The current level of industrialization of the OVD (Outside Vapor Deposition) technology is reviewed. The individual manufacturing steps are described. Product flexibility and fiber quality are given, and trends are discussed.

Since 1972, AT&T has been examining the technology for the mass production of optical fiber. With the invention of the modified chemical vapor deposition (MCVD) process at AT&T Bell Laboratories' development began to define the manufacturing methodology to produce fiber with the highest quality at the lowest cost. By 1979, processes for fiber production had been developed and were then expanded to a full mass production facility at AT&T's Atlanta Works to manufacture connectorized lightguide cable. Since 1980, this facility has produced hundreds of thousands of kilometers of optical fibers using the MCVD process.

In this paper an overview is given of the PCVD process as applied for the large scale production of optical fibres for telecommunication. The specific merits and potentials of the process, such as the profile-independent high deposition rate and excellent controllability are described. The current state of the art of the process, as it is used in the Eindhoven production unit, is a deposition rate of 1 g/min., a preform size of 28 km and a drawing speed of 4 m/s. Fibre characteristics are well within the requirements imposed by the telecommunication market. The PCVD process has also proven to be suited for the production of dispersion optimized single mode fibres and high NA graded index fibres for short distance applications. For both fibre types the high refractive index differences obtained with fluorine doping are exploited. Depending upon market demands all fibre types can be manufactured at the same productivity. Some trends are given towards further increase of productivity and reduction of fibre costs.

Recent progress on the VAD process is described with the performance of VAD fibers in mass production basis. Some of the latest developments of VAD technologies, such as fluorine-added fibers and high-speed deposition techniques are also discussed.

We review the historical development of silica deposition methods using an atmospheric pressure plasma torch, from the Verneuil process for optical grade fused silica ingots to the Axial Lateral Plasma Deposition (ALPD) process for single and multimode fiber production. The present status of ALPD is presented, with special emphasis on index profile control, OH reduction and multiple ingot to preform pulling.

High numerical aperture multimode fibres for local area applications are presented. Using the MCVD technique with inner pressure control it is possible to dope the core only with Ge02 without cracks along the preform. A maximum value of 27% molar concentration of GeO2 has been achieved. Fibres have been made with nm N.A. greater than 0.3 and with a loss below 4 dB/km at λ= 850 nm and below 1 dB/km at 1300 nm.

Graded-index and single-mode fibers can be prepared by a low pressure plasma-impulse-CVD process. During the short plasma pulses, generated by pulsed microwaves, layers of mono molecular thickness are deposited onto the inner side of a quartz-glass tube. Thus a fine profiling of the refractive index distribution is possible and fibers drawn from these preforms are of high quality.

This paper will review the requirements for optical waveguide coatings, materials and coating designs, application and cure technology, and coating characterization techniques needed to assure product performance. Also discussed will be recent developments in high speed coatings, coatings for special applications, and challenges for the future.

The thermomechanical properties of various primary coatings was studied under static and dynamic conditions. Especially two different epoxy-acrylate resins have been investigated with respect to their aging parameters under hostile environments.

Fiber losses increase if hydrogen diffuses inside fibers. The following four methods are effective in preventing these loss increases. (1) Do not use P205 as a dopant, or use a minimum amount if it is used. (2) Use fluorine as a dopant. (3) Use a synthetic cladding layer. (4) Carefully select the material to cover fibers. Plastic materials that do not generate H2 should be selected for cables. Generally, UV curable resines generate less H2 as materials to cover fibers. Using these conditions, loss increases of GI fibers and SM fibers with H2 1 atm., 20 °C at 1.3 and 1.55 μm in 25 years will be <0.01 dB/km.

Characteristics describing the time dependent attenuation coefficient of an optical fiber during and following a very short and intense radiation pulse are analyzed. This problem is important for transmission applications when the fiber is subjected to gamma, electron, or neutron beams. Besides time, the attenuation coefficient is a function of temperature, dose rate, dose, nature of the radiation (n,e,y), fiber composition and purity, pre-existing solid state defects, and wavelength of the transmitted signal. The peak attenuation for a given fiber is mainly determined by the dose rate and pulse length, but temperature and strain (or athermal) annealing also contribute to a partial recovery during the pulse duration. The peak attenuation per unit dose appears to be smaller at high doses, perhaps caused by particle track overlap, which produces a saturation effect. After pulse termination, the attenuation coefficient tends to recover towards its pre-radiation value at different rates, depending upon the factors mentioned above. In particular, ionized electrons relax back to the positive lattice ions at a rate which depends upon initial separation distance and temperature. The initial separation distance is a function of beam energy. Some electrons will encounter a trap in the lattice and may recombine by quantum mechanical tunneling or be removed by photons (hence, absorption). Besides ionization, radiation may induce lattice displacements which in turn produce additional absorption centers. The displacement contribution has a different time constant than that associated with ionization. These topics, as they influence fiber characteristics, are discussed, along with supporting experimental data.

High coupling efficiency, as well as simple connector and coupler technology make plastic clad silica (PCS) fibers very attractive in applications requiring frequent structural adjustment like, for example, Local Area Networks (LAN). However, due to a significant variation of the core-cladding refractive index difference, severe variations of attenuation and band-width within an extended temperature range (-40°C up to +80°C) must be taken into account. The thermal characteristics of PCS fiber are therefore radically different from those of all-glass fibers.

A study has been made of temperature induced losses in monomode fibers with a loose tube, jelly-filled jacket. It has been found that at low temperatures where the fiber has to buckle to fit into the tube, the losses have character of pure bend losses. At high temperatures where the fiber is pressed against the tube wall, the losses are of the microbending type.

Glass as the optical fibre material is very sensitive to the damaging effect of flaws. Fracture mechanics theory shows up this influence and gives the possibility of predicting fiber strength in long-length application and time-to-failure based on examinations of short samples.

The achievement of a good reliability of optical cables has required high fiber strength. The surface treatment of preforms as well as drawing conditions used in order to realize this improvement are described. The fiber drawing has been performed using a graphite resistance furnace which can be operated its top either closed or open. The contamination of fibers with particles produced by the furnace is explained for these two types of configuration. The results are expressed in terms of break frequency on a 10 km reference length, process defect origins and fiber length yields after proof testing at 0.9 % and 1.5 % elongation.

Hermetically-coated optical fibers have been fablicated with aluminum coating using metal dipping method. Mechanical and optical properties of the fibers have been investigated for high temperature applications. Mechanical aging test have been performed at the temperature from 300°C to 400°C by static fatigue mandrel test. Aluminum coated fibers have been proved their fatigue resistivity, there have been no problem at 350°C continuous exposure. Optical aging test have been performed from room temperature to 600 °C. Several kinds of aluminum coated fibers which include different core material or different combination of core and cladding material, have been tested and loss spectra have been evaluated. Ultra-violet absorption tail and OH-absorption peaks appeared in all fiber when heating, but their magnitude and influences for application wavelength seems quite different. In silica core fiber, two absorption increase is small, so that, there is no influences of loss increase at 0.85 μm band. In germanium doped fiber, loss caused by the defect center must be converted to the stable permanent OH when high temperature application.

Based on laboratory investigations and flight test experiences we specify essential parameters and test requirements for airborne fiber optic components suitable for data busses in commercial and military aircraft. Functional and environmental characteristics and reliability requirements are described.

End preparation appears to be one of the most delicate operations during fibre splicing. While, in a mechanical splice, fibres are polished, a cleaving tool is usually used for fusion splicing. A common way of testing fibre breaks is interferometry : the interference pattern perfectly characterises a fibre end. However patterns are so different from one sample to another that comparison remains a difficult task. We propose a simplified method for the measurement of break angle as well as for the evaluation of surface quality : a collimated light beam is directed parallel to the fibre axis onto the cut where it is reflected and diffracted. A diffraction spot, which characterises the fibre end, forms in the infinite plane. The position of the spot gives the break angle with an accuracy of 10'. The size of the diffraction pattern is proportional to the amplitude of the surface defects. The experimental setup has been used for testing cleaving tool efficiency and checking fibre ends before splicing. The technique is convenient enough to give rise to a field control device. A PIN photodiode will be set in the centre of the diffraction image plane. The diffraction pattern is central, small and bright for a good break, the photodiode will then receive maximum, flux. The spot extends with a poor quality break so that the small size photodiode receives less flux. A quick selection of fibre breaks can be made by eliminating samples for which the flux received by the photodiode is less than a minimum value.

This paper describes the qualification procedure of preform analysers with respect to an interferometric reference test method. The index difference of monomode preform is measured with an accuracy of +/- 10-4. Geometrical dimensions and homogeneity are also controlled to adjust the fabrication process.

Fiber transmission properties are statistically quantified to fulfil the requirements of the Spanish Telephone Company (TELEFONI-CA) installations during 1985-88 period, both in the junction and in the long-haul links. As a result, more than 900/0 of the graded index fibre links can be nicely performed with fibres of category II (< 2 dB/Km: > 500 MHz, at 1.3 microms); for the long-haul links more than 900/0 of the elementary cable sections require an attenuation coefficient lower of equal to 0.5 dB/Km (at 1.3 microns), for 140 and 565 Mbit/s digital line systems.

The correlation of the baseband frequency response of single lengths of multimode fibre with that of an assembled set of such fibres has remained one of their most unsatisfactory features. Many approaches to this problem have yielded only partial success in a predictive sense, despite their effectiveness in accounting, after the event, for observed experimental results. The common feature among most predictive procedures is an assumption of 'good behaviour' of the fibres involved. Others have attempted to take into account 'rogue behaviour' (fibres having modes separable into distinctly fast and slow groups) by manipulation of the transfer function, in either time or frequency domains, in an effort to derive a frequency parameter leading to a consistently applicable predictive theory. The motivation, apart from scientific thirst for knowledge and understanding, has been economic. An algorithm enabling a required system performance to be achieved with certainty, without an excessive margin, and using the most readily available, lowest-cost fibre would clearly be a desirable outcome. Has such an outcome been attained? If so, to what extent, and under what conditions? Does the complexity of the available algorithms in any way detract from their practical utility and economic benefit? Are there better, simpler ways of achieving the end result? The author is of the opinion that the problems, scientifically fascinating though they are, are likely to remain largely unsolved, because in the end, they will not be worth solving.

In this paper the bandwidth measurement results on long multimode fibers links opera ting in the second window are shown and the effect of mode coupling and mode-group delay differences are analyzed by means of different techniques based on mode selective launch. Having assumed a ∝ length law for the link bandwidth criteria suitable for industrial application have been developed in order to predict with a good approximation the total bandwidth from the single fiber iengtn values.

Models have been developed to examine the interaction between fiber design parameters and system performance and cost. The design variables have been parameterized and studied using computer analysis. Representative fibers were fabricated and tested to verify the models. This paper reviews the engineering results of this study and presents specific applications already in use in the U.S.

For a future broadband network the possible services are described. The variety of services and the expected subscriber development requires the built-up of an overlay network for the starting phase which ends in a dense broadband subscriber network. The most secure fibre for future possibilities will be the single mode fibre in the trunk and also subscriber network. For the connecting and branching points in the local network solutions are presented.

The single-mode optical fibre network currently being installed as part of the BT modernisation programme is optimised for 1300 nm operation at 140 Mbit/s. The fibre specification is designed to enable upgrading to higher data rates and longer wavelength operation. Fibre properties at 1.3 um and 1.5 um are discussed and measured results reported.

The purpose of this paper is to review our development of single mode fiber testing methods for use in the manufacturing environment. The rationale for selecting particular measurement techniques will be discussed in terms of precision, speed, the relevance of the parameter determined and the ability to integrate the test equipment into an efficient and cost effective measuring station. In addition, the suitability of current industrial measurements for qualifying newer fiber designs, such as dispersion shifted fiber will be addressed.

It is known that several methods are available in order to measure a single parameter on the fiber. In the industry, one is concerned with the efficiency in terms of time and accuracy of the selected method. The paper presents some results obtained in a cable plant and the statistics related to the most important parameters such as attenuation, bandwidth, numerical aperture, geometry of the fiber cross section.

Contained in this paper are empirical results leading to a better understanding of the development of a steady state power distribution in an optical fiber, commonly referred to as Equilibrium Mode (Power) Distribution (EMD). First, three terms are defined and the importance of their measurement given: Transient Loss, Steady State Mandrel (SSM) Prescription, and Mode Power Distribution (MPD) Measurement. Second, the results of the transient loss testing and the mandrel wrap experiment, which determined the SSM (a mandrel which can be used for convenient attenuation measurement), are given for 85μm core fiber (Corning Data Broadband Fiber (dBF)). Third, MPD measurements verifying the conclusions drawn concerning the establishment of EMD or steady state are given as a function of fiber length and for various launch conditions.This paper has two primary conclusions. First, by actual measurement of the relative mode power distribution in sample fibers, steady state or EMD has been shown to have been established. Second, a mode filter in the form of a mandrel around which the sample was wrapped five times has been chosen such that the steady state power distribution was reproduced.

Methods of deriving propagation parameters of optical fibres from measurements of the transmitted near field are described. The transmitted near field of an optical fibre is dependent on the illumination wavelength spectrum and this is used to determine refractive index and mode spot profiles. When an optical fibre is multimoded by illuminating with white light the transmitted near field intensity distribution is directly related to the refractive index profile. This can be used to achieve a reasonably accurate measure of the refractive index profile of both multimode and monomode fibres. Both the theoretical considerations and the practical conditions for performing this measurement are described. When the optical fibre is single-moded the transmitted near field intensity distribution is the mode spot distribution. Many of the propagation characteristics can be determined from this distribution. For standard single mode fibres, such as step index or depressed cladding, the spot size and hence the equivalent step index are easily derived from this distribution. Advanced single mode fibres can be characterised from a more detailed knowledge of this distribution. Other techniques have been suggested and used for these measurements, they are all theoretically interrelated. By considering the relationship between the refractive index profile, the transmitted near field intensity distribution and the far field intensity distribution it is possible to estimate the relative merits and errors of these techniques.

The proper design and implementation of an optical communication system requires a knowledge of the transmission characteristics of the optical sources, fibers and interconnection devices used in the system. In this paper we will briefly review how extrinsic and intrinsic (to the fibers) parameters affect the transmission loss of interconnection devices. We will then describe an experimental splice loss versus transverse offset technique that was used in conjunction with a modal splice loss model to determine the mode power distribution produced in multi-mode graded index fibers by a variety of optical sources (lasers and LED's). A knowledge of the modal power distributions produced in optical fibers by optical sources is needed to determine important parameters such as information transmission capacity, splice loss and transient loss in a multimode fiber system.

Single-mode fiber and cable structures have been designed aiming at preservation of the intrinsic low loss property of the fiber. Apparatuses for fiber measurements and splicing have been developed for use in the field. On the basis of the developmental studies, NTT has carried out the construction of a system of large capacity principal transmission lines covering all of Japan. Performance of the cables was excellent, which made certain of a great progress in the telecommunication networks in Japan.

The advantages and drawbacks of cut-off wavelength and mode-field diameter measurements in single-mode fibres are analyzed with reference to the standardization issues and paying attention to the most recent evolution trends. Each parameter is discussed separately analyzing the various definitions and characterization techniques.

Various chromatic dispersion measurement methods have been explored and developped by many research laboratories or industrial manufacturers all over the world. These methods lead either to test benchs used for optical fiber design studies, or to on-site test equipments. After a non-exhaustive review of the main measurement techniques previously reported, we propose a new method using an electrically tunable external cavity semiconductor laser which leads to a compact high resolution field equipment.

For the multiple-source phase shift method of measuring chromatic dispersion in single mode optical fibers, the magnitude and origins of measurement errors are examined. Calculations show that worst-case errors can be caused primarily by source wavelength uncertainty for measurement of long fiber lengths with high modulation frequencies. For shorter (factory length) fibers and lower modulation frequencies both phase and wavelength uncertainty can be important error sources.

The length dependence of LP11 mode cutoff wavelength and its influence on the chromatic dispersion measurement by phase shift method is investigated both theoretically and experimentally. The effective cutoff wavelength λce decreases linearly in a logarithmic scale of fiber length as fiber length increases. Based on the length dependence of λce, the error in the chromatic dispersion measurement of a single-mode fiber by phase shift method is analyzed, which operates at wavelengths including those shorter than the cutoff wave-length. It is clarified that the measurement error for a 200 m-length fiber with a longer cutoff wavelength of 1.35 μm is less than 0.1 ps/km/nm.

The paper deals with the standardization work carried out within the CCITT (The International Telegraph and Telephone Consultative Committee) both on multimode and single-mode optical fibres. For each of these fibres, also suitable test methods are indicated.

Temperature changes involving the liquid cell of an RNF profiler may seriously affect the radial power flow distribution and, hence, be a practical limitation for absolute measurements. However, we demonstrate that the thermal dependence can be exploited to enhance the spatial resolution of the profiling system. This effect offers significant potential for future development and, in particular, opens up considerable possibilities for single mode fiber profiling.

After a description of the measurement set-up results are given for lasers and single-mode fibers. Distortions of mushroom stripe lasers were analyzed and side lobes of fibers down to 70dB lower than main lobe examined. Lenses have been optimized for laser to single-mode fiber coupling.

Measurement accuracy of mode field diameter (M.F.D.) by far field pattern method and the characteristics of M.F.D. in viewpoint of its definition are examined. It is found that M.F.D. by second moment definition becomes larger than that by gaussian fit defini-tion with a decrease in normalized frequency V.

An accurate spot size determination by means of far-field measurement using a dedicated integrating sphere is reported. The spatial intensity distribution is obtained by varying the spatial filtering as produced by moving the fibre in front of a sphere aperture.

For the dispersion-optimized range of power-law refractive-index profiles in optical fibers the phase and eigenvalue integrals appearing in WKB solutions are solved. These new solutions provide an explicit description of the electromagnetic field distributions and phase constants in weakly guiding fibers with power-law refractive index profiles in terms of usual fiber parameters.

A deliberate employment of mode mixing is utilized to make a source that illuminates a fiber in the simulated "steady state distribution". Mode mixing is achieved by combining power exchange due to bending and micro-bending. Discussing these effects design criteria are derived and a practical solution is given. These mode mixers have been used for fiber evaluation and high quality device assembly.