From serial PMD call: new open issues list
Here is the list of topics raised by comment resolution for the serial ad
hoc to work on, with some additional input from Tuesday's call. This list
needs to be checked against the comment database to see if we have
overlooked anything.
Jitter spec numbers
-------------------
Need refined values for DJ and RJ. We believe that the RIN we have allowed
exceeds the RJ we have allowed (to be checked). Therefore, need to tighten
RIN, or relax RJ and tighten DJ. CDR experts are concerned that not enough
is left at the receiver, after total jitter. RIN penalty in the current
spreadsheet is somewhat underestimated.
Test fibres for 1310 dispersion penalty measurement
---------------------------------------------------
Necessary fibres may not be available. Draw out truth table for when
special fibre is needed?
Link attenuation
----------------
The proposed attenuation used to be 7 dB, made up of 0.5 dB/km + 2 dB for up
to 4 connectors. A value of 6.5 dB has crept into the draft, creating an
inconsistency; we think this came about by not having clear linkage of
connector loss to PMD type and was not intentional. However, GR-253-3 has
0.55 dB/km including connectors, IEC 60793-2? has 0.4 dB excluding
connectors, TIA 568B.3 has 0.5 dB ... We could review what we need to
achieve the 10 km objective. Juergen is going to an ITU-T meeting next
week; can ask the operators what attenuation they envisage. Metro
attenuation is generally worse than long haul.
Items from Tuesday's preliminary list
-------------------------------------
Review and refine 1310 serial risetime removal proposal.
850 nm risetime removal proposal - propose harder mask to represent faster
risetime.
Review table values generally?
Review table values around stressed and nominal receive sensitivity and
receiver eye opening penalty.
Find out whether/why we need an optical return loss (ORL) specification as
ITU-T has, as well as a discrete reflectance specification, for the cable
plant. This one needs input from cabling experts.
Tx maximum power and Rx overload, 1310 serial.
Work with Ben Brown's Test Pattern team with choosing test patterns.
Technical feasibility
One item where demonstrated feasibility, soon, would be desirable is
in trying out the jitter test methodology. Relates to test patterns but
does not depend on knowing the chosen patterns.
Piers