In this paper, an ultrabroadband optical access system for a future multimedia infrastructure, Gigabit-to-the-home (GTTH), is described. GTTH is based on a fiber-to-the-Home (FTTH) architecture and offers a very large downlink capacity of 2.5 Gbps. Because of this large downlink capacity, various broadband communication services and digital broadcasting services with hundreds of video programs can be simultaneously provided. GTTH also has high flexibility due to its multiple-interface architecture based on a bit multiplexing/demultiplexing scheme, which enables hybrid integration of multiple systems regardless of their signal formats, protocols, and transfer modes. A laboratory model of GTTH was developed. The optical network unit was achieved with small size, low power consumption and high receiver sensitivity by using a one-chip receiver IC, a high-speed CMOS demultiplexer, and a super lattice avalanche photo diode, which are newly developed for GTTH. The system demonstrates the service integration capability by simultaneously providing an ATM-based broadband communication service with the transmission capacity of 156 Mbps and multichannel broadcasting services with 300 video programs with MPEG2 format.

In this paper we present the architecture, operation, and the implementation of a prototype system, flexible access system (FAS), developed by CHT T.L. for FITL applications. FAS adopts fiber to the curb (FTTC) topology and provides POTS, ISDN BRA, and T1 services. The system consists of three main parts, namely, fiber central office terminal (FCOT) at switch side, fiber nodes (FN) at customer side, and a monitor station. FCOT and FNs are connected through an STM-1 fiber distribution ring. The hardware structure and the operation principle for each module are described. FAS contains the mux-demux module, POTS, ISDN, and T1 central office/customer service modules, operation modules, and power module. In the design, each central office service module in FCOT simulates as a CPE to terminate signals from switch, and will separate the data and signaling and map them into DS1 format. Mux/demux module will add/drop these DS1, put the operation information in data communication channel (DCC), and convert them to STM-1 optical link. Each customer service module in FN simulates switch line cart to regenerate the signal for customers. The prototype system implementation confirms the benefits of introducing SDH add/drop technologies to FITL system to provide basic telecommunication services. The design experience reveals that the clock synchronization method and noise reduction technique can significantly influence the performance in such a system. The test results that demonstrate and verify the designed functionality and the service quality of the prototype system will be presented.

For cable-to-the-air network providing a seamless access network in both indoor and outdoor, direct optical switching CDMA scheme is newly proposed to multiplex any types of radio signals for two types of bus link such as optical switch connection and optical coupler connection. It is clarified that in the optical switch connection system, by introducing the additional optical gain at each radio base station, the performance can be improved compared with the optical coupler connection system.

Wireless access in hybrid fiber/coax (HFC) networks can integrate two major telecommunication infrastructure, i.e., wireless and HFC networks. This paper studies the technical feasibility of integrating IS-95 code-division-multiple- access (CDMA) wireless networks and HFC networks. Capacity estimation of an HFC optical node-based CDMA cellular systems using a return-path channel model is provided. In addition to multiple access interference in IS-95 CDMA networks, the return-path channel model takes into account ingress noise, impulsive noise, and micro-reflections in a coaxial cable transmission system, and laser diode clipping- induced intermodulation noise in an optical fiber link. Furthermore, in our Monte Carlo simulation, we have assumed slow flat fading for radio propagation, and ideal timing recovery and power control.

We propose and demonstrate a fiber-optic feeder network for microcellular CDMA personal communication service. The proposed network is based on a passive double-star architecture. The relaxed CNR and dynamic range requirements of CDMA signals allowed the use of double-star architecture. The proposed network was demonstrated by using various light sources to examine the possibility of using inexpensive components. The result shows that this network could be implemented cost-effectively by using a Fabry-Perot laser for the downlink and LED's for the uplink.

In practical analog/digital hybrid optical transmission, carriers are modulated by video signals or digital data, and the amplitude of a multiplexed signal composed of these modulation signals is more compressed than that of the carriers. This causes a decrease in the frequency of clipping of the multiplexed signal at the laser threshold. Consequently, the BER of the M-QAM signal in a practical hybrid transmission is lower than that of the experimental results for same optical modulation index (OMI). However, it is difficult to prepare many practical modulation signals for experiments in a laboratory. Therefore, there is demand for a BER analysis method for a multiplexed signal that includes modulation signals needs to evaluate the BER and determine optimum OMI in a practical hybrid transmission. In this paper, we describe such a BER analysis method that can effectively estimate the BER in a practical hybrid transmission by using representative profiles of the modulation signals. In practical systems, a black pictures gives the largest averaged amplitude for the AM-VSB signal, and the most severe conditions for clipping occurrences. However, in such systems, the BER was greatly improved over the BER of a multiplexed signal of carriers. Furthermore, BER degradations due to clipping can be neglected for the AM-VSB signals in setting a practical OMI range.

In this paper, we propose and demonstrate a novel and effective wavelength-matching scheme for wavelength grating routers to match their transmission peaks to the wavelength assignment of the data channels without any reference sources. The unused portion of the amplified spontaneous emission (ASE) power of the Erbium doped fiber amplifiers is utilized as the monitoring light source and fiber Bragg gratings of a particular center reflection wavelength is used as the wavelength reference. The scheme supports in- service monitoring and will not degrade the performance of the data channels. It is also insensitive to dynamic ASE power variations arising form channel ad-drop. We have experimentally demonstrated the proposed wavelength-matching scheme for WGR and analyzed the scheme numerically for design optimization.

In wavelength division multiplexing (WDM) transmission systems, it is a problem that the transmission characteristics are degraded due to the fiber nonlinear effects. These nonlinear effects are caused by the pulse collision due to the group velocity difference between different channels. In this paper, a novel WDM transmission system with the dispersion slope management is proposed. The system manages not only dispersion but also dispersion slope and group velocity differences to suppress the third order dispersion and nonlinear effects. Our numerical results shows that the proposed system can suppress both of them and has good characteristics in long haul transmission.

Homodyne crosstalk causes severe system performance degradation in optical networks by beating with the desire signal. While Gaussian approximation overestimates the performance degradation, for a single dominant crosstalk source, the exact noise probability distribution and closed- form error probability is derived in this paper.

Multi-wavelength cross-connects (WXCs) will play a key role to provide more reconfiguration flexibility and network survivability in wavelength-division-multiplexing (WDM) transport networks. In this paper, we utilize three different fiber-Bragg-grating (FBG) based P-type, S-type, and N-type building blocks with optical circulators and related control devices for constructing large rearrangeably non-blocking N X N WXCs. The P-type building block is composed of certain 'parallel' FBG-element chains placed between the control device of two large mechanical optical switches (OSWs). The S-type building block consists of a 'series' of FBG elements and the control device of 2 X 2 OSWs. The non-switched N-type building block includes a 'series' of FBG elements with appropriate stepping motor or PZT control devices. Large N X N WXC structures, with minimum number of required constitutive elements, based on a three-stage Clos network are then constructed. We investigate their relevant characteristics, compare the required constitutive elements, and estimate the dimension limits for these WXC architectures.

A multiwavelength source is generated by four-wave-mixing in nonlinear dispersion-shifted-fiber with an accurate frequency separation of 100 GHz defined precisely by two seeding wavelengths. The multiwavelength source can find wide applications in dense WDM systems or spectrum-sliced WDM access networks.

Optical circuit-switch WDM network has the characters that the transmission data length at one certain connection can be huge, while the different connects switch can be slow. So a channel with low capacity can be sued to form the control network in order to manage the WDM network. No. 7 signalling network which exists logically in public telephone network is fit for management of the WDM network. The method works practically, because both the telephone network and optical fiber exist in any place of the world. The No. 7 signalling network is high reliable one. The new technology is studied both theoretically and experimentally.

We propose and experimentally demonstrate a novel variable bit-rate limiter (BRL) for ASK optical transmission systems. The BRL operates on the sampling principle with an optical switch inserted anywhere between the terminals in the fiber link. To facilitate the design of the BRL, a theoretical model of bit-error-rate for the proposed BRL is developed, and shows good agreement with the experimental data.

Next generation fiber-optic transmission systems adopt new line codes for considerably simplifying clock recovery . The major concern in implementing such new transmission schemes was the influence of jitter accumulation. Jitter generation and accumulation in a chain of repeaters were studied in a previous paper using a 1B4B type line code 2• Jitter accumulation in a chain of 20 repeaters was tested with two types of line codes. The one is a 1B4B type code which is promising for extremely simple implementation. The other is a 3B4B/1B4B multipled block code which features in simple multiplexing. It is seen that Type A jitter (high frequency component of jitter 3) accumulates in proportion to number of repeaters in a chain N. It is shown that an extremely simple bandpass filtering with a Q as low as 5 can be used in each repeater for drastically reducing the accumulation of Type A jitter. Two types of line coding are tested in a chain of 20 repeaters to prove that the amount ofjitter accumulation can be substantially reduced. Keywords: Digital transmission, Simple implementation, Clock recovery, Jitter accumulation, Jitter suppression

Photonic switching, is an essential synergetic approach in optical networks, providing virtually unlimited communication bandwidth and transparency to the data rate and encoding, has been developed to provide high bandwidth and avoid the repeated optical-to-electrical (O/E) and electrical-to-optical (E/O) signal conversions. The 2 X 2 directional coupler is a common switching element used in photonic switching networks. Due to the imperfect coupling energy in one path through the another path, crosstalk occurs. A faulty switch is defined as a switch that produces crosstalk beyond the acceptable level. A blocking network, say Dilated Omega Networks (DON), are discussed. One of the characteristics of DON is that the input signal and crosstalk signal will not pass through the same output switch. It relaxes the designs of diagnosing fault algorithm compared to that of Dilated Benes Networks, especially for the reduction of test needed, saving time and effort for the cases, such as single-path-multiple-faults, multiple-path- multiple-faults and crosstalk symmetry. Detail proofs and more examples will be addressed in this paper.

Due to the periodic variations of high power solitons, ultra-sharp pulses are created, and the analysis of the propagation behavior is computer intensive because of the requirement of large sampling density, in order to make the analysis more effective, we have implemented a time domain analysis technique to analyze the propagation behavior of high power solitons in an optical fiber. For a relatively small number of sampling points, the distribution of sampling points at different point of the fiber has been controlled effectively according to the shape of soliton pulse. We found that by a real time effective control of sampling distributions, the propagation behavior of high power solitons can be studied more effectively and the required number of sampling points is drastically reduced.

We propose and demonstrate a technique to increase the output power of a DFB laser transmitter from 6 mW to 16 mW, while still keeping the resultant composite second orders (CSOs) and composite triple beats (CTBs) meet standard CATV system requirements. This technique is based on a novel pre- distortion circuit, which can suppress laser's bias- dependent CSOs and CTBs significantly. The operation principle of this pre-distortion circuit is presented in this paper.

Conventional linearization techniques such as electrical pre-distortion or optical feedforward methods have been applied to Mach-Zehnder interferometer (MZI) modulators for CATV transmission systems. However, both methods can only ensure that the output of the MZI modulator meet the standard CATV requirements on composite second-order and composite triple beat, but cannot guarantee a satisfactory end-to-end system linearity. In this paper, we propose and demonstrate a novel linearization technique, which can simultaneously suppress the nonlinear distortions generated in a transmission system and increase the received carrier- to-noise ratio. This technique is based on offsetting the MZI bias voltage from its inflection point. Both analytical and experimental results are presented.

The repeaterless transmission distance of a 1.55 micrometers external modulation system using an erbium-doped fiber power amplifier is generally believed to be limited by stimulated- Brillouin-scattering (SBS). In this paper, by using both analytical and numerical methods, we find that even when the SBS threshold can be increased well beyond approximately 24 mW, the maximum repeaterless transmission distance will still be limited by self-phase-modulation-induced. Our analysis is based on the nonlinear amplitude envelope equation in a lossy, dispersive and nonlinear single-mode- fiber link at 1.5 micrometers , and our numerical approach uses split-step Fourier method.

In this paper, an integrated monitoring architecture base don telephone maintenance networks for the optical CATV networks is proposed for the first time. Part of the field trail result is also shown. The integrated monitoring system includes the element monitoring sub-system (EMS), the RF performance monitoring sub-system (PMS), and the fiber monitoring sub-system (FMS). The EMS can monitor the active devices, including the optical transmitters and optical receivers in optical CATV networks. The PMS can monitor the RF performance of carrier-to-noise ratio, composite second order, composite triple beat, and per channel RF power level. The FMS can monitor the state of the fiber, and locate the fault for the point-to-point and the passive branching fiber networks. Both polling and alarm call-back methods can be executed. All the polling instructions and monitoring data are transmitted through the telephone maintenance networks. These three sub-systems share the common platform based on NT operation system, and user can access these three sub-systems just from a terminal. This integrated monitoring system not only provides the surveillance function, but also offers the administration function with user-friendly interface.

In the future digital video transmission, a multichannel subcarrier multiplexed system with M-ary quadrature- amplitude-modulation (M-QAM) signals over fiber-optic transmission is an excellent candidate. A system using 1.3 micrometers semiconductor optical amplifiers (SOAs) for increasing the span length or for power splitting is very economical at the situations of the penetration of 1.3-micrometers single-mode fiber with the having installed 1.3-micrometers optical transmitters. However, the significant intermodulation distortions (IMDs) resulted from the signal-induced carrier density modulation of SOA limits the system capability, including channel capacity and power budget. A method using the external light injection to the SOA used in a 77-channel 64-QAM signal transmission system as an in-line amplifier is capable of increasing system power budget. From the viewpoint of system design, two approaches are addressed, one is the optimization of the in-line SOA used the external light injection, the other is the consideration of the optical modulation index per channel (OMI/ch) applied to transmitted. From the experimental results, the pumping wavelength and the pump power in approach one are mainly decided by the requirement of BER and the dynamic range of SOA, in addition, choosing the pump resulting in a smaller accompanied signal gain reduction can enlarge the system power budget. In approach, the optimum position of the in- line SOA incorporated is determined by the value of applied OMI/ch.

We present a directly modulated NTSC AM-VSB 80 channels fiber optical CATV long-distance transmission systems by using 1550nm digital DFB laser diodes, half-split-band, wavelength division multiplexing and EDGA-repeated techniques. Good performances of CNR, CSO and CTB are obtained after 80 km standard single mode fiber transmission. The direct transmission systems are simpler and cost lower than externally modulated transmission systems. 8

We propose and verify through computer simulations the use of a precoding technique to eliminate the nonlinear distortions in an M-QAM channel which is transported along with multiple AM-VSB channels by a laser diode. We show that the precoding technique can completely remove the clipping- induced bit-error-rate floor at a cost of a 3 dB signal-to- noise ratio reduction.

We demonstrate an in-service fault-locating and supervisory technique for multi-staged CATV optical distribution networks. The requirements of ultra-high dynamic range of an optical time domain reflectometer (OTDR) to diagnose the network can be tremendously relaxed by the introduction of optical switch and wavelength-division-multiplexing (WDM) technique. Any faults of the fiber links whether before, after, or between optical splitters can be located. The system operation is experimentally demonstrated in a two- staged eight-optical-node CATV distribution network. The degradation of system performance of carrier-to-noise ratio, composite second order, and composite triple beat due to in- service OTDR operation can be neglected. This surveillance system can provide real-time testing, saving the cost of dispatching a technician to perform routine testing and reducing the time taken to identify faults after an outrage occurs.

Long-period grating (LPG) is the transmission type grating, that light is coupled from the LP₀₁ mode to forward- propagation cladding modes LP_lm thus providing the function of wavelength dependent losses in optical fiber systems. We design a wavelength selective grating filter with transmission loss -21.2 dB at wavelength of 1531 nm, which can be used as the amplified spontaneous emission suppressor to improve the gain profile in dual-stage erbium- doped fiber amplifiers.

Fiber Bragg gratings were fabricated by utilizing high pressure deuterium loading for photosensitivity enhancement and a phase mask for holographic exposure. The absorption peak near 1.71 micrometers was used as an indicator to monitor the evolution of deuterium diffusion into and out of a fiber, and the results were consistent with the theoretical modeling. Thermal hysteresis of the fiber Bragg grating was greatly reduced and thermal reliability was increased when annealing was applied.

Fiber grating sensors are widely used in the strain measurements for smart structures such as composite materials, bridges, aircraft, etc. In this paper, three successful approaches implemented to improve the accuracy for the fiber Bragg grating (FBG) strain sensing system are described and followed by experimental results. Firstly, two reference FBGs and a specific algorithm are used for the automatic calibration function. Secondly, the newly developed amplified spontaneous emission light source using Erbium-doped fiber was adopted to eliminate the wavelength reading error caused by the ripple. Thirdly, the pair sensors formed by two FBGs with different reflected wavelength were designed to simultaneously measure the temperature and strain of the structure. The strain reading variation of the sensing system is found to be 3.3 (mu) (epsilon) .

Hundreds of photosensitive fibers have been prepared by the procedure of hydrogen loading in SMF-28 fibers and exposed to the UV lamp. This process can temporarily prevent hydrogen in fibers from outgas and maintain the photosensitivity for 4-5 days. It has been observed that hydrogen will diffuse out the core of fibers within several hours if fiber prepared by hydrogen loading is not exposed to excimer or UV radiation immediately. Such preparation is convenient for mass production of fiber gratings. The photosensitivity of fibers is judged according to the maximum reflectivity of fiber gratings written by 248nm excimer laser and OH-ion absorption at 1380nm wavelength. We also conclude that forming Bragg gratings in fibers of above preparation by use of a (pi) /2 phase shift mask, where the diffraction intensity ratio of plus/minus first order to zero order is 45-55 percent, can reach above 30dB in reflectivity. Notably, the periodic modulation of refractive index in fiber Bragg grating is permanent after the verification of thermal aging test for 60 days.

A hybrid fiber grating-based temperature sensor with simple demodulation scheme has been constructed. Temperature variation of 1 degree C could be linearly resolved in a 40 degree C range by directly measuring the transmission power through a long period fiber grating with a bandwidth of 6.5nm.

A simple method is reported for transformation of a uniform fiber grating into a linear chirped grating and realization of independent tuning of grating's linear chirp degree and central wavelength. This method involves embedding a uniform grating into a textile laminated beam and creating an odd- symmetrical linear strain distribution along the grating versus its center with a three-point-bending and stretching setup. The grating's central wavelength and chirp degree can be tuned by adjusting the horizontal stretching range and vertical bending displacement on the beam independently. A simulated experiment for compensating the dispersion of a standard single-mode fiber over 100km for 10Gbit/s signal at 1550nm window is successfully demonstrated using such a tunable chirped grating with 10 cm in length.

This paper presents progress of collinear beam (CB) AOTFs as wavelength routing switches (WRS) in DWDM networks. Significant improvement in the CBAOTF bandpass response was achieved. The measured performance at 1550nm shows a FWHM of 0.5nm and a sidelobe bandpass below 22dB at a wavelength spacing of 0.8nm from the center wavelength. By using two cascaded dual polarization AOTFs, a 2 X 2 WRS can be constructed that yields an extinction ratio greater than 35dB.

In this paper, we investigate and compare the characteristics of erbium-doped superfluorescent fiber source (SFS) configurations using different flattening technique in terms of large line-width, high output power, and stable mean wavelength. Not only the intrinsic flattening technique by optimizing the erbium-doped fiber length and pump power, but also the extrinsic flattening methods by adding the fiber Bragg grating-notch filter or an appropriate length of the samarium-doped fiber are experimentally studied in double-pass backward configuration. The design criteria of these SFSs using different flattening techniques and their features are presented and discussed.

A single-longitudinal-mode erbium-doped fiber laser with passive multiple-ring-cavity (MRC) is proposed. The laser is fundamentally structured by inserting three different short ring cavities, which serve as mode filters, into the main cavity. When combined with a mode-restricting intracavity fiber Bragg grating, the MRC resonator ensures single- longitudinal-mode laser oscillation. The laser can successfully suppress the side-mode frequencies up to 1 GHz, and provide an output power of 23 mW with a side-mode suppression ratio of 51 dB at 1533 nm. The short-term linewidth of the laser output measured is about 2 kHz. The ability of this fiber laser as an AM transmitter source is also demonstrated.

Compact and stable multiwavelength lasers (MWL) are attractive for wavelength division multiplexing (WDM) system. Recently, multiwavelength Er³⁺-doped fiber lasers have been reported by several groups. In this paper, we proposed and demonstrate a novel fiber MWL by inserting a double-pass Mach-Zehnder (M-Z) interferometer as a comb filter in the fiber laser cavity. The two output ports of the fiber M-Z device are spliced with a piece of Er³⁺-doped fiber in between to provide the optical gain. One of the input ports is connected with a WDM fiber loop which totally reflects light at 1550 nm and transmits light at 980 nm, and the other input port is connected with a partially reflected fiber loop as the output reflector to the laser cavity. 10 lines lasing simultaneously are obtained when the Er³⁺-doped fiber is immersed into liquid nitrogen. By controlling the optical path difference between the two arms in the M-Z device, the lasing peak wavelengths and their spacings can be adjusted to meet the ITU recommendation for WDM applications.

FP resonator has a very nice character which can be easily used as an OADM device. It can drop the selected channels and reflect the other channels which don't satisfy the resonance requirements. It can be sued as both channels drop and bypass device. In this paper, the design of good reflection spectrum is considered. The factors which affect the reflection spectrum of FP resonator are carefully studied. When using two FP resonators which have both nice reflection and transmission spectrum, it can be realized an OADM node without a circulator.

In a wide area WDM network with a huge capacity, dispersion limits the scope of the network. Among new ideas of dispersion compensation, dispersion compensation fiber (DCF) is the best. However, DCF can not balance the dispersion of all channels because of dispersion slope. There is surplus dispersion in some channels, so that further compensation of the surplus dispersion is needed after the using of DCF. The FP resonator is a dispersion-tunable device. The surplus dispersions of all the channels can be compensated after the suing of FP resonator as a tunable receiver. The numerical results show that the radius of WDm network can be made three times as large as it would have been without this device.

One parameter of the optical fiber is the crosstalk caused by its non-zero resistance band. The tunable flat passband filter is needed to improve the two characters, as the channel spacing becomes narrower, and the protocol of WDM network becomes faster optical packet switch one. This is because the flat passband can tolerate shorter channel spacing so that a large capacity and less tuning time are received. The structure of flat passband combines the multimode and basic mode to form the flat passband filter. This device can be designed with large tunable area and a very narrow passband.

We developed planar beamsplitter integrated with an Er-doped waveguide amplifier to compensate for the loss of the beamsplitter. In thin-film waveguide, a vertical dimension is determined by an Er-doped layer thickness, whereas a lateral dimension is defined by a patterning process that usually involves photolithography and etching techniques. Etching of Er-doped glass is known to be a challenging process. Chemical etching usually results in rough surfaces. Dry etching of Er-doped glasses containing alkaline metals shows a low etch-rate problem. Some researchers have used an ion milling technique to form a ridge structure with variable success. However, high temperature reflow is still required to minimize side wall roughness. We have developed a new fabrication process for Er-doped 2D waveguides. The process does not require etching of an Er-doped film, and therefore is simple, economical, and highly reproducible in defining a lateral dimension of a waveguide. Under-cladding silica layer was grown by Microwave Plasma Assisted Chemical Vapor Deposition on silicon substrate. Patterning of waveguide was defined on the under cladding layer. After that, an Er-doped soda-lime silicate glass film was deposited with rf magnetron sputtering, and jointed with core layer region of waveguide beam splitter which has compatible mode profile with optical fiber.A 1.7-cm-long Er- doped waveguide shows 7.2 dB gain at 980 nm pump power of 40 mW. This demonstrates that the amplifier can provide an optical gain sufficiently to compensate for the splitter losses.

We report a newly developed passive fiber depolarizer using cascaded 2 X 2 single-mode fiber couplers in a recirculating ring structure. Narrowband light source is used. The advantages are that these are all-fiber devices; they are simple, and insensitive to environmental fluctuations. Due to the cascaded incoherent recirculating fiber-ring structure, a completely depolarized light has been achieved by averaging all polarization states with equal probability. The degree of polarization as low as -20 dB has been achieved. Design criteria and design principles are discussed. Experimental results show that this new depolarizer is insensitive to the input polarization state and can eliminate the polarization noise in a polarization-sensitive fiber-optic system.

An appropriate analytical model which deals with out-of- plane coupling between an elliptical Gaussian beam and an angled single mode fiber is developed and utilized for two potential applications. The first one is for fast search of the optimal coupling efficiency in an optical sub-assembly packaging process while the second one is for the inspection of concentricity between the up and low tooling in an optical sub-assembly packaging system.

A high-isolation polarization-insensitive optical circulator being reduced in size nd low polarization mode dispersion is disclosed. A transparent optical circulator comprises a first birefringent crystal, a first pari of Faraday polarization rotators having two Faraday polarization rotators aligned in parallel with the propagation direction of an optical signal, a second birefringent crystal, a second pair of Faraday polarization rotators similar to the first pair of Faraday polarization rotators, and a third birefringent crystal. The optical circulator with the reflector and the Faraday rotator together reduce the total size of the circulator, and improve the polarization mode dispersion. The insertion loss is under 1 dB and the polarization dependent loss is under 0.2 dB with change in a polarization state. A novel optical circulator configuration that doesn't employ any polarization beam-splitters (PBS) is introduced. Birefringent crystals are used instead of PBS to achieve a high-isolation polarization-insensitive optical circulator. The isolation of the innovated optical circulator is more than 65 dB at 1550 nm wavelength. Details on the design, fabrication, and characteristics of the circulator are presented.

In this paper, we study a multiple output fiber-optic storage loop including a semiconductor optical amplifier. Seven wavelength division multiplexing channels can use a common fiber-optical storage loop. Such a fiber-optical storage loop can provide optical signal buffer for asynchronous transfer mode switching systems. The dispersion relation of a multiple quantum well semiconductor optical amplifier is used for designing a wavelength division multiplexing fiber-optical storage loop. Theoretical model of the fiber-optical storage loop is built up. Spectra responses of the wavelength division multiplexing fiber- optical storage loop are studied in this paper.

The effects of optical feedback on the characteristics of a double pass backward (DPB) superfluorescent fiber source (SFS) are demonstrated and compared to those of the single pass backward (SPB) SFS. The major effects of the optical feedback are the formation of resonance cavity for a DPB SFS and the reduction of output power for an SPB one. An isolator with an isolation loss larger than 50 dB should be used for both backward SFSs to maintain a stable mean wavelength operation, large output power and broad linewidths.

A formation is presented to study the effect of carrier heating on the second-order harmonic distortion of semiconductor lasers under analogue-amplitude modulation. The results show that the second-order harmonic distortion will be enhanced by the effect of dynamic carrier heating if the modulation frequency is much smaller or much larger than the resonant frequency. On the other hand, it will be suppressed by the effects of dynamic carrier heating if the modulation frequency is close to the resonant frequency. For a typical carrier energy relaxation time within 0.5-2 ps, the results indicate that the effect of dynamic carrier heating alone on the second-order harmonic distortion may be insignificant. The results also suggest that the effect of dynamic carrier heating on the second-order harmonic distortion can become important if the carrier energy relaxation time or the dc bias current are enormously large.

A powerful numerical technique, the quadratic finite element method with Galerkin approach is proposed for analyzing the modal characteristic of single mode optical fiber with arbitrary refractive index profile. A generalized refractive index profile is defined, which can be characterized by few parameters. The closed form solution of the wave equation for optical waveguide is available only for step index profile. Hence for arbitrary refractive index profile, one has to look for approximate method of analysis. Most of the existing numerical approaches are good enough for the calculation of propagation constant, though some of them are not suited for arbitrary refractive index profile. However for the numerical technique can be made highly accurate so as to meet the above requirements, with only moderate computational efforts. The simulated results when compared with exact ones for step index profile, confirms the accuracy of the proposed numerical technique. Examples are given for simple step index fiber, dispersion shifted fiber, W-fiber, quadruply cladded fiber. The numerical approach can also be used in the analysis of multimode fibers. With multiple cladded fiber, we can have a wavelength span over which the dispersion is kept confined within a prescribed level so as to exploit the vast transmission capacity offered by single mode fiber.

The characterization of Z-cut LiNbO₃ optical modulators with novel complex electrode are discussed by using the finite-element method (FEM) and studied experimentally in this paper. The calculation results show that the configuration of the electrode can realize microwave-optical velocity match easily and reduce microwave electrode losses drastically, so the modulators with such electrode are candidate devices for future ultra-high-speed optical fiber transmission systems. The traveling wave electrode consists of upper and lower parts. Because the thickness of the hot electrode can be thicker, the microwave attenuation coefficient of the traveling-wave electrode (alpha) ₀ is able to be less than 0.35 dB/(cm GHz^1/2). The configurations of the electrode are optimized by criteria of the ratio of 3dB optical bandwidth to drive power. The results of optimization show that the properties of the modulators can be improved by increasing the electrode gaps. An electro-optic modulator with bandwidth 100 GHz, half wave voltage 6V is designed. The modulator is being fabricated in our laboratory. Some experimental results of the traveling- wave electrode are obtained. The characteristic of low microwave transmission loss of the electrode is demonstrated. The attenuation coefficient (alpha) ₀ measured by microwave network analyzer is as low as 0.35 dB/(cm GHz^1/2).

Wavelength division multiplexing and demultiplexing (WDDM) techniques are the two key technologies for upgrading optical communication system bandwidth. The use of WDM technologies not only provides high speed optical communication links, but also provide advantages such as higher data rates, format transparency, and self-routing. Over the past twenty years, many kinds of WDDM device technologies have been developed and demonstrated. WDDM devices using dispersive photopolymer or dichromated gelatin volume holographic gratings have been recently reported. In this paper, we report an integrated four-channel multimode fiber compatible WDDM system with four semiconductor lasers operating at 750, 780, 810 and 840 nm, respectively. The device is demonstrated using the combination of graded index lenses, photopolymer based holographic gratings and substrate-guided waves.