Note: Descriptions are shown in the official language in which they were submitted.
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
Method for Connecting Optical Fibers and the
Interconnection
Background of the Invention
Field of the Invention
This invention relates to an improved method for
the interconnection of optical fibers and in one aspect
relates to a new "dry," i.e. gel-less, mechanical optical
fiber interconnection where the fiber ends are optically
aligned and are pressed together and maintained under axial
compression by the splicing element.
Description of the Prior Art
Optical fiber splices are well known and the art
is becoming crowded with elements for aligning the fiber
ends optically and holding the same in alignment. Cleaved
fiber ends are used in most mechanical splices currently
available. These splices contain a coupling medium,
usually a gel or oil, that has the same index of refraction
as the core of the fiber. This index matching material is
used to fill the gap between the pair of fiber end faces
which are to be spliced.
Prior art showing a splice element of the type
used in the present invention is disclosed in USA patents
Nos. 4,824,197 and 5,159,653.
Mechanical splices all contain the index matching
gel materials. Some mechanical splices have a problem with
temperature cycling due to the index of refraction of these
materials changing at varying temperatures, which result in
fluctuations of the optical signal, mainly return loss
increases. Therefore, the mechanical splices, not using
angled cleaving on the fiber ends, presented a problem,
first, in not always meeting temperature cycling
specifications and, second, needing the index matching
materials. However, the mechanical splices are easier for
-1-
CA 02247448 1998-08-25
WO 97/36200 PCTIUS97/02812
the technician in the field to complete. The improvement
of the mechanical splices in a manner to render them
acceptable and readily comparable to the fusion splices is
contemplated by the present invention. Existing techniques
for preparing the fiber ends for connection are employed in
the present invention to produce mechanical, dry, i.e.
gelless splices.
Summary of the Invention
The present invention provides a new and improved
process for splicing optical fibers. The process comprises
the steps of preparing the ends of the fibers to be spliced
for intimate axial compressive contact between the cores,
entering the fiber ends into the opposite ends of a fiber
passageway in a fiber splice element, and placing a
compressive force at the interface of the fiber ends to
retain intimate axial contact of the fiber cores throughout
temperature cycling between 0 C and 40 C. This compressive
contact is afforded by applying axial compressive forces
onto the fibers when in the splice element passageway prior
to actuation of the element, placing a stress onto the
splice element prior to actuation of the element such that
when the stress is removed the splice element will apply an
axial compressive force on the fiber ends, or applying
deforming pressure to the element after actuation to apply
a compressive stress to the ends of the optical fibers to
maintain them in intimate axial compressive contact
throughout the desired temperature ranges. Stress can be
applied to the splice element by heating the element,
bending or stretching the element along the fiber
passageway prior to the insertion of the fiber ends into
the splice element and actuating the same. Also, the
compressive force to the fiber end interface may result by
applying a force to deform the element and forcing the
fiber ends into intimate pressure contact. Further,
combinations of these procedures are also contemplated,
-2-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
i.e. heating the splice element and placing the fiber ends
in the element under spring or mechanically applied
compression contact.
Heating the metallic element to a temperature
between 100 and 120 C prior to actuation of the splice
element, which actuation serves to align and bind the fiber
ends and effectively holds the fiber ends in aligned,
intimate axial pressure contact with each other when the
splice is returned to room temperature. The amount of heat
applied exceeds the temperature used in the temperature
cycling tests of the Bellcore specification, as used
generally industry wide and published by Bellcore (Bell
Communications Research, Morristown, NJ) Document TA-NWT-
000765 and GR-765.
After placing the fiber ends in intimate contact
in the splice element of a FibrlokTM Splice, force can be
applied to the element to distort or stress the material
and cause the material to apply compressive stress on the
fibers to force them into compressive contact.
The interconnection of the present invention is
an optical splice between two single mode optical fibers
comprising a splice element having a longitudinal
passageway for receiving the ends of the fibers approximate
the midpoint of the passageway, and the fiber ends being
placed in axial compression against each other free of any
index matching material.
The splice of the present invention contemplates
splicing two optical fiber ends which have been positioned
and clamped within a metal splice element. Through one of
a variety of methods, the fiber ends are placed and held in
optically aligned intimate axial compression against each
other. The splice of the present invention is free of any
index matching material.
-3-
CA 02247448 2006-03-20
60557-5918
An aspect of the invention provides a process for
using a fiber splice element for splicing optical fibers
having light transmitting cores wherein the fiber splice
element has a coefficient of thermal expansion that is
greater than that of the fibers being spliced comprising the
steps of: a. preparing the ends of the fibers to be spliced
in abutting relationship, b. entering a fiber end into each
of the opposite ends of a fiber alignment passageway in a
fiber splice element until the fiber ends are in contact
with each other, c. applying an axial compressive force by
heating the fiber splice element above 80 degrees C,
actuating the splice element onto the fiber ends, and
cooling the fiber splice element, and d. maintaining the
axial compressive force at the interface of said fiber ends
in the element to retain intimate contact of the fiber cores
throughout temperature cycling between 0 degrees and 40
degrees C.
Another aspect of the invention provides a process
for using a fiber splice element for splicing optical fibers
wherein the fiber splice element has a coefficient of
thermal expansion that is greater than that of the fibers
being spliced comprising the steps of: a. preparing the ends
of the fibers to be spliced, b. entering an end of each of
the fibers to be spliced into opposite ends of a fiber
passageway in a metal fiber splice element until the ends
are in contact with each other near the center of the splice
element, c. heating said metal fiber splice element to a
temperature above 80 C, and d. actuating said splice element
to clamp onto the fiber ends for maintaining the fiber ends
in intimate contact while cooling the splice element,
whereby an axial compressive force is placed on and
maintained at the interface between the ends of the fibers.
-3a-
CA 02247448 2006-07-19
60557-5918
A further aspect of the invention provides an
optical splice between two single mode optical fibers
comprising a splice element having a longitudinal passageway
for receiving the ends of said fibers approximate the
midpoint of said passageway and having a coefficient of
thermal expansion greater than that of the fibers being
spliced, and said fiber ends being placed in axial
compression against each other free of any index matching
material, the axial compression being of a sufficient
magnitude to retain intimate contact of the fiber cores
throughout temperature cycling between 0 C and 40 C, wherein
said axial compression force between the ends of said fibers
is the result of heating said splice element to a
sufficiently high temperature to expand the splice element
prior to inserting the fiber ends into said passageway, then
allowing said splice element to cool after insertion of said
ends into said passageway and further clamping said fibers
within said passageway.
Still another aspect of the invention provides a
tool for use in completing a splice between two optical
fibers by generating and maintaining, in a splice element
and throughout temperature cycling between 0 C and 40 C, a
thermally induced axial compressive force at the interface
between two optical fiber ends so as to retain intimate
contact of the cores of the optical fibers, the tool
comprising: a base, a nest on said base for receiving the
splice element, wherein the fiber splice element has a
coefficient of thermal expansion that is greater than that
of the optical fiber being spliced, heating means in said
nest for heating said splice element placed in said nest,
and means in the base for supporting the fibers with the
fiber ends placed in said splice element, wherein, by
following the process described above said base supports
-3b-
CA 02247448 2006-07-19
60557-5918
means serve the function of applying force to at least one
of said fibers to afford axial compression between the fiber
interfaces in said splice element.
-3c-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
Brief Description of the Drawings
The present invention will be further described
with reference to the accompanying drawing wherein:
Figure 1 is a side elevational view of an end of
an optical fiber that has been cleaved;
Figure 2 is a perspective view of a fiber end
that has been cleaved and beveled;
Figure 3 is a side elevational view of the
cleaved and beveled fiber end of Figure 2;
Figure 4 is a side elevational view of a pair of
cleaved fiber ends placed in contact;
Figure 5 is a side elevational view of a pair of
fibers pressed together where the end of one fiber has been
cleaved and the end of the other fiber has been cleaved and
beveled and the ends have been placed in contact;
Figure 6 is a top plan view of a splicing tool
for practicing the process of the present invention to
obtain the desired m~!chanical splice of the present
invention;
Figure 7 is a graph illustrating the test data of
a splice prepared according to the present invention;
Figure 8 is a top plan view of a splicing tool
incorporating a movable fiber clamp mounted on a ball
slide;
Figure 9 is a longitudinal partial sectional view
of a splice where the splice element has been actuated
according to the present invention;
Figure 10 is a top plan view of a modified
splicing tool having a mechanism for applying various
measurable compressive force on the fibers to apply
compression on the fiber ends in a splice element;
Figure 11 is a front elevational view
illustrating a modified splice element and a modified nest
tool for stretching the splice element prior to actuation
-4-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
to place and maintain a compressive axial force on the
fibers at the fiber interface;
Figure 12 is an end view of the splice element
and tool of Figure 11;
Figure 13 is a front elevational view of a
further embodiment of a modified splice element positioned
on a housing of a splicing tool for insertion of the fiber
ends;
Figure 14 is an end view of the splice element of
Figure 13; and
Figure 15 is a sketch of the splice element having
a bending pressure applied to place compressive stress onto
the fiber ends.
Description of the Preferred Embodiment
The present invention provides an improved
mechanical splice and the novel method of obtaining the
splice. In the drawing, incorporated to illustrate the
novel features of the present invention, like reference
numerals are used to identify like parts throughout the
several views.
While optical fiber ends can be polished, it is
time consuming and difficult to be precise, cleaved fiber
ends are used in most mechanical splice configurations
currently available. Cleaving is a practice of first
scoring or nicking a very small point on the outer
periphery of the cladding of an optical glass~fiber, which
reduces its mechanical strength at that point. The later
application of a tensile or bending load along the length
of the fiber where the score defect was made causes the
fiber to break generally perpendicular to the fiber axis.
The cleaving procedure creates the fiber end face 8, see
Figure 1, which possesses the most easily obtainable defect
free surface mechanically obtainable for the fiber core 9.
Most mechanical splice configurations currently available,
contain a coupling medium, usually a gel or oil, that
-5-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
possesses the same index of refraction as the core 9 of the
fiber 10. This index matching material is used to fill the
gap at the interface between the pair of fiber end faces 8,
see Figure 4, which are spliced. This gap is caused by a
cleft or protrusion 11 which can occur at the point
directly opposite the score mark or nick.
This condition, as well as other undesirable
conditions, usually are caused by inherent problems within
the cleaving procedure. Many different tools are described
in prior art which perform the perpendicular cleaving operation
with varying degrees of accuracy. Accuracy is measured with a
micro-interferometer, which determines both the flatness and the
angularity of the end.
Undesirable features caused by the cleaving
process can be removed by grinding a bevel on the end of the
fiber. The conical geometry of the bevel 13 which has been
ground on the end of a single mode fiber are easily
generated by using existing tools which are operated both
manually and automatically. one such tool for bevel grinding
is described in publication No. WO 95/07794, published 23
March 1995, of Minnesota Mining and Manufacturing Company,
3M, St. Paul, Minnesota. A bevel angle or included angle,
of the cone shaped end portion 13, of between 40 and 160
degrees can be easily obtained by making simple adjustments
to the tools. The change in angle, provides a change in the
end face diameters. The fiber length is easily set with a
simple fixture, and once set, the angles generated are
consistent from fiber to fiber, within a range window of 10
degrees. The end face area or its diameter is determined by
the amount of material removed from the end of the fiber.
On the automatic bevel tool, this is controlled by the
amount of time the fiber is ground. The more time spent
results in a smaller end face diameter. On the manual bevel
tool, this is controlled by the number of revolutions the
fiber is rotated against the abrasive media, more
revolutions results in a smaller end face diameter. Control
-6-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
of the end face diameter is very consistent once the tool
parameters have been set. End face diameter can usually be
maintained within + or - 0.0002 inch, over many beveling
cycles before the abrasive media requires changing due to
wear. The surface area of the end face of a fiber is
greatly reduced by beveling the end of a fiber. Figure 5
illustrates this reduction by comparing a cleaved fiber 15
next to a cleaved and beveled fiber 16. By beveling the end
of the fiber to 0.0015 inch, which was the diameter used
most during splice testing, the surface area of the fiber
end face is reduced by 90%. The imperfections from the
cleaving operation located around the periphery of the end
face have been removed. The edge of the end face has been
strengthened, due to increasing the angles between the end
face and the side of the fiber. The angle used most during
splice testing was a 90 degree included angle, or 45 degrees
from the axis of the fiber.
The reduction in surface area is important for
two reasons. The first involves the fiber cleaving
operation which rarely yields a perfectly flat
perpendicular end face. An angle of up to 1 1/2 degrees
from perpendicular to the fiber axis, usually occurs on the
end face. If a pair of fibers 15 and 16 are mated so that
both have angled end faces, a gap is created between the
fiber cores. Figure 4, illustrates this condition. By
beveling just one of the fibers, fiber 16, keeping the
angular relationship intact, the gap is reduced.by 66%
using a 0.0015 inch (0.04mm) diameter end face on fiber 16.
The second reason involves the amount of axial
pressure required to elastically deform both fiber end
faces to the point that the fiber cores 9 are in intimate
= contact, and a gap does not exist between them. If
pressure on a fiber remains constant, and the surface area
of the fiber's end face is reduced, the total pressure
experienced at the fiber interface increases. This
-7-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
benefit, i.e., increasing fiber end interface pressure to
afford intimate pressure axial contact between the fiber
cores, is important. The amount of force that can be
transmitted down a length of 250 micron buffer coated fiber
all the way to its end face is very small. Using a
standard 1.500 inch (38.1mm) long Fibrlok'"Splice as an
example, the fiber is inserted 1/2 way into the splice,
which would be a distance of 0.750 inch (19mm). This would
be the closest point at which the fiber could be gripped by
some device in order to transmit force longitudinally to
the end face. The diameter of the glass portion of a 250
micron fiber measures 0.005 inch (0.125mm). This is a poor
length to diameter ratio (150 to 1) for the transmission of
force. The splice entrance hole is several thousands of an
inch larger than the outside diameter of the fiber, to
provide clearance, which allows the fiber to bend when put
under axial pressure. If this pressure is too great, the
fiber buckles, and damage or breakage can occur.
Some of the benefits obtained by cleaving and
beveling fibers can be acquired also by polishing the end
of a fiber. Undesirable defects caused by cleaving are
removed. The face surface area is greatly reduced. Fiber
edge strength is usually improved depending on the end face
profile that is used. Control of the end face profile is
dependent on the equipment and procedures used. Polishing
can be obtained by heat polishing or abrasive polishing.
Heat polishing means the melting of the fiber surface and
abrasive polishing involves the removal of glass with an
abrasive media. There are several disadvantages associated
with abrasive polishing which leaves scratches in the end
face of the fiber. Finer and finer abrasive grits are used
to reduce the size and depth of these scratches, improving
surface finish, but scratches will always remain. The
finer the finish desired, the more polishing steps
required, which means more time required.
-8-
CA 02247448 1998-08-25
WO 97/36200 PCTIUS97/02812
Optical fiber splices are expected, by most
customers, to meet standard Bellcore performance
specifications. Limits are prescribed for return loss
throughout a temperature range of -40 C to 80 C. Typical
splices, which contain optical index matching materials,
display poor return loss at and near these temperature
extremes. Intimate contact should be maintained between
0 C and 40 C for indoor applications.
Concerns raised with the index matching materials
in mechanical splices might find solutions by improving the
optical index matching media, or eliminating the optical
index matching media. In order to eliminate the optical
index matching media, the ends of the fibers must be
improved to ensure intimate contact between the end faces. The
fiber end preparation procedures previously described yield
improved fiber ends that allow intimate axial pressure contact
between the end faces and eliminate the need for index matching
materials.
To test these improved fiber end faces, Fibrlokl"
Splices were made in the factory without the application of
index matching gel and/or oil, a "dry" splice. A pair of
125 micron single mode fibers were stripped and cleaved,
within 1 degree of being perpendicular. The fibers were
then beveled with 90 degree included angles, and 0.0015
inch diameter end faces, similar to the fiber 16
illustrated in Figure 5. They were cleaned by moistening a
lint free cloth with isopropanol, surrounding the fiber
with the cloth, pulling it out, and finally making several
wipes of the tip of the fiber against the cloth. The
fibers were then inserted into the splice using the
standard established procedures, and then the splice was
actuated.
A Textronics Fibermaster7" OTDR was used to measure
splice performance during all tests. Measurements were
made by averaging the readings taken from both fibers, one
at each end of the splice, for greater accuracy. Several
-9-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
splice samples were made at room temperature, 26 degrees
C., and results were similar for all samples. Insertion
loss was below -0.2dB, and return loss, (backreflection),
ranged from -20dB to -45dB. The splices were put into a
temperature cycling chamber and put through a Beilcore
patterned temperature test, -40 C to 80 C. One hour at
each temperature with an hour and a half transition time
between temperatures. During this test it was noticed that
the insertion loss remained stable for all splices,
fluctuating less than a tenth of a dB, and the return
losses for all splices would increase to the-18dB range at
temperatures above 40 C. Below 0 C, all splices improved
from their original room temperature measurements, ranging
between -40dB and -60dB.
The explanation for these results is as follows.
The Fibrlok''T' Element is made from aluminum. The thermal
coefficient of expansion for aluminum is 0.0000238 per unit
length per degree C. That number for glass can be between
0.0000102 and 0.00000005, depending on its chemical
composition. Aluminum will therefore expand and contract
at a faster rate than glass. The Fibrlok"" element grips
the glass fibers upon actuation. The fibers are actually
lightly embedded in the elements surfaces, and no slippage
occurs between the fibers and the element. When the test
splices were made at room temperature, the fiber ends were
placed into light nominal contact with each other, without
fiber end face deformation, using the standard splice
assembly procedure which uses forces generated from bowing
the buffer section of the fiber outside of the splice to
ensure fiber end face contact is made inside the splice.
When the temperature inside the thermal test chamber
increased above the temperature at which the splices were
assembled, the aluminum element was expanding at a faster
rate than the glass fiber, until all pressure at the fiber
interface had been relieved and the fibers started to
-10-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
separate. When the fibers separated, high return losses
are measured due to the glass/air/glass interface. This
condition reversed itself when the thermal chamber
temperature dropped below the splice assembly temperature.
As the temperature decreased below room temperature, the
aluminum element contracted at a faster rate than the glass
fiber, resulting in increased pressure at the fiber
interface, reducing and eliminating the glass/air/glass
interface yielding lower return loss readings. The return
loss pattern remained consistent from one cycle to the
next.
These tests lead to the present invention. The
present process involves heating a dry Fibrloka' Splice to a
temperature that was higher than the highest temperature
that the splice would ever be subjected to, inserting the
fibers into the "hot" splice, and actuating it. This
should ensure that the fiber end faces are always in
intimate contact within the target operating temperatures,
which would yield low return losses.
As an example, a commercial Fibrlok' ' Splice
Actuation Tool 17 was modified as illustrated in Figure 6
to apply heat to the splice. The plastic splice retention
nest was machined off the base of the tool, and replaced by
a new nest 20 made from aluminum which housed two 1/8 inch
(3.17mm) diameter by 1 inch (25.4mm) long, electrically
powered, 25 watt cartridge heaters 21 and 22. The heaters
were controlled by a Ogden digital control, accurate to +/1
degree C. The splice nest was designed to surround as much
of the splice as possible without affecting its function,
in order to transfer heat to the splice as fast as
possible. A thermocouple was placed inside the center of a
= metallic FibrlokTM splice element, it was assembled into a
splice, and the splice placed into the modified tool. The
splice was heated four separate times, from room
temperature to 100 C. The average time for the splice
-11-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
element to reach the target temperature was 50 seconds.
When the splice was removed from the tool, it required from
three to four minutes for the splice element to cool back
to room temperature.
The splice was tested and the temperature control
for the first splice test was set at 100 C, 20 degrees
higher than the highest temperature specified in the
Bellcore testing procedure. A dry FibrlokTM Splice was
inserted into the heated aluminum nest and allowed to soak
at a temperature of 100 C for a minute. A pair of fibers
were cleaved, within 1 degree, and beveled to an end face
diameter of 0.002 inch (0.05 mm). The fibers were inserted
into the heated splice, the splice was actuated, allowed to
cool, and placed into the temperature testing chamber. The
first performance measurements taken at room temperature
showed insertion loss was -0.10dB with a total fluctuation
of 0.02dB. Average return loss was -56.7dB with a total
fluctuation of 3.7 dB. The results indicated stability of
the optical signal. The normal buffered fiber restraints
23 and the actuating lever 24 remained on the tool 17.
Another splice was prepared and assembled using
the same parameters as the previous splice test. A graph
illustrating this test data is shown in Figure 7. The
first return loss measurement taken at room temperature was
-57.5dB. The splice was slowly heated back to 100 C,
taking measurements every 5 starting at 60 C. The return
loss slowly decreased until 80 C at which point it
decreased at a faster rate until it peaked at -83dB at
90 C. Between 90 C and 100 C, the return loss rapidly
increased, reaching -36dB at 100 C. The splice was then
allowed to cool. A second heating cycle was performed
having similar results as the first, except the peak was
observed at 94 C. (-93dB). During this cycle, the splice
cooling trend was recorded. During cooling, the return
loss decreased at a fast rate to a peak of -93dB at 83 C.
-12-
CA 02247448 2005-08-09
60557-5918
It then rapidly increased to -65dB at 70 , after which, the
increase was very slow. The cooling cycle was stopped at
-40 C, with a return loss reading of -56.8dB. A third
cycle was performed, and followed the same pattern as the
second, with slightly better results. Looking at the graph
of Figure 7, a rapid decrease in return loss is followed
immediately by a rapid increase, within approximately a 20
range. The peak, or lowest reading, is the transition
point between fiber contact and separation, on the hotter
side of the peak. On the colder side, it is theorized that
the compressive forces being generated at the fiber
interface by the contracting aluminum is causing the glass
density to change, which changes its index of refraction.
This eventually stabilizes, and the increase in return loss
almost flattens out when the temperature was lowered. Note
that the optical signal is fairly stable in the Bellcore
operating temperature test range of -40 C to 80 C.
A second tbol modification is shown in Figure 8,
where a fixed clamping mechanism 25 was attached to the
actuating tool 17 at the left side of the splice nest 20,
and positioned to grip the fiber immediately as it exited
the splice. A second fiber clamp 26 was mounted at the
right side of the splice nest 20 on a linear ball slide
with approximately 0.1 inch (2.5 mm) travel. A compression
spring was in contact with the ball slide, and located
opposite the splice rest 20. A screw was mounted on the
tool base and was used for adjusting the force that the
compression spring placed on the fiber via the ball
slide.
A dry Fibrlokl" Splice was placed into the tool,
and a pair of fibers were prepared having a cleave angle of
less than 1 degree, and bevel diameters of 0.0015 inch
(0.038 mm) on each fiber end. The tool was heated to 100 C
one of the fibers was placed into the left hand side of the
splice approximately half way, and then clamped. The
-13-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
second fiber was placed into the right hand side of the
splice until it made contact with the first fiber, and was
then clamped. The force adjustment screw was rotated until
approximately 0.3 pounds (1.3 Newtons) of compressive force
was generated, and then the splice was actuated to clamp
the fiber ends. Both clamps were then released. Return
loss measured -56.9dB, with the splice still at 100 C in
the tool. The splice was then allowed to cool.
Measurements were taken during this cooling cycle. Results
were posted. The splice was cooled to -40 C. Return loss
had increased to -51.8dB while insertion loss remained at
-0.11dB. The splice was then heated, with measurements
being taken every 5 . The transition zone was reached at
151 C where a measurement of -80.3dB was recorded. Beyond
this temperature the return loss increased rapidly. The
splice was again cooled from a temperature of 165 C. The
transition zone was reached at 134 C with a measurement of
-80.3dB, after which, return loss rapidly increased for a
15 to 20 degree period and then increases were extremely
slow.
The added axial pressure of the compression
spring fiber preloading increased the temperature at which
the fibers separated, while keeping the return losses at
the colder temperatures comparable to previous tests. The
optical signal remained fairly stable within the Bellcore
operating temperature test range.
A series of tests were performed involving
assembling splices using the compression spring preload
method, without the use of heat on the splice element. A
dry FibrlokA' Splice was placed into the tool nest which
remained at room temperature. A pair of fibers were
prepared with less than 1 cleave angle, and bevel
diameters of 0.0015 inch (0.038 mm). The fibers were
gripped in the clamps like the previous test.
Approximately 0.2 pound (0.9 Newtons) of compressive force
-14-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
was applied to the right side fiber, then the splice was
actuated and all clamping forces were removed. The fiber
end face transition zone for heating was between 129 and
134 C. The transition zone for cooling was between 114
and 120 C. After the first heating cycle, all remaining
cycle patterns were very consistent with one another inside
of the Bellcore operating temperature test range.
The body 28 and cap 29 of the Fibrlok' Splice
shown in Figure 9 is formed of a liquid crystal polymer
with 30% glass loading and can withstand the temperatures
described above without deleterious effect and the splice
element 30 is formed of aluminum.
A further modification of the actuation tool 17
is shown in Figure 10. The modification was the addition
of a force gauge 35 placed next to the left side of the
splice nest 20 opposite to the fixed fiber clamping tool
25. The force gauge measurement probe pushed on one end of
a pivot arm which in turn pushed on the moveable fiber
clamp. The pivot was positioned in such a manner to reduce
the forces generated by the force gauge/linear slide
assembly by a ratio of 10 to 1. This will improve the
accuracy of the force measurement on the movable fiber in
making a splice.
Four dry Fibrlok7" Splices were assembled using
the new actuation tool. Fiber end face axial compressive
preload was applied with the force gauge/linear slide
assembly. The heated splice nest 20 was not used, heat was
not applied. Two of the four splices had a beveled fiber
to beveled fiber interface, while the remaining two had a
beveled fiber to a cleaved fiber interface. The four
completed splices were placed into a thermal temperature
cycling chamber for a long duration Bellcore patterned
temperature cycling test. The cycle pattern was
accelerated to a four hour complete cycle, starting with:
(1) 1/2 hour at -40 C, (2) 1 1/2 hours transition time from
-15-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
-40 C to 80 C, (3) 1/2 hour at 80 C, (4) 1 1/2 hours
transition time from 80 C to -40 C, continuously repeating.
The purpose of this test was to verify that dry fiber
interfaces could survive many repetitions of the Bellcore
temperature test pattern, and to investigate consistency of
the optical signal from start to finish. Four splices
assembled with this fixture endured 530 complete cycles,
taking approximately 89 days. Measurements were made with
the OTDR by averaging the readings taken from both fibers,
one at each end of the splice, for greater accuracy.
Splices 1 and 2 had a beveled fiber to beveled
fiber interface, while splices 3 and 4 had a beveled fiber
to a cleaved fiber interface. Splice 1 possessed 0.0015
inch (0.038 mm) diameter end faces and was assembled at
0.22 pounds (0.9 Newtons) of fiber preload force. Total
return loss variation was 9.5dB, while insertion loss
variation was 0.05dB. Splice 2 possessed 0.0015 inch
(0.038 mm) diameter end faces and was assembled at 0.2
pound (0.89 Newtons) of fiber preload force. The total
return loss variation was 14dB, while insertion loss
variation was 0.05dB. Splice 3 possessed one 0.0015 inch
(0.038 mm) diameter end face and was assembled at 0.3 pound
(1.33 Newtons) of fiber preload force. Total return loss
variation was 31dB, while insertion loss variation was
0.11dB. Splice 4 possessed one 0.001 inch (0.025 mm)
diameter end face and was assembled at 0.1 pound (0.445 Newtons)
of fiber preload force. Total loss variation was 6dB, while
insertion loss variation was 0.05dB. Except for splice 3, the
results demonstrated good, stable optical signals with no change
in performance from beginning of the test, to its end. Splice 3
performed poorly at the higher temperatures. Upon closer
examination with the OTDR, it was discovered that at 80 C the
fiber end face transition zone had been reached. During some of
the cycles, the end faces would stay in contact yielding
excellent return loss, while in others they would separate
-16-
CA 02247448 1998-08-25
WO 97/36200 PCT/US97/02812
towards the end of the 80 C time period, yielding poorer return
loss.
Using the actuation tool illustrated by Figure 10
and a dry Fibrlok'*' Splice, a test was performed to
correlate the effect of end face preload force and return
loss. A pair of 250um buffer diameter single mode fibers
were cleaved and beveled to an end face diameter of 0.0015
inch, with a cleave angle of less than one degree. Both
fibers were inserted into the splice, and each was clamped
in its respective holder. The force was increased from
zero (0) pound to the point at which the fiber outside the
splice started to buckle. This usually occurred at around
0.3 pound (1.3 Newtons). Increments as low as several
thousandths of a pound were used at the beginning, where
several hundredths of a pound (a tenth of a Newton or less)
worked well during the latter stages of the test. Several
trials were made with each splice and fiber pair, and
several tests were conducted using different fiber pairs.
The results from the tests were substantially the same.
The lowest return losses usually occurred between 0.012 and
0.026 pound (0.05 and 0.1 Newton).
If a Fibrlok7" Splice, splice element is made from
a material other than aluminum, having a coefficient of
thermal expansion closer to glass, return loss performance
should improve, using the fiber compression process prior
to actuation of the splice. There is not a wide selection
of materials available that are as ductile and cost
effective as aluminum. Copper was close, and was chosen to
test this theory. The coefficient of thermal expansion for
copper is 0.0000141 per unit length per degree C., as
compared to aluminum which is 0.0000238. Several elements
were made, and a dry copper splice was assembled. A pair
of fibers were cleaved and beveled to a 0.0017 inch (0.043
mm) diameter end face. The fiber end faces were preloaded
to 0.25 pounds (1.1 Newton) and the splice was actuated.
-17-
CA 02247448 1998-08-25
WO 97/36200 PCTIUS97/02812
The test showed the performance was surprisingly
good considering the use of copper. Return loss
performance improved after the first heating cycle by
almost lOdB, and remained at that level for the remaining
cycles. The curve from the transition point towards the
cooler temperatures appears flatter compared to aluminum
elements, possibly due to the difference in expansion
rates. The curve from the transition point towards warmer
temperatures follows the same rapid increase as the
aluminum element. The transition point itself was within
the same temperature range as previous tests using aluminum
elements.
Optical fibers that are prepared using a quality
cleaving process and the application of bevel geometry on
one or both fibers, coupled with the generation of
compressive forces at the fiber interface by the use of
either heat or pressure, or both, inside of a"dry," i.e.
gel-less, Fibrlok7" Splice, can yield stable return loss and
insertion loss performance, equivalent to fusion splices,
during Bellcore temperature cycling tests without the use
of index matching materials.
Alternative methods of placing stress on the
splice element to maintain axial compressive forces between
the ends of the fibers upon completion of the splice are
discussed. One such method is to apply compression to the
ends of the actuated splice sufficient to cause internal
plastic deformation of the metal splice element from the
ends toward the fiber interface. The splice element will
thus apply and maintain the compressive forces at the
interface throughout the subsequent heating and cooling
environment.
Further, Figure 11 diagrammatically illustrates a
modified splice element 35 comparable to the FibrlokTM
splice element, modified to remove material along the sides
and form cam surfaces 36 adjacent each end, which cam
-18-
CA 02247448 1998-08-25
WO 97/36200 PCTIUS97/02812
surfaces are angularly related to the axis of the fiber
passageway 38, see Figure 12. The actuation tool is
modified to have a spreading cam 39 formed thereon. The
spreading cam 39 is formed with cam surfaces 40 which are
formed to engage the cam surfaces 36 to stretch the splice
element 35 upon the application of force on the splice
element 35 as illustrated by the arrows 41. The length of
the elastic deformation of the splice element, in the
longitudinal direction, is controlled by the distance the
splice element is forced down upon the spreading cam 39.
Alternately, the cam surfaces on the element could be
designed to shear at a selected force which would result in
the desired extension. After insertion of prepared fiber
ends, the element is closed to clamp the fibers in place,
and the spreading force placed previously on the element is
removed. With the removal of the spreading force, stored
energy in the element causes a contraction of the element
35, placing the opposed optically aligned ends of the
fibers in longitudinal compression or intimate axial
contact.
Figure 12 diagrammatically shows the end view of
the modified element 36 and the spreading cam 39.
Figures 13 and 14 illustrate a further
modification wherein a splice element 45 is rectangular
having a longitudinal V-groove 46 along the upper surface
and the ends of the fibers are placed in the V-groove in
firm contact near the longitudinal center of the splice
element 45 and of the V-groove 46. The fibers are then
firmly secured at the opposite ends of the longitudinal
groove. While fixing the position of the ends of the
element within a housing, a force 48, see Figure 15, is
applied normal to the top surface of the element 45 in order
to induce bending in the element. The bending force 48
applied should be sufficient to cause plastic deformation of
the element 45 into the shape of an arc, with the secured
-19-
CA 02247448 1998-08-25
WO 97/36200 PCTIUS97/02812
fibers on the inside surface. As a result of the plastic
deformation of the element, material along the bottom
surface of the element is elongated, while material along
the top surface, which also contains the V-groove holding
the fibers, is compressed. It is this differential stress
to the element 45 and the compression of the material along
the top surface which places the fiber end faces in axial,
i.e. longitudinal compression. Plastic deformation of the
element maintains the compressive force on the fiber ends at
the interface.
Having thus described the invention it is to be
appreciated that modifications may be made in material or
in some dimensions and not depart from the spirit of the
invention as defined in the appended claims.
-20-
