have
> found that configuring the ami algorithmic model also configures the analog
> parasitics.
>
Why not introduce the new parameters into the discussion at this time?
> Walter
>
> Walter Katz
> 303.449-2308
> Mobile 720.333-1107
> wkatz (AT) sisoft (DOT) com
> www.sisoft.com
>
> -----Original Message-----
> From: si-list-bounce (AT) freelists (DOT) org [mailto:si-list-bounce (AT) freelists (DOT) org]On
> Behalf Of Muranyi, Arpad
> Sent: Tuesday, February 09, 2010 10:33 AM
> To: si-list (AT) freelists (DOT) org
> Subject: [SI-LIST] Re: IBIS-AMI Vendor Support Help
>
> Ken,
>
> [External Model] may not work because it is specifically designed to
> replace the [Model] with a better modeling language content while
> preserving the same connectivity of the [Model] keyword. This means
> that the all of the nodes of [Model] or [External Model] are supposed
> to be connected the same exact way, which would prevent us to put die
> parasitics between the buffer model and the pads.
>
> However, [External Circuit] does not have this restriction, so we
> could potentially make use of that keyword for this purpose, perhaps
> referencing the new (proposed) IBIS-ISS capabilities which allow
> us to use S-parameters. If we went with this route, the question
> is, can we have a [Model] behind the [External Circuit], which is
> something I don't remember for sure now. I need to research the
> spec on that a little more...
>
> Arpad
> ================================================== ======================
>
> -----Original Message-----
> From: Ken Willis [mailto:kwillis (AT) sigrity (DOT) com]
> Sent: Tuesday, February 09, 2010 9:19 AM
> To: 'Todd Westerhoff'; Muranyi, Arpad
> Cc: si-list (AT) freelists (DOT) org
> Subject: RE: [SI-LIST] Re: IBIS-AMI Vendor Support Help
>
> Hi Todd,
>
> On #1, it makes sense to incorporate S-parameters into IO models,
> especially
> since some major IP suppliers have this data today. I think the item
> that
> brings some uncertainty is why this would be pushed into the algorithmic
> (AMI) portion of the model, especially if new syntax/API has to be
> invented
> and supported. The intent of AMI was to capture the adaptive filtering
> behavior of these advanced Serdes devices, which you couldn't really
> model
> well in a circuit model. If the S-parameters are largely parasitic data,
> it
> seems more straightforward to include them in the analog circuit part of
> the
> model.
>
> I understand that there is no standard "IBIS" syntax for S-params right
> now,
> but I would agree with Arpad that Touchstone is widely supported. Maybe
> we
> should be looking at [External Model] > Touchstone or something like
> that.
> But we can continue the discussion in the IBIS forum.
>
> Thanks,
>
> Ken Willis
> Sigrity, Inc.
> 860-871-7070
> kwillis (AT) sigrity (DOT) com
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request (AT) freelists (DOT) org with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> http://www.freelists.org/webpage/si-list
>
> For help:
> si-list-request (AT) freelists (DOT) org with 'help' in the Subject field
>
>
> List technical documents are available at:
> http://www.si-list.net
>
> List archives are viewable at:
> http://www.freelists.org/archives/si-list
>
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request (AT) freelists (DOT) org with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> http://www.freelists.org/webpage/si-list
>
> For help:
> si-list-request (AT) freelists (DOT) org with 'help' in the Subject field
>
>
> List technical documents are available at:
> http://www.si-list.net
>
> List archives are viewable at:
> http://www.freelists.org/archives/si-list
>
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
>
>
>

--
Scott McMorrow
Teraspeed Consulting Group LLC
121 North River Drive
Narragansett, RI 02882
(401) 284-1827 Business
(401) 284-1840 Fax

http://www.teraspeed.com

Teraspeed® is the registered service mark of
Teraspeed Consulting Group LLC

------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request (AT) freelists (DOT) org with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request (AT) freelists (DOT) org with 'help' in the Subject field


List technical documents are available at:
http://www.si-list.net

List archives are viewable at:
http://www.freelists.org/archives/si-list

Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu