Thanks for the clarification, Dennis. I assumed that you were talking about multimode, and that's why I raised the issue. Up until very recently, most implementations of super-gigabit rate systems would use singlemode fiber (SMF) for any appreciable distances beyond floor size, and SMF does not characteristically employ the same type of 'bounce' path, as it were, as multimode fiber (MMF).
I should clarify this. Historically, very high speed (super-622 Mb/s, or OC-12) data rates as a rule of thumb beyond 300 meters (and in a limited number of situations up to 2 km over MMF with special laser assmeblies) only took place over SMF. There is a movement afoot at this time, however, to salvage some of the embedded MMF (and introduce new grades of MMF) that has already been installed in campus and in-building situations. These new distances over MMF will be accomplished through the use of certain types of vertical cavity surface-emitting lasers (VCSELs).
At present, existing MMF distribution systems are rated for up to 100 meters, but through the use of VCSELs distances are proposed of up to 300 meters and 850 meters, under some proposals, and depending on the power ratings of those light sources.
These distances are in sharp contrast to some proposed SMF distances for GbE which portend to carry native GbE and 10 GbE up to 40 km.
This topic, i.e., using improved MMF and VCSELS, will gain in importance with the proliferation of GbE, and within the next year or so, 10 GbE as well, when vendors begin making pre-standards boxes available after Draft #1 for 10GbE is issued (which has a target time frame of 1Q2001).
The significance of this topic is rooted in at least two sources of cost savings: (1) attributed to lower splicing costs, since users could potentially reuse existing MMF which only require mechanical splices (definition below), and (2) due to the lower costs of vcsel's compared to traidional lasers now being used to support 10 Gb/s SONET systems, for example.
These cheaper system designs will allow users to more economically support their next generations of Ethernet and other super-gigabit feeds, such as Fibre Channel; longer-distances of IBM's enterprise system connectivity (ESCON); and high resolution video feeds (both analog and digital), by avoiding the costs associated with additional single mode pulls and fusion splicing, ostensibly.
Note: The 850 meter distance is proposed when a single wavelength is used to support serial transmission. Some VCSEL proposals call for four relatively inexpensive wavelength lasers, using wide wdm techniques, each supporting 2.5 Gb/s. Such a scheme would allow for much greater overall distances which are, as yet, unspecified at a mere fraction of the costs associated with traditional SONET-grade OC-192 laser assemblies.
------
re: splicing, you asked with a :-) ,
" I assume it is not done by butting the fibers up against each other :-) Some form of lenses are used, no?"
no... lenses are almost never used. I say almost because some situations call for mode conversion, or the introduction of passive wdm devices, etc., which do sometimes introduce corrective optics on the interfaces of adjoining sections of fiber, but in the vast majority of plain old ordinary optic fiber splices, they are, as you half-suggested, simply butted up against one another in a mechanical fashion. These are actually called mechanical splices, as opposed to more involved forms of heat-treated splices which are defined as fusion splices. See definitions below.
Increasingly, field splicing techniques are leaning towards tight-tolerance mechanical methods due to improved economics, engendered by simplicity and speed. "Gang" splices, for example, are now achievable in this fashion whereby dozens of fibers are spliced together at once with the actuation of a lever.
As a general rule, however, I've found that mechanical splices are more commonly accepted and used in-house and campus situations where MMF dominates, and fusion forms of splicing still take place "out in the street" by the xLECs whose fiber type domains are almost exclusively SMF. In a recent campus design that I was involved with we specified one third SMF and two thirds MMF for the inter-building runs.
The MMF's were mechanically spliced by campus staff, and the SMFs were done fusion style by some BEL students who were on campus. They did it gratis, I might add, which saved the school a bundle in training and rental fees. (I felt I had to say something nice about BEL after my tirade of two weeks ago when they...)
Sections 5 and 6 in the following Corning tutorial at www.IEC.org provide some basic information on splicing parameters.
webproforum.com
A comprehensive treatment on the physical attributes of optical fiber is given in an on-line tutorial by Siecor, probably one of the best tutorials on this subject on the 'net, IMO, at:
siecor.com
From the Siecor optical fiber glossary at siecor.com :
Mechanical Splicing: Joining two fibers together by permanent or temporary mechanical means (vs. fusion splicing or connectors) to enable a continuous signal. The CamSplice <tm> is a good example of a mechanical splice.
The mechanical splice device:
siecor.com
-----
Fusion Splicing: A permanent joint produced by the application of localized heat sufficient to fuse or melt the ends of the optical fiber, forming a continuous single fiber.
Demo of fusion splice:
siecor.com
Comments and corrections welcome.
FAC |
