RE: [IBIS-Users] Sweep voltage listing query

Abrar,
 
I am going to reply publicly, beaus this is a common question,
and I would like other people to benefit from this reply also.
 
The main reason for the rule of sweeping the IV curves from
-Vcc to 2*Vcc is based on the basic reflection phenomena of
transmission lines.  If you have an open line, the voltage
can double due to reflections.
 
If your driver could drive a rail to rail signal into a T-line
(0 to Vcc, or Vcc to 0), there is a possibility that you could
see a reflection that goes to 2*Vcc or -Vcc.
 
However, this will usually not happen in real life, because
 
1) in order to be able to drive rail to rail, the driver
impedance and T-line impedance ratio would have to be extreme
(low driver Z and high T-line Z)
 
2) in order to get a doubling at the end of the T-line, the
T-line impedance and the receiver's impedance, including its
clamps, would also have to be extreme (low T-line Z and high
receiver Z)
 
For one, these two conditions pose a conflicting requirement
on the T-line impedance, and second, practical designs usually
cannot have extremely low driver and extremely high receiver
impedances.  In addition, most high speed buses today are
terminated anyway, so reflections are not going to double
because of that.
 
So, the bottom line is that the doubling of the signal is only
an extreme or absolute worst case situation, and considering
this background, it really doesn't matter whether the IV curve
is swept -Vcc to 2*Vcc using the typical or maximum supply
voltage conditions.
 
There is another problem.  The IBIS specification doesn't spell
out which voltage we should use for calculating -Vcc or 2*Vcc.
To be safe you could use the max Vcc values, but I usually use
the typical only.  There is no rule for this in the spec.
 
To add another twist to this subject, consider the low voltage
swing devices.  For example, in GTL signaling the high level is
at 1.5 V, and the low level is anywhere between 0 and 550 mV
depending on how strong the driver is.  If the driver was a
super conductor, it could drive to 0 V.  In this case a full
reflection would give you -1.5 V on the low side, and 3.0 V
on the high side.  But GTL is a terminated bus, so you will
never see full reflections.  So considering a 3.3 V I/O supply
on a GTL device, does it really make sense to sweep the IV
curves from -3.3 V to 6.6 V?  Absolutely not.  Yet the IBIS
specification doesn't say anything about that.  If you obey
the letter of the law, you will use up a lot of the 100 points
in areas of the IV curve which will never be used.
 
So my recommendation is to seep the double of the signal swing
to the top and bottom.  In this GTL case this would be no more
than -1.5 V and 3.0 V.  Similar situations can arise with other
low voltage swing buffers, such as LVDS, etc...
 
Since this is illegal according to the IBIS spec, I also
recommend that after you are done, just add to more points
with extrapolation to the ends of the IV curve at -Vcc and
2*Vcc (without going over the maximum limit of 100 points)
to satisfy the spec and avoid any possibilities for the
parser to give you error messages.
 
I hope this helps your understanding of the IV curve ranges.
 
Arpad Muranyi
Intel Corporation
---------------------------------------------------------------
 
________________________________

From: Abrar Ahmed [mailto:abrar_techie@yahoo.com] 
Sent: Tuesday, April 11, 2006 11:05 PM
To: Muranyi, Arpad
Subject: RE: [IBIS-Users] Sweep voltage listing query


Dear Mr.Arpad,
 
You mean to say that the sweep range is fixed at the "typical case" values only?
If that is the case, what is the wisdom behind selecting only typical case ?
What about using "max case" as the sweep range, we get more values, would we not.
 
Regards
Abrar
 
======================================================

"Muranyi, Arpad" <arpad.muranyi@intel.com> wrote:

	Abrar,
	 
	The pullup and power clamp tables are Vcc relative, so
	the origin of these IV tables are at the supply voltage.
	When the supply voltage moves for min and max, the origin
	of the IV tables moves together with it.  The end points
	of the tables only determine the sweep range, which has
	nothing to do with what the supply voltage variation is.
	 
	I hope this clarifies your question.
	 
	Arpad


|------------------------------------------------------------------
|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