The macro-bend performance is defined here as the induced
attenuation at a certain number of full 360 degree turns around a
specific bend diameter usually stated in dB/turn. (See Figure 2). It is also important to note that the macro-bend loss
depends on the specific operating wavelength and increases for longer
wavelengths. For the users of this standard, this leaves open a variety
of combinations that makes it hard to compare the fiber specifications.
It seems intuitive, though, when comparing bend performance to use the
most common and most bend-sensitive operating wavelength of 1550 nm as
well as a bend diameter that best reflects the targeted applications.

Figure 3. Bend Performance.

The performance for most of these bend-tolerant fibers is still
limited, as their fiber design is typically based on legacy processes
for forming refractive-index profiles of single-mode fibers. Figure 3 also depicts the typical bend performance range of
Corning's SMF-28e XB fiber. While specified as a G.657.A bend-improved
fiber, its bend performance clearly exceeds the requirements for
next-generation G.657.B at those bend diameters used in most of today's
hardware and cabling applications. This maximizes the performance of
today's generation of FTTH products.

Again, you will notice that bend-improved fibers have a performance
advantage over the latest standard releases (G.657.B) when used with
today's legacy hardware components such as terminals, closures,
cabinets, pedestals etc., that have bend diameters in the 20 to 30 mm
range.

More important, in order to enable hardware components and fiber
optic cables for the challenging environments of today's emerging
applications without increasing system loss, another level of bend
insensitivity is required. Why do you need this extra bend performance
beyond G.657.B? The most important reason, of course, is the specific
application.

While wrapping fiber cables around a pencil or tying a knot in the
fiber cable are impressive demonstrations, the primary and real
concerns of most installers tend to fall into two categories: stapling
of the cable, and discrete small-radius turns - both techniques used
for installing copper cables. Figure 4 shows an indoor
optical drop cable being stapled to the studs similar to an MDU
installation. The mechanical stress is compensated for by the cable
design, but the bend is still incredibly small – much smaller than
G.657.B requires as illustrated in the X-ray picture shown in Figure 5. Having a series of these staples on a cable without a
truly bend-insensitive fiber in such a drop cable can easily add up to
several dB of incremental loss in the drop section alone.

Talk About Flexibility
There are two issues to consider when planning for deployments that will bend a fiber to the new levels we’ve discussed.

Issue 1: How important is backward compatibility for
installers and network planners? Many commercially available fibers
performing at the required bend diameters below 15 mm are also not
compatible with standard single-mode fibers (ITU-T G.652) and involve
significant complexity to prepare for connectorization/termination. A
new technology is required to overcome this obstacle while achieving
bend insensitivity in the range illustrated in Figure 2.

Issue 2: The next consideration for installers is fiber
mechanical reliability at these relatively small bend diameters. For a
fiber that has undergone the standard proof test stress of 100 kpsi
during manufacture, it is commonly accepted that for bend diameters
greater than 64 mm, the axial stress experienced by the fiber will be
lower than the safe stress level (taken as 1/5 of the proof test
stress), and an operator can be confident that the fiber will
experience no failures due to applied stress in a 25-year lifetime. For
a normal end-to-end transmission line installation of optical fiber,
this 64 mm bend diameter limit is rarely exceeded. However, for FTTx
network installations, the density of fiber ports in service terminals
and/or the nature of indoor installation practices can drive bending of
the fiber well below the 64 mm diameter limit, hence the smaller bend
diameters specified in the G.657 standard. When the safe stress level
is exceeded in network deployments, the fiber moves into a different
regime of reliability consideration where the probability of failure
during the fiber's lifetime becomes finite (non-zero).

Fortunately, the probability of failure is extremely low, even when
subjected to extremely tight bend diameters of the order of 15 mm. For
more than 30 years, detailed research into fiber reliability in tight
bend scenarios has shown that a standard single-mode fiber when
subjected to a 360 degree turn of 15 mm diameter will have only 1 ppm
(one-in-a-million probability) of failure over a 25-year lifetime.

In summary, when choosing the right fiber type for applications
requiring small bend diameter, it is important to differentiate between
existing and new applications. For bend diameters above 20 mm,
state-of-the-art bend-improved fibers can exceed even the G.657.B
standard for bend-tolerant fibers while guaranteeing full backwards
compat-ibility. However, to solve installers problems in tough
environments such as MDUs where the bend diameters can go as low as 10
to 15 mm, a new technology for fiber, cable and hardware is required to
ensure truly bend-insensitive performance while maintaining a high
degree of reliability and enable backwards compatibility.