Font Size: a A A

Theory Analysis And Design Research On Optical Fiber Type Traveling-wave Modulator

Posted on:2005-06-13Degree:DoctorType:Dissertation
Country:ChinaCandidate:J S LiFull Text:PDF
GTID:1118360152980038Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
High speed and broadband electro-optic modulator is a key component in thefields of fiber communication system, CATV fiber transmission, solitoncommunication, and fiber sensor, etc. With the rapid development of communication,it is urgent to develop a high speed and broadband modulator. In the dissertation, the limitation of Ti:LiNbO3 wave-guide modulator has beenanalyzed. A novel design project —— LiNbO3 fiber type traveling-wave modulator,which can guarantee high speed and broadband modulation, is firstly proposed. Bothfinite-element method and neural network have been introduced to analyze andcompare the main characteristics including the modulating bandwidth, effectiverefractive index, half-wave voltage, characteristic impedance, transmissionattenuation of microwave, and insertion loss, of modulator with three different type ofelectrodes: ACPS, SCPS, CPW. The results show that the characteristics of themodulator are ahead of the world, for its two main advantages: broad bandwidth andlow half-wave voltage. In addition, we have the intellectual property right on themodulator. The neural network is firstly used to model both passive and active opticaldevices. Through the RBF neural network, many devices can be modeled, such as thefiber direction coupler, the noise characteristic of laser, and so on. Furthermore, theTi:LiNbO3 wave guide modulator with ridge structure researched by NTT lab issimulated using the GRNN, and the error is analyzed. It is identified the neuralnetwork model for the modulator is feasible and reliable. Since it has the advantages of rapidity, accuracy, expansibility, ANN model isused to simulate the fiber type traveling-wave modulator. Moreover, the simulationresult of GRNN is compared with the calculation of FEM and conformal transfer withGreen function approximation. The results of GRNN simulation are well agreementwith that of FEM calculation, and the error is tiny. The trained model has also beenapplied to optimize the structure dimensions of modulator. Through the theory analysis and GRNN simulation, the optimal CPW electrodedimensions are gotten, and the performance index of fiber type traveling-wavemodulator shows as follows: ?f =250GHz, Vπ =1.14V, Z =55.78Ω. Finally, thewhole manufacture process of fiber type traveling-wave modulator is introduced. II摘 要Specially, the technological process of LiNbO3 fiber M-Z interferometer is introducedin detail, which is a wholly new innovation. Each technological step is described, andsome steps are given the specific technological parameters and figure. All of theseprovide an important guidance for the subsequent fabrication and measurement of thefiber type traveling-wave modulator.
Keywords/Search Tags:lithium niobate fiber, fiber type traveling-wave modulator, finite element method, neural network, half-wave voltage
PDF Full Text Request
Related items