Re: [10GMMF] TP3, 3.6dB ISI penalty criteria
This question Robert Lingle asks of why some fibers meeting the FDDI
spec of 500MHz.km were excluded from consideration from the sieve was
raised during Sudeep's presentation in Vancouver bhoja_1_0105. I don't
think it is justified to exclude these fibers. I agree with Robert that
we need to understand whether including these fibers affects the
results, but I worry about excluding them in any case because it sets a
precedent.
The rationale for considering this needs to be better documented.
The IEEE link model assumes a correlation between bandwidth and ISI
corresponding to a Gaussian pulse. This is one of the few known
weaknesses of the link model (which is an extremely useful tool),
because pulse splitting due to intermodal dispersion can result in
non-Gaussian pulses. In the TIA OM3 specification development we treated
bandwidth and ISI separately with more detailed models to address this
issue. The TIA modeling showed that the -3dB BW does not correlate with
ISI. This was also seen in the early 802.3aq modeling by Phyworks and
the two cases are summarized in a Portland presentation
abbott_2_0704.pdf slide 8 of 14.
There has been discussion of simple alternate BW metrics which give a
better correlation with ISI -- the TIA OM3 work suggests using the 1.5dB
BW. I presented a paper at the NIST SOFM in Sept 2004 suggesting that
measuring the amplitude of the transfer function |H(f)| at a fixed
frequency f_bit approximately 1/2 the bit rate would would give an
improved correlation. I went over these ideas in my Portland
presentation abbott_3_0704.pdf
It was firmly established during the 1GbE development that the installed
base of MM fiber often showed significant pulse splitting particularly
with center launches.
The potential issue for the TP3 sieve is that these are exactly the
issues which EDC needs to address.
John A.
-----Original Message-----
From: Lingle, Jr, Robert (Robert) [mailto:rlingle@OFSOPTICS.COM]
Sent: Friday, February 04, 2005 12:07 AM
To: STDS-802-3-10GMMF@LISTSERV.IEEE.ORG
Subject: [10GMMF] TP3, 3.6dB ISI penalty criteria
All,
I am embarrassed to say that I missed the fact in Sudeep's 12/20 TP3
presentation, there was added "to the sieve" a requirement that each IPR
meet a 3.6dB ISI penalty screening criterion (based on ewen_1_1104). I
didn't realize it till looking at the Vancouver presentation,
bhoja_1_0105,
page 2.
Sudeep,
Would it be too much trouble to produce plots showing which IPR's were
not
included in the pre-, post-, and quasi-symmetric cases due to that
criteria?
Robert
Robert Lingle, Jr
Fiber Design and Development
OFS R&D, Atlanta, GA
-----Original Message-----
From: Jim McVey [mailto:jim.mcvey@FINISAR.COM]
Sent: Monday, January 31, 2005 7:00 PM
To: STDS-802-3-10GMMF@LISTSERV.IEEE.ORG
Subject: [10GMMF] TP3 No meeting notice: February 1, 2005 No meeting
Hello all TP3 people,
Based on conversations with various people at the Vancouver meeting, I
am
proposing that we discontinue the weekly TP3 call until the work
required by
our sub-group requires further calls.
At Vancouver, we completed the receiver test sections with the simple
and
comprehensive stressed receiver tests and the receiver jitter tolerance
test. We set parameters in support of these tests. I know that there
is
some continuing e-mail traffic about these parameters, but I feel that
it
has not coalesced yet. I would suggest that we use the e-mail reflector,
work on studies and presentations in support of aspects of the
discussions,
and when ready move for the TP3 conference call to happen again.
If you feel strongly about the call now, or think it is required in the
future, then please send me your agenda items.
Thank you all of you for your help in making so much progress in
advancing
the 802.3aq standard.
- Jim McVey
+ 1 650 740 7732