Hello Everyone,
I went over the minutes of last week's ATM meeting to refresh my memory
on the unfinished conversation we had on DC_Offset, so that I could gather
some thoughts on the subject and try to lead the conversation in a way so
that we could find a solution.
On thought that jumped at me is the discussion we had on the various
PAM4_***Threshold and the Rx_Receiver_Sensitivity parameters and their
relationship to DC_Offset. So I looked those parameters up in the spec,
because I wanted to find some details on them. For example, I think we
sort of implied in the discussion last week that Rx_Receiver_Sensitivity was
defined relative to an unnamed threshold level at the core logic input.
I am wondering whether that is actually true. If the input to AMI is based on
a difference waveform, there is no DC component in it. But isn't the output
of the Rx GetWave also a difference waveform? And as a result, isn't the
eye diagram also built from that difference waveform? If this is all true,
isn't
the Rx_Receiver_Sensitivity also applied to that difference waveform that
comes out of Rx GetWave? Note the emphasis on difference waveform.
I couldn't find any mention of differential to single ended conversion at
the "decision point" in the IBIS specification, which implies (right or wrong)
that the decision point is also making its decision on a difference waveform.
Having said all this, I think we cannot conclude that the
Rx_Reveicer_Sensitivity
parameter is actually defined with respect to an unnamed DC threshold voltage
level. For this reason, I don't think our proposed DC_Offset is related to it
in
this way.
I have similar questions about the PAM4_***Threshold parameters, but to
keep this email short, I am not going to get into that now.
Questions/comments welcome.
Thanks,
Arpad
================================================================