All,
I assume that we will go forward with something like "[Standard]
<standard> <buss>".
a) clock forwarding
a. The mechanism is straightforward (use the clock_times vector to
input clock times to the DLL
b. Documenting how the EDA tool generates this vector, and how the
model AMI_GetWave consumes this vector, and outputs clock_times is the
much larger problem.
c. This should be tabled until someone writes a BIRD with these
details resolved
b) moving Vref
a. This is specified by the [Standard]
c) rise/fall may be asymmetric
a. gets worse at higher data rate
i.
This statement may not be true
b. can be handled by EDA tool without changing the DLL footprint
i.
Cadence and SiSoft agree to this
ii.
Keysight says the AMI Flows are incorrect
c. IBIS says for statistical (Init) flows.
i. The
"analog" portion of the channel is characterized by means of an impulse
response leveraging the IBIS constructs for device models defined in
Sections 6.1, 6.2 and 6.3.
ii. The
first column of the impulse response matrix is the impulse response for a
through channel, a channel that serves as a communication path between a
transmitter/receiver pair.
d. IBIS says for time domain (GetWave) flows.
i.
Step 1. The EDA tool obtains the impulse response for the analog channel.
This represents the combined impulse response of the transmitter's analog
output, the channel and the receiver's analog front end.
ii. Step
6a. If Tx GetWave_Exists is True and Rx GetWave_Exists is True, the output
of Step 5 is convolved with the output of Step 1 by the EDA tool and the
result is passed on to Step 7.
e. Conclusions
i. The
IBIS flows talk about a single Impulse Response of the channel.
ii. We
should all agree that there are two Impulse Responses of the channel:
1. Rising Impulse Response (derivative of rising step response)
2. Falling Impulse Response (derivative of falling step response)
iii. One
clarification to the standard is to define the single Impulse Response
that is used in the flows to be derived by "Melding" the Rising and
Falling Impulse Responses.
1. Do we need to define how to "Meld" the two impulse responses, or
leave that to the EDA tool, since we do not tell how the EDA toll should
generate the Impulse Response for the existing flows.
iv. Are
there other ways that we should document in IBIS-AMI how to handle
asymmetric Rising and Falling Impulse Responses.
Walter
Walter Katz
<mailto:wkatz@xxxxxxxxxx> wkatz@xxxxxxxxxx
Office 978.461-0449 x 133
Mobile 720.417-3762
