Subject: re: RE: [IBIS-Users] Information about IBIS Warnings .
From: Lynne Green (lgreen@cadence.com)
Date: Thu Nov 21 2002 - 11:45:10 PST
re: questions on SPECCTRAQuest warnings.
1) SPECCTRAQuest checks area under the curves. The warning tells you
that you should
visually check the curves to verify the V-t columns are in the correct
typ/min/max order.
Note that V-t curves can be viewed in Model Integrity, released in
SPECCTRAQuest 14.2.
4) The "hump" in the V-t curve probably exceeded 1%. The warning just
tells you that you
should visually check the V-t curves. This is probably not an
indication of an error in the
model. This is not related to any I-V mismatch problems.
5) To model a _change_ in pullup resistance, _IF_ the change is
resistance is "small", you
can try adding or subtracting the current delta_I caused by the change
in resistance. This
has the advantage of preserving some of the non-linearity of a pullup
resistance implemented
as a MOSFET.
delta_I = -delta_R * V / (R*R)
delta_I = -delta_R * I / R
delta_I = -delta_R * I * I / V
Note that this does NOT apply to DETECTING resistor terminators. For
further information on
detecting these, see Arpad's "JEDEC" presentation at
http://www.eda.org/pub/ibis/training/.
Best regards,
Lynne
Dr. Lynne Green
Senior Member of Consulting Staff
Cadence Design Systems, Inc.
"All the world's an analog stage, whereon digital plays bit parts."
Fred Field , Hughes Aircraft
-----------------------------------------------------
-----Original Message-----
From: Ingraham, Andrew [mailto:Andrew.Ingraham@hp.com]
Sent: Thursday, November 21, 2002 7:47 AM
To: whiz kid
Cc: ibis-users@eda.org
Subject: RE: [IBIS-Users] Information about IBIS Warnings .
> I have few questions about what the warnings mean. I
> am using Specctraquest for my simulations.
>
> 1) PullDown : Overall Typical Area exceeds Maximun
> Area. What has area goto do ????
If it is talking about transistor area, then it must be coming from a
non-IBIS model in your simulation.
But perhaps it is checking the areas under the curves in the IBIS model.
Generally speaking, I(V) ought to be bigger in the Max column than the
Typ column. But exceptions may exist at certain points or ranges (since
Vcc also changes, etc.). So Specctraquest may be comparing the areas
under the I(V) curves to make sure the Max data is indeed "stronger"
than the Typ data, and it sees that it isn't. That's my guess, but it's
only a rough guess.
> 2) Duplicate Points removed. This may sound dumb but
> why will the curves have duplicate points in the first
> place.
Round-off? (Closely-spaced points became the same.)
Or, if two or more tables were combined, resulting in overlap.
> 3) TV curves should asymptotically approach a final
> value- Point added. what does asymptotically mean.
Asymptotically means the curve should continually approach the final
value = its asymptote. This is not a precise definition. It means that
the curve would eventually reach the final value if you could look out
along the curve toward infinite time. Technically it approaches but
never quite reaches it, in finite time.
Asymptotically often also implies that the difference between the curve
and its asymptote always gets smaller as you move farther out (i.e., the
polarity of the slope does not change).
> 4) Falling wave form Fall1 Non monotic dv=-- at time
> dt=-- exceeds 1% of voltage span. Do we need to worry
> that the calculated value is off compared to the VI
> curves . How is the best way to fix the above warning.
Yes; if I understand correctly, this warning message (or its earlier
version which was slightly different) has come up often in discussion
groups.
The V-T waveforms and I-V curves should be consistent with one another.
If the V-T waveform begins, or ends, at a value that is not consistent
with the I-V tables, then something is wrong, and "all bets are off." In
other words, what comes out of the simulator may be garbage, because it
is trying to satisfy both requirements simultaneously, which it can't.
It means there is something wrong with the IBIS model, and should be
fixed.
> 5) Also, I have a open drain IBIS model with a built
> in Pull up resistor modelled into it. Is there a way
> to create a new model with out the pull up so I can
> play with different values of the Pullup in my
> simulation. How do I go about creating this new model.
The pull-up resistor's effects are probably in the Clamp curves.
If you know the typ/min/max resistor values that were used when the
model was created, you could try calculating the resistor current vs.
voltage, and subtract it from the tables. Then add your own pull-up
resistor currents.
But this assumes the pull-up resistor was linear ... i.e., a real
resistor and not a pull-up FET which is how many ICs today implement
on-die pull-ups.
If you assume that nearly all of the current in the Clamp tables,
between GND and VCC, was due to the pull-up resistor, then you might try
zero'ing the current in the [GND Clamp] table for positive voltages. But
this is kludgy and not very accurate.
I wouldn't know how to modify the V-T waveforms, so maybe it's better to
just remove them from the model and rely on [Ramp] data instead. But
the old [Ramp] data (especially dv/dt_r) probably isn't right either!
It is probably better to go to the source of the IBIS models and have
new models created.
Regards,
Andy
|------------------------------------------------------------------
|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 : Thu Nov 21 2002 - 11:54:46 PST