[SI-LIST] Re: What frequency are IBIS models good to?
- From: "Muranyi, Arpad" <arpad.muranyi@xxxxxxxxx>
- To: "si-list" <si-list@xxxxxxxxxxxxx>
- Date: Tue, 14 Sep 2004 12:19:48 -0700
Eric,
It is true that IBIS (up to version 4.0) was basically made up
of DC IV curves, but there is a little more to it than just that.
IBIS also has a parameter to represent the die-capacitance, called
C_comp. This includes everything that you see when you look into
the die at the pad. C_comp can be split into 4 parts which are
connected to the respective reference nodes that the corresponding
IV curves are connected to. So if the pullup IV curve is connected
to the pullup reference, it can have its parallel C_comp_pullup,
and so on. This feature comes handy when someone is interested
in power delivery simulations.
One of the biggest drawback of C_comp for very high speed
simulations is that it is a single and constant number (aside
from the 4-way splitting mentioned above). I.e., you can't=20
describe the voltage or frequency dependence. For this reason,
the IBIS model maker has to pick the best value for C_comp that
is valid for the frequency of operation and the signaling levels
of the buffer. This can be done, but may not be sufficient in
every situation.
This is where the latest version of the IBIS specification v4.1
comes handy. In this version we added new language extensions to
IBIS with which you can basically describe a buffer to any level
of accuracy. For example, using VHDL-AMS, or Verilog-AMS you can
write your own capacitance equation(s) any whichever way you want
to. You can include charge conservation, frequency and voltage
dependency, or anything you need. You could even model the piezo
effects due to mechanical vibration if you want to, just to name
an extreme example which was recently mentioned on the SI-list.
(Even though I doubt that this effect would appear on the die).
It is now strictly up to the model writer to make a model as
accurate as possible.
Arpad Muranyi
Intel Corporation
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx]
On Behalf Of Eric Goodill
Sent: Tuesday, September 14, 2004 11:55 AM
To: si-list
Subject: [SI-LIST] What frequency are IBIS models good to?
Howdy,
We were having a discussion around work the other day about the maximum=20
frequency any given SPICE or IBIS model is good to. For SPICE circuit=20
models, it seemed pretty clear that the frequency accuracy of the model=20
depends on how good a job the model builder did at including parasitics=20
and so forth that come into play as the frequency increases.
Our understanding of IBIS models is not great, but we understand them to
be essentially I-V curves. One fellow asserted that this info is=20
gathered essentially at DC and thus could not represent the=20
high-frequency (> 500 MHz) behavior of a driver.
Comments?
-Eric
--=20
Eric Goodill Redback Networks
ericg@xxxxxxxxxxx 300 Holger Way
voice: (408) 750-5319 San Jose, CA 95134-1362
------------------------------------------------------------------
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 FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
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:
