In the May 2001 Electrical Contractor, my article mentioned that a high laser-bandwidth 50-micron fiber was being considered for use for LANs less than 300 meters in length. So where does it stand, now that five months has past?
The industry name “850 laser optimized 50/125 mm fiber” has been agreed upon. And, if that is not enough of a mouthful, its official standard reference will be, “TIA/EIA-492AAAC,” which will be the detail fiber specification number once it is balloted and approved.
Some manufacturers’ press releases have reported on “hero” experiments using pre-standard fiber and 10 Gbps transceivers operating in excess of 300 meters, which causes one to ask, “What is the holdup?”
There is a world of difference between demonstrating technical feasibility based on a controlled experiment and having testing procedures and specifications in place for both the fiber and the transceivers, which will ensure interoperability and performance in the industry across multiple vendors.
As in all fiber applications the system performance is dependent upon both the transceiver and fiber characteristics and specifications. While these issues have been resolved for speeds up to 1 Gbps, the need to have well-defined and controlled test procedures and specifications to support 10 Gbps becomes critical.
The process has been complicated because of the “give and take” approach needed between the transceiver and fiber specifications. The wider the transceiver specifications, the tighter the fiber specifications have to be to support 10 Gbps over 300 meters and also vice-versa; therefore, the optimum middle ground has to be found between the two components.
As the draft specifications accurately state, “When a multimode fiber is used with laser sources, its bandwidth may vary, depending on the details of the modal structure of the lasers, of the modal delay structure of the fiber, and of the coupling between the laser and the fiber modes.”
Progress has been made in providing these test procedures and specifications; however, additional work must be done and final agreement reached between TIA and IEEE, and with the international standards organizations of IEC and ISO.
One of the most critical areas of agreement needed is the launch condition of the transmitter. Until this was completed, little could be done to finalize the test procedures for the fiber, because they had to take into consideration the launch conditions that would be used with the fiber.
With the publication of TIA FOTP 203, Launched Power Distribution Measurement Procedure for Graded-Index Multimode Fiber Transmitters, specifying the use of encircled flux to measure launch power, the door was opened to establish the fiber test procedures.
TIA FOTP 220 “Differential Mode Delay Measurement of Multimode Fiber in the Time Domain,” establishes these test procedures, has been approved, and is in the final editing stage.
With these two critical test procedures in place, the issue remains to specify the required performance of the transceivers using FOTP 203 and the fiber using FOTP 220 that will allow for a 10 Gbps system to operate for 300 meters. Present proposals for the transmitter specifications using encircled flux (FOTP 203) would have less than 30 percent of the launch power within the radius of 4.5mm and greater than 86 percent of the launch power within a radius of 19mm.
In laymen’s terms, this specification ensures that the launch power is distributed relatively equally over a launch spot the size of 23mm and is not overly concentrated in the center of the fiber.
Proposals for fiber differential mode delay (DMD) are a little more complicated in that the required DMD varies based on the radial offset, distance off center in microns, that is used when the measurement is made.
Present proposals allow for six different combinations (DMD templates) based on radial offset and required resulting DMD. Rest assured that significant and careful studies using fiber and transmitters from several different manufacturers, as well as extensive simulations, have been completed showing that, when you combine transmitters and fiber to meet the proposed specifications, a minimum effective modal bandwidth-length product of 2,000 MHz•km will be achieved, which is what is needed to support 10 Gbps over 300 meters.
So what remains—finally, agreement among the various standards organizations and manufacturers? Basically, IEEE has reach agreement on the encircled flux specification for a 10 Gbps transceiver and TIA has to reach agreement on the DMD templates, which would be published in the fiber detail specification, TIA/EIA-492AAAC.
Then almost certainly TIA TR-42.8 would approve an amendment to TIA/EIA-568-B.3 to recognize the specifications for 850 laser-optimized 50/125mm fiber cable and TIA TR-42.1 would approve an amendment to add this fiber the recognized list of media for premises cabling systems in TIA/EIA-568-B.1.
All of the above is not expected to be completed until March 2002 at the earliest and probably not until mid-2002. Until then, users are advised to consider using standard-based multimode fiber, either 62.5 or 50 mm, with the understanding that IEEE is also working on wide wavelength division multiplexed (WWDM) solutions that should allow users to use 10 Gbps Ethernet over the installed base of optical fiber for 300 meters.
BEAM is director of systems marketing at AMP NETCONNECT Systems. He can be reached at (336) 727-5784 or firstname.lastname@example.org.