Subject: RE: [IBIS-Users] Question regarding Dynamic Clamp
From: Muranyi, Arpad (arpad.muranyi@intel.com)
Date: Wed May 07 2003 - 14:12:21 PDT
Tim,
The original purpose of the dynamic clamp in the IBIS specification
was to address some special clamping circuits we had in some of our
buffers. These clamps would look like a diode, but their reference
voltage (or knee voltage) would be shifted "in" towards the rails
for some time when the pad would see a transition. This would
effectively reduce overshooting signals without increasing the
DC (or leakage) currents for the rest of the time.
Later we started to make buffers which had on-die termination, but
in such a way that the termination was actually one or the other
transistor left on when the buffer was in the receive mode. In
SPICE it would be easy just to leave the buffer driving when it
is actually receiving, but some tools may not let you do that where
the definition of the stimulus is not accessible (or flexible enough)
to the user. It turns out that the dynamic clamp submodel allows
us to make models for this case also (even though this was not the
original purpose of it). I suspect that your buffer may fall into
this category.
Such buffers can be modeled by making a normal I/O buffer first,
and then making a copy of the pullup or pulldown IV curve over into
the submodel (depending on which way the buffer terminates when it is
in receive mode). If the buffer terminates to ground with its
pulldown transistor, I copy the pulldown IV curve to the submodel,
but I call it [GND Clamp] to be static. If I make this submodel
"non-driving" then this submodel (which includes its clamp) will
only be added to the main buffer in the receive mode. Since the
[GND Clamp] of this submodel is a copy of the pulldown IV curve of
the main buffer, this mechanism will be the same as if the main buffer
was driving low during the time it is receiving.
Regarding the error, the submodel is not required to have all four
IV curves. You can have only one [*** Clamp] keyword, and it is
legal. That's why the parser doesn't complain. If your particular
tool cannot handle that, you need to discuss that with the tool
vendor. They either didn't understand how to implement it, or
have technical difficulties. (If you want, you can tell me in a
private mail which vendor it is, we do have discussions with some
vendors regarding IBIS implementations).
I hope this answers your questions.
Arpad Muranyi
Intel Corporation
===================================================================
-----Original Message-----
From: Timothy Coyle [mailto:timothy.coyle@sun.com]
Sent: Wednesday, May 07, 2003 7:54 AM
To: IBIS List
Subject: [IBIS-Users] Question regarding Dynamic Clamp
Hi,
I know this topic has been discussed recently, but I'm not sure I understand how the Dynamic Clamp is suppose to work. I have an I/O buffer that calls a non-driving Dynamic Power Clamp. I have power clamp data in the I/O buffer model, and more power clamp data in the submodel.
This is for a on-die pullup to VDD. It seems the power clamp in the I/O model contains the pullup, and the submodel power clamp does not.
(this seems to be the power clamp when output is Hi-Z) The IBIS file passes the parser fine, but when I try to use a certain ibis to
specific model format translator, I get an error saying that the submodel dynamic clamp needs to have a pullup and/or a pulldown curve.
Has anyone published any articles on dynamic clamps or know of where any more info is located?
Regards,
Tim
|------------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org
|with just 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 written 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 just 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 written 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 : Wed May 07 2003 - 14:20:43 PDT