Subject: RE: [IBIS] waveform fixtures for differential models
From: Muranyi, Arpad (arpad.muranyi@intel.com)
Date: Tue Sep 30 2003 - 08:35:24 PDT
Mike,
I was playing with the equations a few months ago when I wrote a
buffer model using IBIS data in VHDL-AMS. The presentation is
posted on the IBIS web site:
http://www.eda.org/pub/ibis/summits/jun03a/muranyi1.pdf
One of the goals I had was to find out why we got such a bad
correlation when the V_fixture values of the Vt curves were
different from the actual termination voltage in the simulation.
http://www.eda.org/pub/ibis/summits/sep01/muranyi1.pdf
I also wanted to answer the question you are raising: What
should we use for a DDR buffer where the Vtt is Vcc/2. Should
we make the model with V_fixture=Vtt and have only two Vt curves,
or should we use Vcc and GND for the Vt curves and have four?
It turned out that the miscorrelation is inherent in this
two equations two unknowns system (and it is not a tool
implementation problem). Also, I found out that as long as
one of the V_fixture values matches the actual termination
voltage in the simulation, you will get very accurate results,
and it really doesn't matter what the other V_fixture value is.
But you are still going to get better results if you have four
Vt curves instead of two. I also found that the results are less
inaccurate if the termination of the simulation is between the
two V_fixture values and worse when it is outside them. In
addition, we found (with Luca's help) that using multiple Vt
curves can help in making a more general buffer model that can
simulate accurately with arbitrary termination voltages if the
Vt curves use several V_fixture values between Vcc and GND.
So, to answer your question, I would suggest that you make your
model with at least four Vt curves. If you can only make four,
than make one of the two V_fixture values the same as your
termination voltage in the simulation. The other V_fixture
can be Vcc or GND, or anything else you like. If you can make
more than four Vt curves, then space them around the area of
interest or between Vcc and GND evenly.
I know of two tools (there may be more) which support multiple
Vt curves, so you could actually utilize a model made this way
if you work with those tools. (Cadence and Hyperlynx).
I hope this helps,
Arpad
================================================================
-----Original Message-----
From: owner-ibis@eda.org [mailto:owner-ibis@eda.org]On Behalf Of Mike
LaBonte
Sent: Monday, September 29, 2003 1:34 PM
To: IBIS
Subject: [IBIS] waveform fixtures for differential models
We have encountered IBIS models with only 1 rising and 1 falling waveform,
each with V_fixture set to 1/2 of power voltage. These are for differential
buffers where the termination voltage is effectively the midpoint, so it
seems to make sense. My question is, will all IBIS simulators simulate this
model correctly? Or are there simulators that absolutely require waveforms
with V_fixture at both power and ground levels?
Mike LaBonte
-----------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org
|with the appropriate command message(s) in the body:
|
| help
| subscribe ibis <optional e-mail address, if different>
| subscribe ibis-users <optional e-mail address, if different>
| unsubscribe ibis <optional e-mail address, if different>
| unsubscribe ibis-users <optional e-mail address, if different>
|
|or email a request to ibis-request@eda.org.
|
|IBIS reflector archives exist under:
|
| http://www.eda.org/pub/ibis/email_archive/ Recent
| http://www.eda.org/pub/ibis/users_archive/ Recent
| http://www.eda.org/pub/ibis/email/ E-mail since 1993
-----------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org
|with the appropriate command message(s) in the body:
|
| help
| subscribe ibis <optional e-mail address, if different>
| subscribe ibis-users <optional e-mail address, if different>
| unsubscribe ibis <optional e-mail address, if different>
| unsubscribe ibis-users <optional e-mail address, if different>
|
|or email a request to ibis-request@eda.org.
|
|IBIS reflector archives exist under:
|
| http://www.eda.org/pub/ibis/email_archive/ Recent
| http://www.eda.org/pub/ibis/users_archive/ Recent
| http://www.eda.org/pub/ibis/email/ E-mail since 1993
This archive was generated by hypermail 2b28 : Tue Sep 30 2003 - 08:36:38 PDT