|
Dial-in details
-
Date: Thurs, Mar 31, 05
-
Time: 9:00 AM Pacific
-
Duration: 1:30 max
-
Number: 401-694-1515
-
Access code: 421721#
Proposed agenda
Tom
----- Original Message -----
Sent: Friday, March 25, 2005 5:45
PM
Subject: [10GMMF] Minutes for TP2 call
3/24/05
Attendance (no order)
- John Ewen
- Ernie Bergmann
- Mike Dudek
- Greg Lecheminant
- Piers Dawe
- Joe Gwinn
- David Cunningham
- Vivek Telang
- Lew Aronson
- Tom Lindsay
- Jan Peeters Weem
- John Jaeger
- Matt Traverso
Agenda
Discussions
- Uploading waveforms
- All waveforms should be
- (re)sampled to 16 samples per UI average.
- exactly one pattern repetition long; for example, a prbs
2^9-1 waveform would include 16*511 samples.
- time-aligned to the binary sequence as given on the web; for
example, start with a run of 9 ones for 2^9-1.
- exact alignment is not critical, but should be within a fraction a
bit.
- Waveforms should be in actual optical power units as from TP2.
- OMA and SteadyStateZero values must be included.
- This will promote common format and simplify usage.
- Simulation results
- See http://grouper.ieee.org/groups/802/3/aq/public/upload/TP2measuredwaveformpenalties_finiteEQsTP3D1.1stressors.pdf.
- Vivek will try to reproduce results (I have since sent a few
pre-processed waveforms to Vivek)
- Preliminary observations
- Measured waveforms, although having lower penalties than the
Gaussian waveform for very long equalizers, often show higher penalties
for shorter EQs.
- The increases are fairly small (a few tenths of a dB), but then most
of the waveforms analyzed are "good" eyes; we need more results from a
wider range of waveform characteristics.
- 2 pre-emphasized eyes showed reduced penalties (relative to
Gaussian) in almost all EQ configurations.
- Penalties are not always monotonic with EQ length (an explanation
has been subsequently sent to the reflector).
- The Gaussian waveform settles with 3 or more feedback taps; in
several cases, measured waveforms benefit from more than 3 FB
taps.
- Future work
- More waveforms (see above).
- Try new IPRs, at least pre-cursor, developed for matching coverage
across range of EQs
- Replot penalties vs. PIE-D value for each IPR.
- Other suggestions and analyses encouraged.
- Eventually - assess impact, if any, to budget, TP3 stresses, TP2
limits.
- Other WG priorities and tasks
- Back reflection method and spec
- Mike and Piers to complete proposal.
- Leaning towards using MMF and -12 dB spec, but details still need
work.
- Hope to review in a couple of weeks.
- AC coupling for RIN
- Tom agreed to work this. I will try to float a completed comment in
1-2 meetings.
- AC coupling should be straightforward with frequency spectrum method
(elevated to primary status in Atlanta); due to DC coupling, scope
method users may overestimate value.
- Plan is to use 1 MHz.
- Jitter waveform (measurement example)
- No real action required. Tom created waveform for D1.1, but after
deadline, so change not accepted.
- I will send to group for review.
- Jitter magnitude
- No current motion to change spec or method.
- TWDP editorial changes
- Comment 96 (Tom) has most of editorial changes required - submit
next time.
- Suggestion to break up comments - accepted.
- TWDP technical
- Focus - track work above with measured waveforms, pre-emphasis, etc.
- OMA definition - related to previous line. In current TWDP approach,
OMA has 1:1 effect on penalty results.
- Test patterns
- Tom needs to write up what is required - action for next
call.
- This task does not relate to prbs9 patterns, but to in-service
testing.
- Eye mask
- Suggestion to eliminate mask, but any changes to the mask should
follow TWDP and measured waveforms work.
- Would like to eliminate "beauty contest", but if LRM does not
provide a mask, then customer will create his own which could be a worse
situation.
- Need peak power limit to protect Rx overload? (carryover topic,
still need inputs).
Next call
Date: Thurs, Mar 31, 05 (regular
day/time)
Time: 9:00 AM Pacific
Duration: 1:30 max
Number: 401-694-1515
Access code:
421721#
Thanks,
Tom Lindsay ClariPhy Communications tom.lindsay@clariphy.comphone:
(425) 608-0209 or (949) 480-9210 cell: (206) 790-3240 fax: (425)
608-0232
----- Original Message -----
Sent: Wednesday, March 23, 2005 12:40
PM
Subject: Re: [10GMMF] TP2 call
3/24/05
All - here is a proposed agenda for tomorrow's call.
- Attendance
- Set agenda
- Discussions
- Preliminary simulation results of measured waveform penalties vs. EQ
lengths for three D1.1 TP3 channels
- Format for uploaded measured waveforms (see below)
- Prioritize, assign WG task list (see below)
- Other proposed topics?
- Next call
Measured waveforms - before uploading, can each waveform be
pre-processed, where pre-processing includes resampling to #nn samples per
UI, truncated, and time-aligned? If so,
- Can suppliers of waveforms they take responsibility for their
pre-processing?
- Set #nn = 16.
- Can the waveform be aligned to the pattern or vice versa?
- Noise
95 (back reflection spec or method/setup)
- 98 (AC coupling?)
Jitter
TWDP
-
96 (editorial updates to code)
-
146 (channels, EQ lengths, limits)
83 (OMA definitions – critical)
Test patterns (no comment #)
Mask
Others?
Tom
----- Original Message -----
Sent: Wednesday, March 23, 2005 8:43
AM
Subject: [10GMMF] TP2 call
3/24/05
Call details:
Date: Thurs, Mar 24, 05 (regular
day/time)
Time: 9:00 AM Pacific
Duration: 1:30 max
Number: 401-694-1515
Access code:
421721#
Thanks,
Tom Lindsay ClariPhy Communications tom.lindsay@clariphy.comphone:
(425) 608-0209 or (949) 480-9210 cell: (206) 790-3240 fax: (425)
608-0232
|
