Dear P802.3DJ Task Force colleagues,
During last week’s meeting I raised some questions on Roberto’s presentation (https://www.ieee802.org/3/dj/public/23_07/rodes_3dj_01a_2307.pdf) regarding the
mentioning of max mean DGD.
After checking with Pete Anslow and Mark Kimber I would like to provide the following clarification between “Max mean DGD” in the Table for Transmitter compliance channel specifications (e.g. Table 151-12) and “DGD_max” in the Table for
Fiber optic cabling (channel) characteristics (e.g. Table 151-13).
The examples mentioned (clause 151 for 400GBASE-FR4 and 400GBASE-LR4-6) provide the following specifications.
The characteristics of this compliance channel are created specifically for the TDECQ test (or before for the TDP test) of a transmitter. In such a test we want to measure impairments due to transmitter effects associated with ISI and Chromatic
dispersion and as such it is important that any effects due to PMD are minimized so that the associated penalty is negligible. The penalty due to PMD/DGD is captured separately as a contribution to “Allocation for penalties (for maximum TDECQ)”, in the same
way as a penalty due to MPI.
Therefore here you will see a requirement for 0.8 Max mean DGD, which is the same for 2 km and 6 km. In Clauses where PMDs are provided for 40 km (e.g. Clause 122) this Max mean DGD is also 0.8 ps.
The Table for Fiber optic cabling characteristics define the worst case channel over which a PMD needs to operate with sufficiently low BER.
In this Table you will see a value DGD_max, which is depending on the maximum transmission distance.
Based upon the analysis in Pete Anslow’s presentation
https://www.ieee802.org/3/cu/public/cu_adhoc/cu_archive/anslow_3cu_adhoc_051519.pdf. The “usual” value of 8 ps for 10 km links, was reduced to 5 ps for 400GBASE-LR4. Since the maximum distance was reduced from 10 km to 6 km the DGD_max was reduced from
5 ps to 4 ps.
There is no relation between the Max mean DGD of 0.8 ps for the TDECQ compliance channel and a value for DGD_max in the fiber optical cabling (channel) characteristics.
In the case of 800GBASE-LR4 it may be appropriate to agree on a Max mean DGD value of less than 0.8 ps, as Roberto Rodes indicated in an offline discussion, because 0.8 ps may not provide a sufficiently low DGD penalty.
For the 800GBASE-LR4 specification the TF will need to agree on a value for DGD_max in the channel specification.
Kind regards,
Peter Stassar
To unsubscribe from the STDS-802-3-B400G list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-B400G&A=1