Ambrish,
Thanks for your reply. I would like to clarify a few details.
First, I would like to understand what exactly you mean by "bit by bit channel
simulation".
Are you referring to what we call the "GetWave flow" in the IBIS-AMI
specification, or
something similar, but not necessarily exactly the same?
Here is why I am asking:
You say "Channel simulation required an accurate characterization of the
channel". That
is true for any simulation, if we want good results. But an accurate channel
characterization
that results in a single channel impulse response will not include the effects
of different rising
and falling edges of the Tx, no matter how accurate it is. So the trick is
hiding somewhere
else.
The second half of that sentence says "and the analog front end of the Tx/Rx
that was
modeled in the IBIS model (which had asymmetric rise/fall edges)". Are you
essentially
saying that the "Channel simulation required ... the analog front end of the
Tx/Rx that
was modeled in the IBIS model (which had asymmetric rise/fall edges)", in other
words,
were you using the Tx analog model (whose rising and falling edges are not
symmetric)
in the channel simulation instead of the ideal, digital stimulus to include the
asymmetry
effects?
Thanks,
Arpad
=====================================================================
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx] ;
On Behalf Of Ambrish Varma
Sent: Friday, February 8, 2019 6:57 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: A question about asymmetric edges in AMI simulations
Arpad,
Walter is correct. There was no AMI model used in the simulation. The
comparison was between
a) simulating a random bitstream with time domain circuit simulation and
b) same bitstream with bit by bit channel simulation.
Channel simulation required an accurate characterization of the channel and the
analog front end of the Tx/Rx that was modeled in the IBIS model (which had
asymmetric rise/fall edges).
There was no equalization modeled or simulated.
Thanks,
Ambrish.
From: ibis-macro-bounce@xxxxxxxxxxxxx<mailto:ibis-macro-bounce@xxxxxxxxxxxxx>
<ibis-macro-bounce@xxxxxxxxxxxxx<mailto:ibis-macro-bounce@xxxxxxxxxxxxx>> On
Behalf Of Walter Katz
Sent: Friday, February 8, 2019 6:27 PM
To: Arpad_Muranyi@xxxxxxxxxx<mailto:Arpad_Muranyi@xxxxxxxxxx>;
ibis-macro@xxxxxxxxxxxxx<mailto:ibis-macro@xxxxxxxxxxxxx>
Subject: [ibis-macro] Re: A question about asymmetric edges in AMI simulations
EXTERNAL MAIL
Arpad,
What they did had nothing to do with AMI modeling. They (and I) just used the
AMI flow of generating a "Melded Impulse Response" of the channel, and then
convolving it with their "Melded" Impulse response.
We both had no Tx Equalization, and we generated waveforms at the Rx Pad.
Walter
From: ibis-macro-bounce@xxxxxxxxxxxxx<mailto:ibis-macro-bounce@xxxxxxxxxxxxx>
<ibis-macro-bounce@xxxxxxxxxxxxx<mailto:ibis-macro-bounce@xxxxxxxxxxxxx>> On
Behalf Of Muranyi, Arpad
Sent: Friday, February 8, 2019 4:26 PM
To: ibis-macro@xxxxxxxxxxxxx<mailto:ibis-macro@xxxxxxxxxxxxx>
Subject: [ibis-macro] A question about asymmetric edges in AMI simulations
Ken, Kumar, Ambrish,
I was impressed with the presentation you gave at the IBIS Summit at DesignCon
on the
topic of "Channel Simulation Using IBIS Models with Asymmetric Rising and
Falling Edges".
http://www.ibis.org/summits/feb19/willis1.pdf<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ibis.org_summits_feb19_willis1.pdf&d=DwMFAg&c=aUq983L2pue2FqKFoP6PGHMJQyoJ7kl3s3GZ-_haXqY&r=0Gavo25Aqgu_F0CPGlNHGxaE1UHA4ptnU3q-kd8FzDQ&m=Hv1hSgpqzKJ7xgNALhVIIW7j9bPWLhynLTsZsnJZwQg&s=nc_aPjLbPhE6Z_XNgZ4aHEMaEOZmJOv7Am8XvQRsOuk&e=>
On slide 2 you mentioned that you were using Micron's y11a.ibs file for the
simulations.
I looked at that file, and noted that it does not have any [Model]s with the
[Algorithmic
Model] keyword. I don't see anything in the slides about what AMI model you
were
using for the "Channel Simulation". I vaguely remember that in the discussion
that followed
the presentation you mentioned that it was a standard AMI model of some sort.
Could you
tell us whose/which AMI models (DLL) you were using to generate the simulation
waveforms
in the presentation?
Thanks,
Arpad
========================================================================