According to the request in today’s ATM meeting, I am forwarding my comments
on BIRD158.6 draft2 to the ATM reflector. Questions, comments, suggestions are
all welcome…
Thanks,
Arpad
=================================================================
From: Muranyi, Arpad
Sent: Wednesday, June 28, 2017 6:14 PM
To: 'radek_biernacki@xxxxxxxxxxxx' <radek_biernacki@xxxxxxxxxxxx>
Subject: Comments on BIRD158.6 draft2
Radek,
I am sending this to you privately, because I have a lot of things to pick on
and don’t want to do it publicly to avoid any embarrassment, etc… Maybe
after we converge on something we can go public…
The following sentence sounds a little awkward to me:
“This section discusses alternative analog buffer modeling devised specifically
for AMI applictions.”
Maybe something along these lines would be better (but I would leave it up
to native English speaking people to find the best wording):
“This section discusses an alternative analog buffer modeling technique,
specifically designed for AMI applications.”
The voltage labeling on the “transmitter analog circuit” drawing and the text
below it seems to be “out of touch” with each other.
“The voltages of the voltage sources correspond to V=Tx_V for logic level 1,
and V=-Tx_V for logic level 0.”
I don’t see any “Tx_V” in the drawing. Also, the “Vdc” notation in the drawing
seems to be misleading because these sources are NOT DC voltage sources,
these are bit patter sources. Also, the drawing uses “V/2” for the value of
these
sources, but I don’t see “V” being mentioned in the text anywhere… In addition,
this paragraph doesn’t mention Tx_R at all, and I think it should explain that
too.
I think the text and drawing should match.
My next comment is a minor thing, but may eliminate misinterpretations later.
I would change this sentence:
“Ports 1 and 3 are at the stimulus source side, and ports 2 and 4 are connected
to the buffer terminals.”
to:
“Ports 1 and 3 are at the stimulus source side, and ports 2 and 4 are the
transmitter buffer’s output terminals.”
In the receiver analog circuit description section, I would change this
sentence:
“Ports 1 and 3 are connected to the buffer terminals, and ports 2 and 4 serve
as the differential input to the Rx algorithmic model.”
to:
“Ports 1 and 3 are the receiver buffer’s input terminals, and the waveforms
obtained at ports 2 and 4 represent
the channel’s response that is used for the AMI simulations.”
We can work some more on the exact wording, especially the 2nd half of this
sentence, but the main point here is that the waveform at ports 2 and 4 are
not the input to the Rx AMI model. (If anything they are the input to the
Tx AMI_Init function)…
The following two paragraphs have multiple problems:
“The IBIS AMI flow requires that the EDA tool generates the impulse response of
the entire analog circuitry from Tx to Rx algorithmic models. Typically, the
Touchstone file data specified here is to be used for either the Tx analog
buffer excluding the Tx package model and/or the Rx analog buffer model
excluding the Rx package model. In the preceding sentence the term “package”
may mean just the package or the package together with the on-die interconnect.
The following figure illustrates the corresponding entire setup when both Tx
and Rx use the Ts4file parameter.”
“Please note that the package data is added to the channel by the user using
user’s own data or using some other vendor’s data. This means that the Tx or
the Rx analog circuits specified in the AMI file are to be used in lieu of the
analog buffer model. This can be modified by another new reserved AMI parameter
Ts4file_Boundary. In any case the package and possibly the on-die interconnect
data associated with the IBIS model pointing to this AMI file via the
[Algorithmic Model] keyword shall not be automatically incorporated into the
above schematic by the EDA tool.”
The 1st sentence may sound better this way:
“The IBIS AMI flow requires that the EDA tool generates the impulse response of
the entire analog circuitry including the Tx and Rx analog buffer models.”
The rest of this paragraph and the following paragraph could be combined and
simplified like this for easier reading:
“Typically, the Touchstone file data specified here will be used to describe
only the analog behavior of the buffer itself, excluding the effects of the
package and/or the on-die interconnect, as illustrated in the following figure.
This way the Tx or Rx analog circuits specified in the AMI file may be used as
a direct replacement for the corresponding analog model described by the
[Model] keyword. However, the model maker may choose to include the package
and/or on-die interconnect model in the Touchstone file data. The new reserved
AMI parameter Ts4file_Boundary shall be used by the model maker to inform the
EDA tool about the content of the Touchstone file. If the model maker includes
the package effects in the Touchstone file (i.e. Ts4file_Boundary is set to
“pin”), the EDA tool must ignore the package model in the IBIS file. If the
package effects are not included in the Touchstone file (i.e. Ts4file_Boundary
is set to “buffer” or “pad”), the EDA tool shall provide a selection mechanism
for the user to decide whether to automatically include the applicable package
model from the IBIS file or not. This gives the user an option to manually add
other package and/or on-die interconnect models to their simulation setup.”
The content of the next paragraph should be moved up to the text below
the transmitter analog circuit and the text below the receiver analog circuit.
“For Tx models that have the reserved parameter Ts4file, the reserved parameter
Tx_V is required and the reserved parameter Tx_R is optional. For Rx models
that have the reserved parameter Ts4file, the reserved parameter Rx_R is
optional. In other words, for a Tx buffer, the transmitter circuit defines the
analog buffer model between the zero impedance stimulus input voltage source
and the buffer terminals. For an Rx buffer, the receiver circuit defines the
analog buffer model between the buffer terminals and a high impedance probe at
the input to the Rx Algorithmic model.”
I am not sure what the purpose of the last paragraph is:
“Given that the Touchstone 4-port model is LTI there are many methods of
generating the impulse response to be used in AMI modeling that will give the
identical result within numerical accuracy of the technique chosen. One
technique commonly used in EDA tool simulation is to generate the impulse
response by applying a step stimulus and calculating the time derivative of the
response. When both Tx and Rx Ts4file parameters are present the impulse
response is measured between the SRC_pos/SRC_neg input and a high impedance
differential probe between ports 2 and 4 of the Rx 4-port network. When only
one of Tx or Rx .ami files has the Ts4file parameter present then the other
component’s contribution to the impulse response comes from the model details
provided under the [Model] keyword referencing the .ami file without the
Ts4file parameter.”
The 1st sentence really doesn’t say anything. The 2nd sentence basically
contradicts the purpose of this entire approach, which is to be able to do
the channel response without a time domain step response simulation.
The 3rd sentence is kind of obvious, but could be combined with what I
suggested in my change in the receiver analog model text. The last sentence
doesn’t say what it really should say, which is that in the mixed model case
we have to resort back to the time domain step response simulation approach…
I didn’t read the rest of the document yet. But in order to make it easier
for you, I am attaching my version of this BIRD including all the changes I
suggested above. I did an accept all changes on your document, and turned
on the tracking after that, so the marked changes you see are only my changes.
Let me know if you want to get together on the phone to discuss these
suggestions.
Thanks,
Arpad
================================================================
From: ibis-macro-bounce@xxxxxxxxxxxxx<mailto:ibis-macro-bounce@xxxxxxxxxxxxx>
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Curtis Clark
Sent: Wednesday, June 28, 2017 12:46 PM
To: IBIS-ATM <ibis-macro@xxxxxxxxxxxxx<mailto:ibis-macro@xxxxxxxxxxxxx>>
Subject: [ibis-macro] Minutes from the 27 June ibis-atm meeting
Minutes from the 27 June ibis-atm meeting are attached.
The following document, which was discussed during the meeting, has been posted
to the ATM work archives.
DATE
AUTHOR<http://ibis.org/atm_wip/archive-author.html>
ORGANIZATION<http://ibis.org/atm_wip/archive-org.html>
TITLE<http://ibis.org/atm_wip/archive-title.html>
FORMATS
27-JUN-2017
Radek Biernacki
Keysight Technologies
BIRD 158.6 draft 2
(zip<http://ibis.org/atm_wip/archive/20170627/radekbiernacki/BIRD_158_6_draft_2.zip>)(docx<http://ibis.org/atm_wip/archive/20170627/radekbiernacki/BIRD%20158.6%20draft%202/BIRD_158.6_draft2.docx>)
Attachment:
BIRD_158.6_draft2_AM.docx
Description: BIRD_158.6_draft2_AM.docx
