This week's call will focus on progress on the actions from last week's
call.
Does anyone have any presentations on progress? Please let me know in
advance if possible.
I'd like to put some dates on specific efforts if possible.
*******
Dial-in details
Proposed agenda
----- Original Message -----
Sent: Friday, December 03, 2004 11:36
AM
Subject: [10GMMF] Notes from TP2 meeting
12/2/04
The TP2 call was held as scheduled.
Attendees
- Jan Peeters Weem
- Nick Weiner
- Tom Lindsay
- John Jaeger
- Greg Lecheminant
- Martin Lobel
- Pavel Zivny
- Lew Aronson
- Raj Savara
- Abhijit Shanbhag
- Others who did not ID?
Agenda discussed and approved (see below)
Tasks and distribution of work (per attachment on meeting announcement TL
email, 12/1 sent 21:41)
- Review D1.0 schedule
- Will be out for review and comment ~12/6
- Comments due 1/7/05
- Our objective - resolve as many details as possible by/before
comment deadline
- Waveform capture method
- Pattern definition
- John Ewen has led this work so far and has volunteered to continue
to address the listed questions.
- Pavel would like to be involved.
- John has analyzed BnBi for a PRBS7-like pattern and has summarized
the info into PPT format. Tom to ask John if he is willing to distribute
this.
- Need help from instrumentation companies on timing of pattern key.
This relates to next section below. Greg and Pavel to work with John.
- Define PRBS pattern for component test option (currently PRBS7).
- ACTIONS - John and Tom to work on pattern definition. Should
include lab and sim work. Greg and Pavel to consider needs for timing of
pattern key.
- Trigger block
- Agreement that this should describe only what is required, not the
details of how it is done - may need only a couple of descriptive
sentences.
- Jitter tracking to 4 MHz probably not required, but tracking to
avoid drift is advised.
- ACTIONS - Pavel to draft some words, review by Greg
and Lew.
- Waveforms
- Tom interested in sharing some ClariPhy waveforms for purposes
of corroborating EDC simulation.
- Others should capture waveforms for penalty analysis that
represent range that they would like to consider for LRM. I encourage we
"stretch the envelope", as EDC brings a paradigm shift, and we're still
learning what is good or not good.
- Waveforms can be simulated or experimental.
- Lew agreed to capture waveforms, but Tom action to be very specific
about parameters and details.
- ACTIONS - Tom to provide waveform acquisition details to
others. All to capture waveforms and distribute for penalty
analysis.
- Corroboration of EDC simulation
- Who else has capability?
- Martin interested in comparing results, but first action is to agree
on definitions, references, architecture, and other basic approach and
assumptions. Tom will attempt to document what ClariPhy is doing.
- Suggested that ClariPhy make its code available to small group to
validate; once agreed, publish to wider group. Tom to check with
ClariPhy on timing. Goal to get something before Vancouver meeting.
- ACTIONS - Tom to document definitions, references,
architecture, and other basic approach and assumptions. Tom to check
with ClariPhy on timing for sharing/review of code.
- Noise and jitter control (review of attachment sent in 12/1 TL email,
sent 21:44)
- RIN and SNR are redundant, although SNR may have better chance in
system environment. Not resolved.
- General consensus around measurement of RMS jitter. Pavel also
proposes a pk-pk limit with a specific sample count. Pavel to propose
details and justification.
- Proposal to use same pattern (instead of or an option to square
wave) as for waveform capture, but run scope in different modes (such as
persistence mode on a specific edge or edges. Tek and Agilent (Greg) to
discuss this and make recommendations.
- Trigger recovery block also needs description for this test. Should
include 4 MHz tracking to reduce effects of low frequency jitter.
- ACTIONS - Pavel to propose and justify
pk-pk jitter metric. Pavel and Greg to propose mode for scope and
pattern. Not discussed, but the trigger recovery block for this test
also needs description - can the same team as above (Greg, Pavel, Lew)
please tackle this one too?
- Statistical quality method (as introduced in dawe_1_1104)
- Several interested in studying this to determine if it has value and
if so, how to implement it.
- No specific actions from call.
- All HW quality method
- Many still concerned that there would be way too much variability;
also, given that we can't even seem to settle on a definition of PIE-D,
how could we ever agree?
- Feasibility is a question for EDC suppliers.
- No specific actions from call.
Next call 12/9, same coordinates
- I will be traveling and may have little time for prep for the call, so
hopefully others can make progress on the actions and report on the
next call. Please let me know in advance of any agenda
items.
|