[802.3ae_Serial] From Serial PMD 6 Sept: adoption of new link model, J cal, MPN,
Mode partition noise over new MMF: link model vs. reality
---------------------------------------------------------
The predicted MPN, about 3/4 dB plus knock-on penalties, is enough to be
concerning. However, this is based on a continuous spectrum analysis with
MPN k(OMA) = 0.5. Actually, VCSEL modes will be separated by O(0.5 um) so
the spectrum is very far from Gaussian, with typically 3 modes in operation.
This may mean that the MPN would be better represented by a different
(lower?) k value. It would be good to see some experimental work on the
subject, in Copenhagen, to allow us to choose a good MPN k(OMA) number.
Proposed adoption of new link model
-----------------------------------
I intend to propose that we adopt
http://www.ieee802.org/3/ae/public/adhoc/serial_pmd/documents/10GEPBud3_1_14
.xls as our standard model. As I wrote on Fri 17 Aug. 2001 "Updated link
models"
http://www.ieee802.org/3/10G_study/public/serial_adhoc/email/msg00444.html ,
it has several features that our "official" model 2.4.1 lacks and which we
need for our work. There was concern that, with not everyone able to attend
in Copenhagen, we might not be able to have a thorough debate on this.
However, the new model has definite benefits and we have not yet found any
significant demerits, so approving it is not likely to be a retrograde step.
The new link model seemed a little pessimistic. Petar suggested I compare
it with Gigabit numbers.
[After the meeting, I compared the new model with the 1GE model for 400
MHz.km, 500m, 1.25 GBd. Using the 1GE numbers for every variable, the
result was just 0.15 dB more optimistic at eye centre, and 2.3 more
pessimistic at the eye corners (which aren't part of the standard any more).
So, the biggest difference is in accounting for jitter and mistiming, which
are real enough effects.]
Jitter measurement calibration
------------------------------
Opinion that jitter calibration was very hard and not very reproducible (at
a different location). greg_lecheminant@xxxxxxxxxxx, an expert in scopes
and BERTs, offered to take an interest in the measurement issues involved.
Next phone conference
---------------------
On 11 Sept, back to the usual Tuesday slot:
15:15 GMT = 4:15 p.m. BST = 17:15 CET = 11:15 a.m. EDT = 8:15 a.m.
PDT, Tuesday
+1(816)650-0631 Access code 39209
The following week, many of us will be in Copenhagen, so no phone conference
then.
Piers