I am particularly pleased to have this opportunity to address a group of experts assembled to discuss a subject that has been vital to the technology which has been my principal concern for more than two decades. Had it not been for considerable earlier work in the field of reliability it is likely that many of the advances in international telecommunications technology could not have been brought to market with affordable risks and I might have been engaged all these years in building railroads and bridges, my secret fervor, instead of building submarine cables and operating international circuits.

An overview is given of the various standards committees addressing tests for fiber environmental performance and reliability. Applicable FOTPs of the EIA are are reviewed and improvements are suggested. Fiber strength theory is clarified in some detail. A tentative new standards working group should address a number of reliability issues in which standards are now lacking.

The strength and fatigue data for Corning's single mode fiber (standard fiber) whose surface composition is 100% SiO₂ are compared with those of Corning's HFR fiber (developmental fiber) whose surface composition is predominantly SiO₂ but doped with a minor amount of another oxide. The modification of surface composition of the latter fiber results in a substantially higher fatigue constant n which is principally responsible for observed improvements in both the short-term and long-term reliability of HFR fiber relative to standard fiber. The strength data, both in tension and bending, in various environments are compared for short gage lengths of both types of fiber and found to be consistent with their compositional differences. The dynamic fatigue constant of 10 m long fibers, under tensile loading in 100% RH, is found to be 30% higher for HFR fiber than that for standard fiber, consistent with the measurements for bulk glasses of same compositions. Similarly, the static fatigue data in two point bending of both types of fiber in 80°C liquid water yield orders of magnitude longer lifetimes for HFR fiber than for standard fiber for a given strain level. The corresponding n value in bending is also higher for HFR fiber than for standard fiber. Thus, both the tensile and bending fatigue data are consistent with surface compositional differences pointing out another way for improving fiber reliability in service. Furthermore, the bending fatigue data for HFR fiber do not show the presence of a "knee" in failure time vs. flexural strain plot, unlike standard fiber.

The goal of this work is to determine an allowable bending strain (bend radius) for optical fiber cable which is to be held in bending for long periods of time. The particular interest is in military and aerospace applications. In these applications it is desirable to use space efficiently and to route cable with as tight bending radii as possible. However, reliability cannot be compromised. In determining the lifetime of fiber it has been shown that the fiber must be tested under conditions that resemble those of actual use as closely as possible. For this reason static fatigue measurements are made in bending, and not in tension. In order to correlate static fatigue data with strength we have found that the strength must also be measured in bending. Most importantly for making reliable predictions the measurements must include long term tests; extrapolations from short term data are not reliable.

Mechanical reliability is a very important factor in determining the practicality of employing optical fiber systems in specific applications. High strength optical fibers are readily achievable, but silica fiber with a polymer coating will lose its strength over time when subjected to moist or wet environments. Fatigue testing has become the primary way to estimate mechanical reliability and predict lifetimes. There are several methods to measure fatigue and the disparities between them may affect the safety factor of systems designed on the basis of their respective results. To measure these differences, our laboratories have tested fiber fatigue by the most prevalent methods. Four different approaches were employed to determine the fatigue character of the fiber. Static fatigue was measured by mandrel wrappping 1.1 meter lengths, by two-point bending, and by direct loading 0.5 meter lengths. Dynamic fatigue on 1.0 meter lengths was measured using seven strain rates. These data, analyses and the fatigue parameters determined by each technique, are reported and the results compared. Implications of the test and analysis methods on lifetime predictions and other reliability issues are also discussed.

The thermal characteristics of excess signal loss in unjacketed multimode optical fibers are examined experimentally and theoretically. An unjacketed optical fiber is repeatedly wound under several different tensions and excess optical loss is measured at various temperatures using an Optical Time Domain Reflecto-meter. It is found that excess optical loss increases below 0°C. The fiber buckling model, used in the thermal analysis of jacketed fibers, is considered for the theoretical evaluation of fiber buckling caused by the mismatch of thermal expansion coefficients between the fiber and coating. The mode coupling theory is used to analyze microbending of the fiber caused by lateral shrinkage of the coating and Young's modulus increase at low temperature in the presence of initial fiber imperfections. A comparison of the data and theoretical results indicates that imperfections at the crossovers are the prime contributor to excess loss at low temperature. Results of the data show that excess optical loss at low temperature increases with winding tension.

The degradation mechanisms and reliability assurance strategies of InGaAsP semiconductor lasers that are currently being used in many commercial lightwave systems are discussed.

Today a user of laser diodes faces a complex problem when assessing the performance characteristics and expected life of different products. This paper describes the various practices that have proliferated in the absence of needed standards. In addition, a number of quality and reliability implications related to these practices are explored.

The environment can affect optical and electronic properties of optoelectronic devices. Here, noticeable alterations in spectral shape and wavelength of peak emission in light emitting diodes are shown to occur with changes in environmental gas and pressure. A relationship between band bending and surface charge density is derived.

Electrical and optical properties of light-emitting diodes and photodiodes can change significantly in vacuum. These changes involve alterations in forward and reverse current-voltage and capacitance-reverse voltage characteristics, time response, spectral characteristics, quantum efficiency, diode ideality factor, etc. Here, attention is centered on alterations in electrical properties of imaging CCDs in vacuum. Such alterations include increased dark signal, increased responsivity at longer wavelengths in particular, and non-linearity of response.

Monolithic laser arrays, used in a "cold stand-by" redundancy mode, offer a means of achieving high availability with high packaging density in fiber-optic computer data link transmitters. A model is proposed for studying the possible transmitter availability improvement, compared to the single laser case, and some theoretical results are described.

This paper summarizes the development of high-power semiconductor diode lasers. Both structural and material considerations, and lifetest methodology appropriate for the high power devices are discussed. Also included are the lifetest data of representative state-of-the-art high-power diode lasers.

Extensive environmental tests followed by a high temperature, powered on life test have been performed on PINFET receivers manufactured at PCO, Inc. These receivers have completed over 6400 hours of powered on life test at 85 Degrees Celsius. Life time calculations cannot be performed since 63% failures have not occurred yet. This paper addresses initial failure modes of earlier construction devices, subsequent design and manufacturing improvements and the latest el ectro-optical, life and environmental performance of these high reliability PINFET receivers.

This paper contains a discussion of Bellcore's proposed channel availability requirement for fiber optic transport systems used in distribution feeder applications. A representative model of the distribution network is first constructed which includes a high capacity fiber optic system running to a distribution hub. An overall availability requirement, consistent with present availability requirements, is applied to this model. The allocation to the fiber system is determined by truncation of the current interoffice channel availability requirement which is prorated with distance. Further allocation of downtime to the electronic hardware reliability failures yields an objective that can be directly compared with the results of electronic hardware reliability models. This requirement is now included in Bellcore's proposed "Generic Reliability Assurance Requirements for Fiber Optic Transport Systems."

The life of optical fibers under stress for an extended period of time is limited by static fatigue caused by stress corrosion in the presence of moisture. In order to predict the life of wound optical fibers, it is necessary to accelerate the aging process by simulating the storage environment (stress, temperature, and humidity) in a short period of time. Existing environmental test systems have been proven useful in the simulation of the storage environment; however, the data is limited due to the manual mode of operation. An automated environmental simulation model is developed to control, collect, process, and analyze optical loss data while measuring temperature and humidity. The environmental conditions for optical fibers wound for various applications are simulated in order to understand the interrelationships between wound fiber parameters including spool composition/design, winding tension, adhesives, and fiber cable design. Experimental investigations are carried out to expose wound optical fiber to simulated environments while monitoring changes in the optical and mechanical characteristics of the fibers. Based on the preliminary results of the data obtained, the automated simulation system is proven acceptable for performing routine modeling and evaluations. The automated system is a valuable instrument to aid in the characterization of optical fibers.

As we have already mentionned in the past, OTDR is a very interesting tool to monitor several ON/OFF sensors multiplexed along a single fiber. Indeed, Time Domain Multiplexing allows to easily monitor the time position of optical pulses reflected into a fiber line by optical microswitches or by any optical device whose reflection coefficient depends on a physical parameter (temperature, pressure, and son on). Moreover, such a technique allows to use very well known components such as pulsed laser diodes, couplers, avalanche photodiode and particularly multimode fiber whose commercial availability and reliability are now well established.

This paper will review the state-of-the-art status of optical fiber splices and connectors from the standpoint of component reliability. We discuss the reliability criteria used for different types of systems and also how system configuration is both affected by and affects component reliability.

The longterm reliability of fused fiber optic couplers is becoming an increasingly critical issue as the number of fiber optic systems being installed continues to rise. Components used in such systems are expected to meet their performance specifications over time periods measured in years. Maintaining such periods of stable and lasting environmental performance requires that particularly strong attention be paid to the component failure modes. In the case of fused fiber couplers, the specific steps of the manufacturing process and the design of the component package can interact in complex ways to influence environmental sensitivity and instigate performance failures.

The panel discussion held Tuesday morning (8:00 AM to 12:00 AM) on August 18, 1987 was tape-recorded. The following summary is transcribed from the tapes and edited by the chairman.

Test procedures and standards are currently being developed to validate and verify the effects of ionizing radiation on optical fiber components. The results from these tests can vary significantly with variations in the actual test setup and test requirements. It is important to recognize these sensitivities and monitor or control them to eliminate any inconsistencies in the test results. This paper will address the effects of ionizing radiation on optical fiber components, the development of test procedures and standards, and the comparison of a few test parameters to the final test results.

Glass optical fibers have high strengths but under tensile stress and in the presence of aqueous media their strength diminishes through fatigue. Analogous to the semiconductor field, where the electronic properties of chips are protected by passivating coatings, optical fiber strength can be protected by passivating the glass surface. Optical fibers have been developed with such a primary coating/cladding, namely HCS^R fibers. These fibers not only have high initial strength but also have excellent fatigue resistance. Their stability to environmental exposure is demonstrated by comparing static fatigue results for exposure to hot water, to boiling water, to steam and to acid/base solutions ranging from 10 M acid to 10 M base to results obtained in ambient water. HCS fibers maintain high static fatigue parameters at elevated temperatures and exposure to these temperatures under low stress does not change their fatigue behavior. Except at the extremes in pH, the static fatigue behavior is unchanged across the acid/base spectrum. This insensitivity to environmental changes and the good fit of the data to the power law model permits the prediction of lifetimes for long term exposure to aqueous environments.

We report an experimental and theoretical investigation of the effects of doping and processing on precursor defects in graded index multimode fibers. Fabrication parameters that significantly influence radiation sensitivity have been identified. In particular, we examined the role of fluorine doping in defect formation and its relationship to radiation sensitivity. The experimental effort included fiber fabrication and radiation-induced loss measurements on graded index, Ge-doped core fibers. Fluorine was added to the core and/or the cladding of test fibers. Two critical parameters, barrier layer thickness and core dopants, have been identified and correlate with induced loss. In addition, the reproducibility of both fiber fabrication and measurement with respect to induced loss has been tested and found to be excellent. Induced loss was found to be proportional to Ge concentration in the core; however, the trend with fluorine doping was less clear. The experimental results are consistent with molecular dynamics simulations which indicate the types and numbers of structural defects in the glasses. The simulations revealed significant differences in defect types and concentrations among glass corn-positions that included pure silica, Ge-doped silica, and Ge/F-codoped silica. Fluorine codoping decreases the number of germanium-related defects but increases the number of defects associated with silicon.

Modern ships are more and more fully equipped with auxilliary control and supervising sys-tems. These systems, in turn, are more and more frequently built with the aid of fibre optics. Fibre optics ensures proper reliability of such a system in a harsh marine environment. This environment includes: marine climate containing salt spray, mechanical vibrations and shocks, electromagnetic and magnetic interferences, changing temperatures etc. All of these harmful reactions may influence the system /worst case marine environment/ during a single sailing cruise. We will consider here the worst case environment for a marine-grade fibre optic system.