AMAZON multi-meters discounts AMAZON oscilloscope discounts
Introduction
This section will identify the main issues involved in connecting fibers together and to optical devices, such as sources and detectors. This can be done using splices or connectors. A splice is a permanent connection and a connector is used where the connection needs to be connected and disconnected repeatedly. A device used to connect three or more fibers or devices is called a coupler. This section will examine the following areas:
• Different splicing techniques
• Step by step fiber jointing procedures
• Practical connector types
• Step by step procedures for attaching connectors to fibers
• The operation of optical couplers
1. Optical connection issues
The main parameter of concern when connecting two optical devices together is the attenuation; that is, the fraction of the optical power lost in the connection process. This attenuation is the sum of losses caused by a number of factors, the main ones being:
• Lateral misalignment of the fiber cores
• Differences in core diameters
• Misalignment of the fiber axes
• Numerical aperture differences of the fibers
• Reflection from the ends of fibers
• Spacing of the ends of the fibers
• End finish and cleanliness of fibers
Theoretical analysis of the losses caused by these factors is complicated by the fact that the distribution of the power across the face of the fiber is usually unknown and varies according to the type and length of fibers, method of excitation etc. The following discussion of idealized connections will illustrate the sensitivity of the attenuation to various loss mechanisms.
1.1 Lateral misalignment of fiber cores
Here, it is assumed that fibers of the same diameter are displaced by a distance d, and are otherwise perfectly aligned as shown in FIG. 1. For simple, worst-case analysis, it is assumed that the power is uniformly distributed across the fiber cores.
FIG. 1 Lateral misalignment of fibers
The coupling efficiency η is calculated from the ratio of the overlapping area to the core area. For small displacements where d is less than 20% of the radius of the core, η can be approximated as:
η = 1-(2d/θa)
Attenuation = 10 log {1-(2d/θa)}
This equation is plotted in FIG. 2 for a step index multimode fiber, and shows how important the lateral alignment of fibers in the connection process is. For a step index multimode fiber, the assumption of uniform power distribution over the fiber is not unreasonable, but it overestimates the loss for graded index fibers where the distribution is less uniform and concentrated more on the axis.
FIG. 2 Typical lateral alignment loss
1.2 Differences in core diameters
Similar mismatches of the emitting and collecting areas of the interface can occur if fibers are perfectly aligned but have differing core diameters. For fibers with emitting diameter of d1 and collecting diameter d2, the loss is given by the formula:
Consider the example of a nominal 50 µm fiber with typical manufacturing tolerance of ± 3µm. It is possible to have a worst case mismatch of a fiber with 53 µm core jointed to one of 47 µm. Assuming uniform distribution of light over the fiber diameter, this results in a loss of 21% of the light and equates to about 1 dB loss. This is a conservative worst case estimate, as the light is generally not uniformly distributed.
This effect is severe if different types of fiber are connected, such as a 62.5 µm multimode fiber connected to a 50 µm fiber where a loss of more than 1.9 dB can be expected. Note that there is no significant loss incurred going from a small fiber to a large one, which can collect the entire incident light.
1.3 Angular fiber misalignment
When the axes of fibers are not aligned, the light enters the second fiber at greater angles and depending on the numerical aperture NA, some of the rays are unable to be confined to the core. This is illustrated in FIG. 3.
Loss = 10 log (d1/d2)2 dB
= 20 log (d1/d2) dB
FIG. 3 Angular misalignment
The coupling efficiency η for small angular misalignments of θ radians is given by:
The loss calculated as:
This loss equation is plotted in FIG. 4 for some multimode step index fibers. The loss can be seen to decrease with the larger numerical apertures, attributable to the fact that at a large NA, the radiation is distributed over wide angles so small angular errors affect less of the total power. The effect of an index matching liquid of refractive index No = 1.5 on a glass fiber is also illustrated in FIG. 4. This shows that the angular misalignment loss increases (although the end separation loss as discussed in section 1.6 is decreased) compared to no index matching (No = 1).
FIG. 4 Typical angular misalignment losses
Angular Misalignment (θ Degrees)
1.4 Numerical aperture differences
Differences in the numerical aperture (NA) of fibers can also contribute to the interface loss. If the receiving fiber has a smaller NA than the source fiber, then light will enter the fiber outside the acceptance angle as shown in FIG. 5. Then such light will not be confined to the core, and subsequently, will leak away.
FIG. 5 Numerical aperture differences
The loss can be quantified by the following formula:
This equation is plotted in FIG. 6 and it assumes a uniform modal distribution. This applies with equal validity to multimode step and graded index fibers. There is no loss if the receiving fiber has a greater NA than the source fiber.
FIG. 6 Effect of different NAs
Loss = 10 log (NA1/NA2) 2
for NA1 > NA2
1.5 Reflection at the end of fibers
When there is a gap between the fibers, Fresnel reflection, as discussed in section 3.1.8, takes place at each of the air-fiber boundaries. If it is assumed for simplicity that all the light is aligned to go straight through between media of refractive indices N1 and N2, then the loss is given by the formula:
where r is the reflection coefficient and is equal to:
For an air-glass interface, (N1 is glass at 1.5 and N2 is air at 1.0) this gives a 4% reflectance (0.177 dB); thus both reflecting surfaces add 0.35 dB. This can be reduced by the use of an index matching gel, with refractive index closer to that of glass, but not necessarily a precise match. For a gel of refractive index 1.4, with glass of refractive index 1.5, the above formula shows the Fresnel loss to be 0.005 dB per surface or 0.01 dB total.
As was seen in FIG. 4, such index matching increases the sensitivity to angular misalignment. Many practical mechanical splices and connectors incorporate index matching gels. These can be thick, viscous liquids like glycerin or silicon grease for unglued joints or transparent epoxy glues.
Reflection from the end of the fibers is also referred to as back reflection. Optical return loss (OPL) is given by the ratio of the back reflection to the input power:
Optical return loss measurements are discussed in Section 9. Reflection from the ends of the fibers can be a major contributor of errors on high bit-rate singlemode systems.
This occurs when the reflected light interferes with the laser diode, causing mode hopping. This can be reduced by the use of non-perpendicular cleaves, on mechanical splices, to ensure the back reflectors are absorbed in the fiber cladding. At connectors, a similar technique is used to prevent the reflections causing interference. The end of the fibers are normally polished to a slightly convex surface by the action of a polishing puck moved in a figure 8 pattern on an abrasive sheet.
1.6 End separation of fibers
When the fibers are separated, losses occur due to the spreading of the light from the fibers, as illustrated in FIG. 7. The light exits the source fiber in a conical beam with the spreading angle dependent on the NA. With a gap separating the fibers, some of the transmitted light is not intercepted by the receiving fiber. Fibers with large NA have greater separation losses because their beams diverge more quickly. Index matching liquid placed in the gap reduces the amount of spreading, and hence the loss.
Loss = 10 log (1- r) dB
r = [(N1 - N2)/(N1 + N2)] 2
OPL (dB) = -10 log (reflected power/input power)
FIG. 7 Index matching gel reduces end separation effects
The loss for small separation (s) is given by the formula:
Where N0 is the refractive index of the index matching gel and a is the core radius of the fiber. A uniform power distribution is assumed. This equation is plotted as FIG. 8 for differing types of fiber, and the beneficial effect of the index matching gel (N0 = 1.5) can be seen. For singlemode fibers, a similar analysis shows that the fiber end gap is even less critical, with a gap of 10 times the core radius producing a loss of less than 0.4 dB.
Loss = -10 log {1 - (s NA)/4aN0 } dB
FIG. 8 Typical end separation losses
1.7 End finish and cleanliness of fibers
The end of the fiber needs a smooth finish across its core to prevent light being scattered at the rough surface. This is achieved by cleaning the ends of glass fibers or polishing the fiber.
The fibers need to be clean as particles of dirt can scatter or absorb light. Joint preparation always involves cleaning the fibers, normally with a lint free cloth and isopropyl alcohol.
1.8 Connection loss summary
We have seen in the above discussion of the various loss mechanisms involved in connecting multimode fibers that the most important is the axial misalignment of the fibers.
With connectors, the minimum loss across the glass/air interface between them will always be about 0.35 dB unless index matching gel is used.
2. Fiber end preparation
A necessary preliminary for connecting optical fibers together or to a connector is to prepare the end of the fiber. There are three basic methods; the scribe and break method, the cleavering method and the lap and polish method. The scribe and break technique and the cleavering tool are usually used prior to splicing of glass fibers, while the lap and polish approach is needed for plastic fibers and attachment of connectors.
2.1 Preparation of glass fibers
The first step in preparing the glass fiber is to carefully bare the fiber, removing any plastic jacket material, strength members and buffering etc. This can be done using wire strippers, razor blades, or similar sharp tools. Great care needs to be exercised to avoid damaging the surface of the cladding. Chemicals, such as methylene chloride, can also be used for the stripping process. Some of these are dangerous to handle and require special precautions. After stripping, the bare fiber is chemically cleaned, using typical industrial solvents like isopropyl alcohol.
2.2 Scribe and break method
The cleaned glass cladding is now nicked by a hard cleaving tool, with typically a diamond, sapphire or tungsten-carbide blade. The fiber is kept under moderate tension while the cleaving blade passes across it and the tension is then increased until the fiber breaks. When performed properly, this produces a flat, mirror-finish surface approximately (within +/- 3 degrees) perpendicular to the fiber axis. This is the quickest and cheapest method of preparing the fiber ends for connection, although the result is generally least effective. The end should be carefully inspected to verify that a smooth clean break has been produced.
2.3 Cleavering tool
This is a very similar method to the scribe and break method, but is done with a very precise mechanical cleavering tool. The end of the fiber is placed into the tool and held fast. Then a gentle action cross-sectional pressure is applied and a clean flat break is achieved. This generally produces a far better result than the scribe and break method.
2.4 Lap and polish method
The lap and polish method is used where the cleaving approach is not practical; for example, in attaching connectors to the fibers. Most connectors have polishing attachments and procedures unique to that design. A generalized connector attachment procedure will now be discussed as an illustration.
The fiber is first inserted into a metal, plastic, or ceramic ferrule, whose purpose is to hold the fiber accurately in position, protect it from damage, and mechanically position the fiber into precise alignment within a matching connector. Connector construction is discussed in detail in section 4.
The bare fiber and its protective jacket are permanently cemented into the ferrule with epoxy around the protruding fiber. The fiber is scribed and broken to create a reasonably straight edge. The ferrule is fitted to a removable lapping tool, which is used to hold the ferrule perpendicular throughout the grinding process. The fiber is ground down with successively finer grades of abrasive to achieve a polished surface flush with the end of the ferrule. Water is used with the abrasive paper to lubricate the fiber and flush away the residues of glass and abrasive. Final lapping is done with polishing paste or lapping paper having abrasive particles of less than 1-micron diameter.
2.5 Preparation for plastic fibers
Plastic clad silica and all plastic fibers usually need to use the lap and polish technique because the plastic does not fracture as precisely as glass. The outer coverings of the fiber need to be removed by the same methods as for glass fibers detailed in section 2.1. The fiber and its jacket are secured in removable polishing clamp or in the connector ferrule and polished to the required extent, in a similar manner to that detailed above.
3. Splicing fibers
Two basic techniques are used for splicing of fibers; fusion splicing or mechanical splicing. With mechanical splicing, the fibers are held together in an alignment structure, using an adhesive or mechanical pressure. With the fusion splicing technique, the fibers are welded together, requiring expensive equipment but will produce consistently lower loss splices with low consumable costs. Mechanical splicers require lower capital cost equipment but have a high consumable cost per splice. Today, fusion splicing is the main technique for joining fibers. It is far better joining with significantly lower loss. Over the long term, it is also far more reliable.
3.1 Fusion splicing
Fusion splices are made by melting the end faces of the prepared fibers and fusing the fibers together. Practical field fusion splicing machines use an electric arc to heat the fibers. Factory splicing machines often use a small hydrogen flame. The splicing process needs to precisely pre-align the fibers, then heat their ends to the required temperature and move the softened fiber ends together sufficiently to form the fusion joint, whilst maintaining their precise alignment. During fusion, surface tension tends to naturally align the fiber axes minimizing any losses caused by lateral misalignment as discussed in section 1.1. Properly made fusion splices are as strong as the original fibers. Production fibers breaking under the proof test are simply fusion spliced for repair by the manufacturer. Such factory splices have typically less than 0.1 dB loss and have a tensile strength comparable to that of the original fiber. Commercial field splicing equipment, in skilled hands, can consistently produce splices with losses less than 0.1 dB. Fusion splicing equipment typically provides the following features:
• A fusion welder
This is normally an electric arc and its electrode spacing and arc timing need to be adjustable to suit the fibers being fused.
• Fiber holders and positioners
Devices to rigidly hold the fibers, accurately move them in three-dimensions so that the fiber cores are precisely aligned with each other and with the splicer electrodes. The fiber ends need to be brought together in alignment to complete the splice once the ends have been melted in the arc.
• Alignment devices
These are devices to ensure correct alignment of the fibers. Manual systems use a microscope or video camera to magnify the fiber ends by at least 50 times, to enable the operator to see the fibers while manually aligning them. Automatic systems use computer controlled positioners to align the fibers so as to optimize the light transmission.
• Optical performance checking
This is a method for checking the quality of the splice by checking the optical power transmitted across the splice. With automatic fusion splicing machines, this is usually done by coupling light into one fiber by tightly bending it around a light injecting post and similarly, coupling the light out of the fiber on the other side. This is also used to allow automatic alignment as described above as well as checking the loss of the finished splice. The simplest fusion splicers do not include this feature.
These basic components of a fusion splicer are shown in FIG. 9.
FIG. 9 Fusion splicer components
3.2 Fusion splicing process
The basic steps involved in making a fusion splice are as follows:
Fiber end preparation The protective plastic fiber jacket on both fibers must be stripped and cleaned for the appropriate distance; both fiber ends are then cleaved so the end faces are approximately perpendicular, (±3 degrees) to the fiber axis.
Fiber alignment
Each fiber end is clamped into a micro-positioner on the fusion splicer. The fiber ends are then brought into alignment: automatic splicers use computer control to optimize the light transmission as described above. Manual splicers rely on the operator to align the fibers using the micro-positioners, microscopic viewer, and mirror to inspect the alternative axis.
Arc cleaning
The ends of the fibers are separated at appropriate distance and a moderate arc is used for about one second to clean the fiber ends and round their edges.
Pre-fusion
Some splicers then bring the two molten ends of the fibers into contact to pre-fuse the fibers. The fibers are then checked again for transmission effectiveness, to confirm their alignment.
Fusion
The main fusing arc then melts the fiber ends and the molten fibers are brought closer together to complete the splice. The surface tension of the molten glass tends to align the fibers, thereby minimizing any lateral offset. The power, spacing, and timing of the arc are critical to achieve the correct temperature for the particular type of fiber.
Protection
When cool, the splice area can be coated with a plastic coating such as RTV or epoxy, for atmospheric protection. Some form of mechanical protection, such as a heat-shrink sleeve or mechanical clip is then fitted.
3.3 Mechanical splicing
Mechanical splicing involves many different approaches for bringing the two ends of the fibers into alignment and then clamping them within a jointing structure or gluing them together. Mechanical splices are generally used for short-term fixes only. Longer term fixes are provided by using fusion splices. Mechanical splices generally rely on aligning the outer diameters of the fiber cladding and assume that the cores are concentric with the outside of the cladding. This is not always the case, particularly with singlemode fibers. Some systems therefore allow active alignment where the fiber loss is monitored and the fibers rotated within the jointing structure to minimize the splice loss. Various mechanical structures are used to align the fibers, including V-grooves, sleeves, 3-rods and various proprietary clamping structures.
Some of these are shown in FIG. 10.
FIG. 10 Splice alignment structures
Some of the main mechanical splice categories are as follows:
Capillary mechanical splice
Here, the prepared fiber ends are inserted into a capillary tube matched to the outer diameter of the fiber. This is illustrated in FIG. 11. An index-matching gel or liquid may be used to reduce reflections from the butted ends of the fibers. Often, an adhesive is used to hold the fibers into the capillary tube, although compression from an external strain relief clamp or simply friction can be used. This provides a cheap, simple, easy to install splice, which is able to compensate for minor differences in the fiber diameters.
This is the most common technique for mechanical splices.
FIG. 11 Capillary mechanical splice
Polished ferrule mechanical splice
The polished ferrule mechanical splice, or rotary splice is a more complex jointing system capable of precise alignment of the fibers. The process is shown in FIG. 12. Here each of the cleaned fiber ends is inserted into a ferrule which mounts the fiber slightly off center. The fibers are glued into the ferrules, then their ends cleaved and polished flush with the ferrule tip. The two ferrules are butted together, with index-matching gel, inside an alignment sleeve. The ferrules are rotated to minimize the splice loss, and then they are fixed in place. This method is more complex and time consuming but can result in lower loss splices. It is well suited for splicing polarization-sensitive fibers.
FIG. 12 Polished ferrule mechanical splice
Elastomeric mechanical splices
The elastomeric splice uses a V-shaped groove in a flexible plastic insert for the alignment of the fibers, as shown in FIG. 13. The top of the splice sleeve exerts pressure to clamp the fiber into the V-shaped alignment groove. An index-matching gel or epoxy glue is inserted into the splice. The prepared ends of the fibers are then inserted halfway into the splice until they meet. These splices have typical losses of about 0.25 dB, so are suitable for the less critical applications, and for emergency service restoration.
FIG. 13 Elastomeric mechanical splices
Multiple fiber V-groove splicing
Multiple fiber ribbon cables can be simultaneously spliced by using suitably grooved plates. The ends of the ribbons are prepared by stripping the sheath, cleaning then polishing the fiber ends as a unit. The two ribbons are mated together on the grooved plates with an index-matching liquid. Each fiber fits into a separate groove in the plate and a matching plate fitted on top to hold the fibers in alignment, as shown in FIG. 14. The technique requires tight tolerances on the grooves in the plates. Suitable precision can be attained by etching the grooves in silicon chips.
FIG. 14 Multiple fiber splicing
4. Connectors
4.1 Connector properties
Connectors are used to make flexible interconnections between optical devices. The optical parameters discussed in section 2 need to be achieved in a unit, which is to be repeatedly connected and disconnected. A good connector needs the following mechanical properties to ensure consistently low loss throughout its lifetime:
• Repeatability
The coupling efficiency of the connectors should not change much with repeated matings.
• Predictability
The loss of the connector with a particular type of fiber should be consistent and relatively insensitive to the skill of the installer.
• Long life
The connector loss should not degrade over time, nor should repeated matings of the connector cause damage or degradation.
• High strength
The connector needs to be able to withstand significant stresses in use. This can arise through normal mating and unmating of the connectors, to abnormal forces on the cable and/or connector caused by bumping into connectors or tripping over cables etc.
• Environmental protection
The connector needs to provide protection to the optical interface from dirt, moisture, chemicals, temperature changes, vibration etc.
• Installation simplicity
The preparation of the fiber and its fitting into the connector should be relatively quick and easy.
• Easy to use
Connecting and disconnecting the connection should be simple, requiring minimal force or dexterity.
• Economical
Connectors should be reasonably priced. Precision components are required to achieve good performance components. As a rule, cheaper connectors, often plastic, are not precise for good performance.
Connectors have significantly greater losses than splices since it is much more difficult to repeatedly align the fibers with the required degree of precision. Active alignment, as was used to minimize some splice losses, is not possible. As was seen in section 2, axial misalignment of the fibers contributes most of the loss at any connection. Consequently, connector loss can be expected to be in the range from 0.2 to over 3 dB.
4.2 General connector construction
There are many different types of connector. The basic concepts in the connector design are illustrated in FIG. 15 and described below.
Most connector designs produce a butt joint with the fiber ends as close together as possible. The fiber is mounted in a ferrule with a hole size to closely match the fiber cladding diameter. The ferrule is typically made of metal or ceramic and its purpose is to center and align the fiber as well as provide mechanical protection to the end of the fiber.
The fiber is normally glued into the ferrule then the end cut and polished to be flush with the face of the ferrule.
The ferrules on two connectors mate with a precise, smooth fitting sleeve, also known as an adapter or coupling receptacle, which provides the necessary axial and angular alignment. The ferrules and sleeves may be tapered, as in biconical connectors, in which case plastic connectors may be used. This is because there is little abrasive wear when tapered components are repeatedly connected and disconnected.
FIG. 15 General connector construction
The ferrules are mounted in the connector body, normally metal or plastic, with provision for strain relief of the fiber. The cable strength member and jacket are usually attached to the body and a strain-relief boot can provide additional protection to the junction with the connector. The connector body also usually requires a mechanism to secure it to the coupling adapter. This can take the form of a screwed connection (SMA, FC and biconic types), a spring-loaded bayonet connection (ST and SC connector). An alternative type of connector follows a lensed approach. This is shown diagrammatically in FIG. 16. A lens is used to collimate the beam emerging from the end of the fiber. The fiber-lens distance is equal to the focal length of the lens. This produces a parallel beam of the lens diameter. This arrangement, when matched with a similar connector, has less sensitivity to lateral offset and gap between the connectors. In addition, it allows glass windows to be fitted over the lens to protect against dirt and scratches. Such connectors are more expensive and are used where rugged performance is critical, such as for military use.
FIG. 16 Lensed connector
4.3 Common connector types
There are many proprietary designs for connectors. Some standardization is now emerging and the International Standards Organization (ISO), the International Electrotechnical Commission (IEC) and the Telecommunications Industries Association (TIA) have all endorsed one connector type, the SC connector. Two other types of connector are widely used: the ST connector in the data and telecommunications industries, and the duplex FDDI connector in local area networks. These connectors are described in detail below.
SC connector
The SC connector is the most commonly used connector in industry today. The SC connector is shown in FIG. 17. This is built with a cylindrical ceramic ferrule, which mates with a coupling receptacle. The connector has a square cross-section for high packing density on equipment, and has a push-pull latching mechanism. The ISO and TIA have adopted a polarized duplex version as standard and this is now being used as a low-cost FDDI connector. The SC connector has a specified loss of less than 0.6 dB (typically 0.3 dB) for both singlemode and multimode fibers and a typical return loss of 45 dB.
FIG. 17 SC connector
ST connector The ST connector is shown in FIG. 18. This is an older standard used for data communications. This is also built with a cylindrical ceramic ferrule, which mates with a coupling receptacle. The connector has a round cross-section and is secured by twisting to engage it in the spring-loaded bayonet coupling. Since it relies on the internal spring to hold the ferrules together, optical contact can be lost if a force greater than about one kilogram is applied to the connector.
FIG. 18 ST connector [Connector Body Bayonet Fitting to Spring Loaded
Coupling Ceramic Ferrule Fibre, Fiber]
FDDI connector
The fiber distributed data interface (FDDI) system developed its own standard duplex connector. This is shown in FIG. 19. The FDDI connector is seen less frequently in industry simply because the FDDI standard is becoming less and less common. It is similar to the SC and ST connectors using a cylindrical ceramic ferrule design, which mates with a coupling receptacle. It is keyed so that it can only be installed with one polarity. This is essential for the FDDI system, which uses a pair of unidirectional fibers, to ensure accidental plugging of the transmit fiber into the receiver does not happen. Polarized duplex SC connectors are now being adopted for the low-cost FDDI standard.
FC connector
This is one of the earlier type of connectors that is rarely used today. It is similar to the ST connector but has a screw in bayonet.
FIG. 19 FDDI connector
4.4 Connector handling
Most fiber optic connectors are designed for indoor use. It is very important to protect them from contamination. The optical performance can be badly degraded by the presence of dirt or dust on the fiber ends. Even though a single dust particle would be 10 microns only in diameter, it could scatter or absorb the light and totally disrupt a single-mode system. Connectors and patch panels are normally supplied with protective caps. These should always be fitted whenever the connectors are not mated. These not only protect from dust and dirt, but also provide protection to the vulnerable, polished end of the fiber. Compressed air sprays are available for cleaning connectors and adapters, without needing to physically touch the mating surfaces. Take care not to touch the end of the connector ferrules as the oil from your fingers can cause dirt to stick on the fiber end. Cleanse connectors with lint-free wipes and isopropyl alcohol.
Durability of the connectors is important throughout their lifetime. Typical fiber connectors for indoor use are specified for 500 to 1000 mating cycles, and the attenuation is typically specified not to change by more than 0.2 dB throughout that lifetime.
Repeated connection and disconnection of the connectors can wear the mechanical components and introduce contamination to the optical path. Connectors for outdoor use require to be hermetically sealed. The lensed connectors described in section 4.2 are well suited to this application, with an optical window on the end of the connector providing the environmental protection. The much greater diameter of the light beam at the interface reduces the effects of any surface contamination on the optical performance.
4.5 Pig-tail
The most common method of installing connectors on cables today is to buy a piece of fiber with a connector already installed. A piece of patch cord, which may be 1, 2, or 3 meters in length, and has a connector installed onto it in a factory. This is referred to as 'pig-tail.' The pig-tail is then simply spliced onto the incoming cable.
Doing it this way guarantees a high quality connection. The factory process is carried out by robots and ensures that the end result is of a consistently high quality. If the installer then uses the process of fusion splicing to install the patch cord to the incoming fiber, this ensures a very low loss overall.
5. Optical couplers
Optical couplers or splitters and combiners are used to connect three or more fibers or other optical devices. These are devices, which split the input power to a number of outputs. While the splitting of the light is done passively, active couplers include optical amplifiers, which boost the signal before or after the splitting process. Coupler configuration depends on the number of ports and whether each of these is unidirectional, (also called directional couplers), or bi-directional. This leads to three main types of couplers, the T coupler, tree coupler, and star coupler, as shown in FIG. 20.
FIG. 20 Passive coupler configurations
The T coupler normally has one input and two outputs. The output power from the coupler ports can be designed for particular applications, such as 20% at one port and 80% at the other. Tree couplers have a one-to-many configuration, with one input to many outputs or vice versa. Star couplers can have multiple inputs and multiple outputs.
The type of fiber used with couplers is important: some couplers are designed for singlemode fibers only, because the light needs to be concentrated in a small area. Other types of couplers are more efficient with large core multimode fibers.
Practical couplers can take many forms:
• Fused fiber couplers use a pair of fibers with their cores placed in close proximity, so that light can be transferred between the cores. This is illustrated in FIG. 21. This produces a directional coupler.
FIG. 21 Fused fiber directional coupler
• Transmissive star couplers are made by fusing a number of fibers (as many as 64) together to form a mixing region, where the light from whichever source fiber is spread to all the output fibers. The operating principle is shown in FIG. 22.
• Reflective star couplers are used if the signals on the fibers are bidirectional, and the light enters the mixing zone from any fiber and is reflected to emerge from all the fibers, including its input fiber. This is shown in FIG. 23.
FIG. 23 Reflective star coupler