Subject: RE: [IBIS-Users] C_comp issue?
From: Muranyi, Arpad (arpad.muranyi@intel.com)
Date: Mon Nov 11 2002 - 16:45:24 PST
Fred,
I think you are correct. What we often forget in these
discussions is to define our frame of reference. So we end
up talking about something and someone else reading the same
words may still misunderstand it because they see it from a
different point of view.
To reiterate then, I agree, that if I keep everything the same
in a transistor level model and change only the amount of
capacitance in it, for example, by tweaking the capacitance
parameters in the process file (Cj and the like), than that
model will give different waveforms, or Vt curves.
I think we also agree that when we simulate with a particular
IBIS model, then no matter what the C_comp parameter says, the
output should reproduce the same waveforms when the model is
driving and no reflections are present.
In case anyone wonders, this difference is by design, and not
because IBIS is inferior or incapable to reproduce the real
world. This is why:
When we obtain a waveform from a transistor model, the shape of
that waveform includes the capacitive effects. In IBIS simulators
this waveform is basically used as if it was a scaling coefficient
as a function of time to scale the size of the IV curve to make that
also vary as a function of time (to account for the transients).
(Remember the IV curve by itself is purely steady state and contains
no transient information).
IV curves are usually implemented as some sort of a current source.
Whether they are scaled with respect to time or not, they are a pure
current source, not including any capacitive parasitics. So if we
want the driver or receiver to exhibit any capacitance, we need to
have a capacitor in parallel with this current source. But if we
did that, then the shape of the Vt curve would also get slowed down,
which we don't want, because that would "double count" the effects
of the capacitance. So IBIS simulators usually have a compensation
algorithm to avoid this double counting, yet preserving the capacitance
so that it is still visible to the signals coming from the outside world.
I hope this helps clarifying anyone's doubts.
Arpad Muranyi
Intel Corporation
======================================================================
-----Original Message-----
From: Fred Balistreri [mailto:fred@apsimtech.com]
Sent: Monday, November 11, 2002 4:00 PM
To: Muranyi, Arpad
Cc: ibis-users@eda.org
Subject: Re: [IBIS-Users] C_comp issue?
I think that perhaps there is confusion. It is agreed that the
simulators should reproduce the v/t curves no matter what the
c_comp that is specified. What is not agreed is that for the
same model if you have two different c_comps then the v/t curve
will be the same. I think what you mean is that for a given
ibis model the v/t curves contain the c_comp. In other words
what ever c_comp is given in the ibis model, that is what
produced the v/t curve.
Given that you have the same transistor level model,
then changing the capacitance in the
original netlist will give you a different v/t curve and
a different c_comp and therefore a different IBIS model.
What it sounds that you are suggesting is that no matter what
the physical capacitance is, that the waveform does not change.
That surely is not true.
Best regards,
"Muranyi, Arpad" wrote:
>
> Todd,
>
> Your description is absolutely correct, and Stephen Peters and Bill Panos
> are only partially correct. Thank you for the excruciating and unambiguous
> details in your description.
>
> All,
>
> To reiterate what Todd said, C_comp is the capacitance you see when you look
> into the die. Package is not included. The waveforms in the model (Vt curves)
> are not effected by C_comp, i.e. the model should reproduce the same waveforms
> if simulated using the same R_fixture and V_fixture combinations that are
> in the Vt curves REGARDLESS of what C_comp is.
>
> Waveforms will only change due to C_comp if the signal comes from the outside
> world, either due to reflections while the driver is driving, or when it is
> in the receiver mode.
>
> If a particular tool does not give these results it is time to talk to the
> tool vendor. Most tools I know of do it this way.
>
> Arpad Muranyi
> Intel Corporation
> ===========================================================================
>
> -----Original Message-----
> From: Todd Westerhoff [mailto:twesterh@cisco.com]
> Sent: Monday, November 11, 2002 1:34 PM
> To: ibis-users@eda.org
> Subject: RE: [IBIS-Users] C_comp issue?
>
> There are a few points here worth mentioning - if only to verify that I
> understand how this stuff is *supposed* to work.
>
> If you vary the C_comp value of the *receiver* on a net - then yes, I expect
> to see the waveforms change, as you're effectively changing the loading the
> receiver places on the net.
>
> However, if you vary the value of C_comp for the *driving* device, then the
> fact that you've changed C_comp will *only* affect any reflections on the
> net that get back to the driver. In other words, changing the driving
> device's C_comp value does *NOT* affect its actual behavior as a driver;
> only its behavior as a load is affected. This is true whether the driver is
> acting as a receiver (i.e. an I/O in the "input" state), or whether the
> driver is absorbing reflections (i.e. the driver is "quiet" because it has
> stopped sweeping the V-T curve).
>
> ... all of which is a long-winded way of saying that IBIS simulators place
> C_comp in parallel with the driver's output and then compensate its presence
> out; thereby reproducing the V-T behavior from the IBIS model. So, changing
> the value of C_comp doesn't actually change the output's *driving* behavior.
>
> Am I understanding this correctly, or have I gotten myself confused again?
>
> Todd ;-}
>
> Todd Westerhoff
> Signal Integrity Engineer
> Cisco Systems
> 250 Apollo Drive - Chelmsford, MA - 01824
> email:twesterh@cisco.com
> ph: 978-497-0272
> ============================================
>
> "When did the choices get so hard, with so much more at stake?
> Life gets mighty precious when there's less of it to waste"
>
> - Bonnie Raitt, "Nick of Time"
>
> -----Original Message-----
> From: owner-ibis-users@eda.org [mailto:owner-ibis-users@eda.org]On
> Behalf Of Peters, Stephen
> Sent: Monday, November 11, 2002 3:47 PM
> To: 'Eric Hsu'; ibis-users@eda.org
> Subject: RE: [IBIS-Users] C_comp issue?
>
> Hello Eric:
>
> C_comp represents the I/O (input or ouptut) capacitance of the driver or
> reciever itself. It is the capacitace seen when standing on the pad of the
> die and looking back into the I/O cell.
>
> Changing C_comp will effect a number of things, but yes -- C_comp has a
> direct effect on a drivers rise or fall time and waveforms.
>
> Regards,
> Stephen Peters
> Intel Corp.
>
> -----Original Message-----
> From: Eric Hsu [mailto:ehsu@netlogicmicro.com]
> Sent: Monday, November 11, 2002 12:00 PM
> To: ibis-users@eda.org
> Cc: ehsu@netlogicmicro.com
> Subject: [IBIS-Users] C_comp issue?
>
> Hi All,
>
> What is the [c_comp] role in ibis? If changing [c_comp] value, will I
> get different results in rising/falling waveform?
>
> Best Regards,
>
> -eric
>
> |------------------------------------------------------------------
> |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
>
> |------------------------------------------------------------------
> |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
-- Fred Balistreri fred@apsimtech.comhttp://www.apsimtech.com |------------------------------------------------------------------ |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 : Mon Nov 11 2002 - 16:53:22 PST