Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [STDS-802-3-400G] worst-case dispersion test bed



Pete,

Yes, I meant G.653. 

Paul

 

From: Anslow, Peter [mailto:panslow@xxxxxxxxx]
Sent: Tuesday, August 01, 2017 12:28 PM
To: Kolesar, Paul <PKOLESAR@xxxxxxxxxxxxx>; STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: RE: worst-case dispersion test bed

 

Paul,

 

Thanks for this.  Where you say “G.953”, this should be “G.653”.  The standard is publicly available at:

http://www.itu.int/rec/T-REC-G.653-201007-I/en

 

The zero dispersion wavelength for these fibers is about 1500 nm, so only negative dispersion can be obtained this way.

 

According to I.2 and Table I.1 of G.653, the typical dispersion is around -17 ps/nm/km at 1309 nm, so the lengths required for the worst case dispersion in the P802.3bs draft are:

200GBASE-FR4 ~ 700 m

400GBAASE-FR8 ~600 m

200GBASE-LR4 ~ 1.7 km

400GBASE-LR8 ~3 km

 

Regards,

Pete Anslow | Senior Standards Advisor
43-51 Worship Street | London, EC2A 2DX, UK
Direct +44 2070 125535
|

 

From: Kolesar, Paul [mailto:PKOLESAR@xxxxxxxxxxxxx]
Sent: 01 August 2017 17:14
To: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: [STDS-802-3-400G] worst-case dispersion test bed

 

On today's SM ad-hoc call, issues regarding measurements with worst-case dispersion were raised.  It is often difficult to obtain worst-case dispersion fiber using the fiber types specified in the 802.3 standard.  Other fiber types can potentially be used for this purpose, such as dispersion-shifted fibers that move the zero-dispersion wavelength (ZDW) near 1550 nm.  When operated near 1310 nm these fibers produce substantial dispersion, but the sign of the dispersion will likely be either only positive or only negative.  (At this point, I have not bothered to calculate which one, and it may depend on the manufacturer and fiber category.)   But a glass half full is better than an empty one. 

 

Some dispersion-shifted fibers have cut-off wavelengths longer than 1300 nm.  These are not assured to be single-mode at the wavelengths of interest.  As a work-around, one can apply mode suppression loops to rid the fiber of the higher order modes after each connection and splice in the test bed.  Another way is to use IEC B2 fibers, a.k.a. ITU G.953 fibers.  These are dispersion shifted, but also offer cabled cut-off wavelength maxima of 1270 nm, so remain single-mode at the wavelengths of interest.  There are two sub-categories, IEC B2_a and B2_b (ITU G.953.A and 953.B), each having different dispersion specs.  The mode field diameter of these fibers is smaller than others, but should still couple adequately to the source. 

 

Obtain the ZDW and ZDS of the fiber sample from the manufacturer.  From these the dispersion at the operating wavelength can be calculated as a function of fiber length.  Select the fiber length that produces the worst-case dispersion for that operating wavelength. 

 

Hope this helps.

 

Paul Kolesar