An observer at the November standards meeting of Working Group TR-41.8.1 of the Telecommunications Industry Association (TIA--Arlington, VA) would likely have left the quarterly session optimistic that the group would soon choose a duplex fiber-optic connector the size of an 8-pin modular connector (or of an RJ-45) to include in its next revision of the Commercial Building Telecommunications Cabling Standard, the current version of which is TIA/EIA-568A. Selecting a small-form-factor fiber-optic connector, after all, has been a major mission of the working group for the last year.
The week-long February meeting opened with the MT-RJ connector, a joint product of a consortium of AMP Inc. (Harrisburg, PA), Siecor Corp. (Hickory, NC), Hewlett-Packard Co. (San Jose, CA), USConec (hickory, NC), and Fujikura (Atlanta, GA), being the front-runner in a competition that included four other vendors and their products--IBM (White Plains, NY), Siecor, and Siemens (Berlin, Germany) with the SCDC/SCQC, Panduit Corp. (Tinley Park, IL) with the OPTIJACK, Lucent Technologies (Murray Hill, NJ) with its OptiSPEED LC, and 3M Telecom Systems Div. (Austin, TX) with the Volition VF-45. After a long day of task group meetings, the AMP delegation rallied for a 10:00 PM meeting to go over its presentation, scheduled to be given the following day before the entire Working Group TR-41.8.1 This meeting would determine whether or not the MT-RJ would be included in the next revision of the commercial building wiring standard. And as the AMP team rehearsed, some of the fiber-optic connector manufacturers whose products had not topped the list in November were lobbying for support to kill the MT-RJ connector as a TIA standard.
The February 25 meeting started at 11 AM with a presentation by Donna Ballast, the standards representatives for BICSI (Tampa, FL). Ballast discussed a BICSI survey in which 761 people in the cabling industry voiced their opinions about the need for another, smaller fiber-optic connector in a standard that already recommends the SC connector. The bottom lime, once the votes were tabulated, was that 66% of the people surveyed wanted to add a small-form-factor fiber-optic connector. However, 66% of the TR-41.8.1 committee had to vote for the MT-RJ in order to ratify this type of connector as a standard recommendation.
Tony Bean of AMP presented the MT-RJ connector to the committee, spending half an hour on a formal presentation and then opening the floor for questions. More than a dozen questions were asked most of them dealing with patents and licensing. The major concern was the mechanical-splicing mechanism of the MT-RJ, the patent rights to which AMP had not offered to other manufacturers. This proprietary mechanism joins the horizontal fiber cable to the MT-RJ, which is then connected to the work-are outlet. Without patent rights to this mechanism, it was asked, how could another manufacturer make this type of connector, even after a license agreement was in place? AMP replied that the company was not going to keep other manufacturers from developing their own splicing mechanisms for the MT-RJ connector. In fact, it was noted that several manufacturers already had their own splicing processes; these manufacturers could continue to use their own methods without infringing on AMP's patent.
Another issue that arose was a letter, circulated to the committee and incorporated into the minutes, in which a contractor complained that he could not get MT-RJ connectors from AMP. The company responded that it had set up a technical-support team to field just such questions--and particularly questions about the MT-RJ. At present, the patch cords that use these connectors are available, but the separate jacks will not be available until the second quarter of 1998. AMP's Beam commented that the company's technical-support team had always been quick to respond to questions from the industry regarding the MT-RJ, and he said he would research the letter circulated at the meeting.
Still another concern was the fact that AMP had only a very small base of installed connectors. However, the technology used to connect the MT-RJ connector has been available for years, according to the company, and so its time-tested. In addition, even though there were not many installations to prove the technology, it was pointed out that a time-tested installed base was not a requirement of the fiber-optic connector competition at this stage of the selection process.
At about noontime, a ballot was handed out to voting members of the working group, and at 5 PM the votes were collected to be tallied. The next morning, the results, including a list of companies participating in the selection process, would be distributed.
At 8 AM on February 26, the moment of the truth was at hand. Its so happened that the working group could not meet in the same room it had occupied the day before, the only other room available that could house the group was the hotel bar. This room, looking like a set for Saturday Night Fever, right down to a large silver disco ball hanging from the ceiling, housed the 90 cabling-industry engineers awaiting the results of the voting. And when the results were announced, the MT-RJ had not made it. The connector received a 59% positive vote, but needed a 66% majority vote to be adapted as a standard.
At that point, the same questions that were asked and the same comments that were made at the November meeting started popping up again. For instance, one manufacturer of hubs and network interface cards said "This will make it hard for end users now. There will be a minimum of five different possible interfaces that can go on these boards, and this will be confusing." An industry consultant added that this decision adds another level of complexity to both network designs and end-user networking solutions.
More to the point are the time and effort wasted in studying and trying to solve the small-form-factor fiber-optic connector problem. A decision to let the market decide among the five connector manufacturers would have made sense a year ago, but since then the TIA working group has invested heavily in this issue. The working group includes representatives of the four-connector manufacturers who lost in the competition, as well as the one that won it. Each manufacturer undoubtedly thought that its product would win.
Fiber-optic task group
The big topic in the discussion of fiber optics in project number PN-3523 was a technical contribution pertaining to backbone and centralized cabling as it relates to emerging networking applications. The central issue is that telecommunications systems bulletin TSB-72, which defines centralized fiber-optic cable administration, has a 300-meter maximum distance limit on cable running from the work-area outlet to the centralized crossconnect, while the Gigabit Ethernet short-wavelength option (1000Base-SX) has a maximum distance of 260 meters.
According to a 1996 survey conducted by Compaq Computer Co. for the IEEE 802.3z Gigabut Ethernet committee of the Institute of Electrical and Electronics Engineers, the maximum transmission distance is supported just as well by a 200-meter limit as by a 300-meter figure? After it came to light that an AT&T-commissioned technical study that provided the data influencing the choice of a 300-meter limit in TBS-72, the working group decided not to lower the 300-meter requirement to 260 meters, even though many participants expressed doubt that 300 meters was a practical length for field installations.
Instead, the PN-3523 Working Group is researching the possibility of specifying optical fiber with a band-width of 200 megahertz per kilometer, a limit set by a factor known as intersymbol interference (ISI). Using this criterion will mean that the supportable distance of 1000Base-SX is limited primarily by the bandwidth will translate directly into reduced ISI and, in turn, will permit longer transmission distance.
US TAG for JTC1/sc25/WG3
Representatives providing liaison with international standards-making groups reported that, with regard to premises wiring standards, it appears that the band width limit for Category 6 and Class E will be raised from 200 to 250 MHz, while for Category 7 and Class F, the limit will be raised from 600 to 750 MHz. This action results from the IEEE's suggestion that an upper-frequency limit at least 25% higher than the specified channel bandwidth be set.
UTP systems task group The PN-3727 Working Group looking unshielded twisted-pair (UTP) issues devoted its time to technical presentations on far-end crosstalk, equal-level far-end crosstalk, and return loss in order to tie down the parameters needed for Draft 6A of "Additional Transmission Performance Specifications for 100-ohm, 4-Pair Category 5 Cabling," as well as for Draft 6B of the enhanced Category 5 cabling document.
It appears that the TR-41.8.1 committee is in a rush to turn these drafts into addenda to the TIA/EIA-568A standard, so that it will be aligned with the Gigabit Ethernet standard now under development. Within the committee there were heated discussions about the FEXT, ELFEXT, and return-loss parameters, and especially about the validity of the measurements values that these parameters must be realistic and achievable in channel specifications.
Residential cabling standard
Out for ballot, with a comment period that expired this past March, is Standards Proposal 3490, the proposed revision of EIA/TIA-570, the Residential and Light Commercial Telecommunications Wiring Standard. If approved, this revision will be published as TIA/EIA-570A.
The American National Standards Institute approved TIA/EIA-569A, the Commercial Building Wiring Standard for Telecommunications Pathways and Spaces, in February. This standard can be purchased from Global Engineering Documents in the United States and Canada by calling (800) 854-7179 and internationally by calling 00111-303-397-7956.