Fiber Maintainence

Fiber Maintainence

The diameter of the core in an optical fiber is very small and any irregularity (such as a join) can result in significant loss of power. To get the maximum light transfer from the cut end of one fiber into another, both ends must be cut precisely square and polished flat. They must then be butted together so that there is minimal air (or water) between the butted ends and these ends must match up nearly exactly.

In a practical situation outside the laboratory this is very difficult to do. In the early days of optical data communication (1983), one industry standard specified an optical cable for use in the office environment which was step-index 100/140 μm in diameter. The 100 micron core is very wide and is certainly not the best size for communication. However, at the time it was the best specification for making joins in the field. (This specification is still supported by some systems - including FDDI.)

Joining fibres together is not a trivial task. Light travelling in a fibre is not “like” electricity travelling in a wire except in the most superficial way. Light travelling in a fibre is a guided wave and the fibre is a waveguide. Any imperfection or irregularity (such as a join) is a potential source of loss and of noise. The problem (obvious as it is) is that the dimensions of a fibre are tiny and accuracy of alignment is critical.

  

There are three general ways of joining fibres:

1.     By fusion splicing (a type of weld)

2. Use of index matching epoxy glues

3. With mechanical connectors of different types

The common requirement of all three methods is that the cores must be aligned. However, we can't always line the cores up - we line the fibres themselves up. This is not the same. The core is not always in the centre of the fibre. The manufacturers try very hard but there is always a variation. Difference between the axis of the core and the axis of the cladding is expressed as the “concentricity” of the fibre. This means that unless you do something to align the cores whilst making the join, there is a random misalignment (and hence loss) imposed by the concentricity error. The major improvement in connector losses observed between the middle 1980's and the 1990's was due generally to better fibre manufacture as much as to better connectors and connection techniques.

Fusion Splicing

A fibre join is a type of weld. The fibre ends are cut, polished, butted up to one another and fused by heat. (Incidentally, with silica fibers you need quite a high temperature - much higher than the melting point of ordinary soda glass.) In practice, a light loss of only .1 dB is the current budget for power loss in a single-mode fiber join. But it should be realised that .1 dB is quite a lot in that it represents the total loss of one half of a kilometer of cable.

Fusion Splicing Schematic

A device setup for fusion splicing is illustrated in above.

1. Each fibre is stripped of its primary coating and the end cleaved such that it is square.

2. The fibre ends are positioned a few mm from one another and clamped to positioning blocks. There is often a groove provided in the mounting block to aid in correct alignment.

3. The fibre ends are then aligned with one another and brought closer together.

4. When alignment is satisfactory an electric arc is started between the two electrodes and the fibres brought into contact. Heat from the arc melts the glass and the join is made. There are two major issues here - alignment of the fibre and precise control of the heating arc.

Fiber Alignment using optical feedback

Fibre Alignment

When fibre ends are melted and touched together there are significant effects caused by the surface tension that tend to align the outside of the cladding. This is very convenient when joining multimode fibres but in the case of SM it can cause a number of problems. A common method of fibre alignment is illustrated in Figure. The primary coating is stripped from the fibre for several cm from the end.

·         When it is mounted in the fusion splicer the fibre is bent tightly around two mandrels (one at each end).

·         Light from a laser (or LED) is focused onto a spot on the fibre bend such that some of it enters the core in a guided mode.

·         At the other side of the splicer there is an optical detector positioned to capture light radiated from the tight bend.

·         One end of the fibre is moved (by moving the mounting block with a piezo-electric actuator) until the output of the detector is at a maximum.

This method works reasonably well but with SM fibres the surface tension effect can change the position as the join is made! Another problem is the fact that the primary coating has to be removed for a long distance either side of the join. This makes the join hard to protect from later damage. Many current automatic fusion splicers use a visual method of positioning. The fibre ends are examined (magnified) through a digital imaging process. Initially, this is displayed on a screen on the splicer so the operator can

check easily for faults in the fibre endfaces. You can clearly see the fibre endface and even the cores this way. Once the operator is satisfied that the ends look okay, a microcomputer in the splicer examines the images of the fibre ends and aligns them automatically. This method works remarkably well and can produce very low-loss splices. A big advantage is that you don't need to strip the fibre back more than a

few cm from the splice.

Control of Heating

Precisely how the fibre is melted and joined is a very important issue. To make a good mechanical join you should melt each fibre end completely and allow for intimate mixing of the glass from the two fibres. However, a join that does this is likely to have a strong perturbation of the refractive index and hence high loss. If you heat only a very thin layer on the endface of each fibre you get the best optical properties but the mechanical characteristics of the join are not good. Once the splice is made it must be protected. This is usually done by covering it with a sleeve of heat-shrink material and then applying gentle heat. The material contracts around the fibre and protects the splice. A metal strengthening pin is often integrated into the side of the wrapping to provide additional mechanical strength.

Cleaving the Fibre

Before a splice of any kind is made or a connector fitted it is of critical importance that the fibre ends be cut square. (Except for the unusual case where we are making a diagonal splice.) In the case of connectors the ends are often polished to a desired shape later but we need to start off with square cut fibre ends. The established technique for ensuring a good square fibre end is called “cleaving”.

This is just the same technique we use when cutting glass to replace a broken window.

1. A scratch or nick is made in the side of the fibre. This destroys the local surface tension and gives the glass a point from which to crack.

2. A stress is applied to the fibre so that it will crack across its diameter. You can apply this stress in many ways. An older technique called for the fibre to be bent around a rod. Unfortunately this had a habit of creating a “lip” on the side of the fibre opposite to the nick. In todays world you use a little machine which delivers a short sharp blow to the fibre in exactly the right place. Usually a good square cleave is obtained.

Mechanical Splicing

In this technique the fibre ends are cleaved and polished, aligned with one another and the gap between filled with an epoxy resin which has the same RI as the fibre core. There are various ways of aligning the fibre ends but we consider the splice to be “mechanical” if the outside of the fibre cladding is aligned without reference to alignment of the cores. There are many ways of aligning the outside of the fibres:

·         One common method here is to use a glass tube into which each end of the fibre is pushed. A small amount of the epoxy resin is placed on the end of one of the fibres before insertion. Usually there is a small hole in the tube at the point of the join so that excess epoxy can escape.

·         There are many other methods of obtaining mechanical alignment of the outside of the fibres. V-groves, slots and alignment rods are all used. In addition heat-shrink elastomer tubes are also used sometimes.

This splicing technique is the lowest cost but it is also not very good. The quality of the join depends on:

1. The concentricity of the fibre

2. The accuracy of the outside diameter of the fibre

3. The circularity of the outside of the fibre

4. The tolerances and precision of the alignment device used

However, this makes a solid, permanent connection and is used for fibre-to-fibre joins in many situations.

In similar techniques epoxy glues are often used for pigtailing microoptic devices (like circulators). There is a wide range of epoxies available which will meet most requirements. In a recent experimental situation a suitable special purpose epoxy could not be found. However, a dental (uv cured) epoxy was available (made for filling teeth) which had almost exactly the needed RI and a very low coefficient of expansion (you don't want your fillings falling out!). However, there is significant doubt about the long term stability of epoxy resins. Resins (might) break down and cause scattering over time. We don't know. Recently there has been a major shift in the industry away from the use of epoxy resins for just this reason.

Mechanical Splicing with Alignment and Bonding

This process is very similar to straight mechanical splicing but the fibres are actively positioned in the same way as with fusion splicing. The cleaved fibres are inserted into silica sleeves and bonded in place. The sleeve ends with the fibres exposed are then polished to get a very accurate surface. After this the sleeves are actively aligned so that the maximum optical power is transferred. They are then bonded with epoxy and covered with another protective sleeve. In reality, the role of the inner sleeves is simply to provide rigidity and bulk to the fibre to make handling and positioning easier and gluing of the endfaces

mechanically strong. This technique provides very high-quality splices but it is very time consuming (and

hence costly) to perform.

Losses in Fibre Joins

Losses in fibre joins are commonly classified into two kinds:

1. Extrinsic losses are those caused by factors concerned in joining the fibre but are unrelated to the properties of the fibre itself.

2. Intrinsic losses are losses caused by some property inherent in the construction of the fibre.

Extrinsic Losses

Longitudinal Misalignment

Lateral Misalignment

Angular Misalignment

Fibre end not cut square

Fibre end irregular or rough

Sources of Loss due to Misalignment in Fibre Joins. Because of the fact that these losses are caused by factors external to the fibre itself they are called “Extrinsic” losses. As shown in Figure  above there are many ways to make a bad fibre join. This applies whether a fused join is to be made or a connector is to be used. Losses and reflections for the different types of mismatches vary but all are to be avoided.

Longitudinal Misalignment

Longitudinal misalignment (or endface separation) has two loss effects. The first is just loss of signal power caused by the fact that light exiting one fibre endface diffuses outwards and (depending on the amount of separation) some of it will not be within the NA of the other fibre and hence cannot enter it in a guided mode. The second effect is that the separated endfaces themselves constitute a Fabry-Perot interferometer. Depending on the wavelength and the exact distance between the endfaces the attenuation can vary between zero and 100%.

Lateral Misalignment

Lateral misalignment is a major potential source of signal loss in all fibres but especially in single-mode fibres. A lateral displacement of one micron in an otherwise perfect join will result in a loss of .2 dB of signal. A displacement of 2.5 microns results in a loss of just more than 1 dB! Fibre End Not Cut Square

If the fibre end is not cut square then you can't mate the two surfaces closely together.

Angular Misalignment

This problem is worst in single-mode fibres due to the very small mode field and the low RI contrast (low NA). A misalignment of only 1 degree produces a loss of .2 dB. A misalignment of 2 degrees causes a loss of around 1 dB!

Fibre End Irregular or Rough

Rough ends on the fibre scatter the light and prevent close contact between the fibre ends. Most of the above comments apply to losses when connectors are used rather than when a fused join is made. In the case of a fused join, most of the above faults create a constriction in the fibre itself and a random perturbation of the RI. Losses in this context are hard to predict quantitatively but can be very large.

Intrinsic Losses

Core Concentricity Core Shape (Ellipticity)

Core Diameter Cladding Diameter

Sources of Intrinsic Loss in Fibre Joins Losses that are caused by factors involving the fibre itself are called “intrinsic losses”. The major ones are summarised below:

Concentricity Error

As mentioned earlier, one of the major causes of loss in fibre joins is concentricity error in the fibre. Concentricity error comes about when the axis of the core and that of the total fibre itself are not exactly aligned. That is, the core is not exactly centred in the fibre. Even assuming that the fibres are lined up exactly on the outside, concentricity error will cause the cores to be misaligned. Concentricity error is a problem for both SM and MM fibre but it is a significantly greater problem in SM fibres. However, vast improvements in fibre manufacture have been made and major fibre manufacturers have recently (1997) announced big improvements in this area.

Core Shape (Ellipticity)

No matter how precise the manufacture the core will always exhibit a (hopefully very slight) ellipticity. When a fibre is cut and re-joined the orientation of the core will usually not be the same and some light will be lost. This is not a big problem with MM fibre. With SM fibre, any ellipticity causes the fibre to be birefringent. That is, the fibre will exhibit different RIs to orthogonal polarisations of light travelling through it. A join in this case can be a source of birefringent noise.

Core Diameter

In MM fibre light is obviously lost (some modes escape into the cladding) when a core of a larger diameter is joined to one of a smaller diameter. This happens in the natural situation of every join where the diameter of the fibre core cannot ever be exact. There is always a difference however slight. Note that if the fibres are aligned correctly the loss will occur only when light passes from the larger diameter fibre to the smaller diameter one. Light travelling in the other direction (from smaller to larger) is not lost. In the situation where two fibres of different specifications (with different diameters) are being joined with a connector, then a lot of light is usually lost. This is a common situation, where fibres with a 62.5 μm core can be connected to fibres with a 50 μm core. This happens often because most available data communications equipment is pigtailed using 62.5 micron MM fibre. Some users have installed 50 micron MM cabling and so a mismatch is inevitable. Loss of light in this situation (about 3 dB) is unavoidable. Again, this happens only in the direction where light travels from the larger diameter core to the smaller one.

Mode Field Diameter

In SM fibres the actual core diameter is not very relevant in considering joins. The diameter of the “Mode Field” (generally larger than the core diameter) is the important parameter.

Cladding Diameter

When fibres are joined we line the fibres up with each other using the outside of the fibre (you can't see the core). This means that at some point on the outside of the cladding both fibres must align with each other. If the outside diameters of these claddings are different from one another then the cores cannot be aligned.

Numerical Aperture

When MM fibres of different NAs are joined some modes that were possible in the fibre of higher NA cannot travel in the fibre of lower NA. These will enter the cladding and ultimately be lost. Thus some optical power will be lost. Loss from this source will occur in only one direction (from the higher NA fibre to the lower NA one). Light travelling in the opposite direction will be retained. There is another source of loss and noise here. Fibres with different NAs usually have different RIs in the core or the cladding or both. When you join fibres of different RIs the RI changes at the join. The join then becomes a partial mirror and some light will be reflected back down the fibre. This can cause noise (as well as loss) due to the phenomenon of “Return Loss Variation”.

Refractive Index Profile

Differences in RI profile in the joined fibres can cause the same effects as described above for numerical aperture.