In the B-ISDN using satellite between geo-satellites and earth stations, the laser having high security and broad band width has to be used as a carrier for transmitting massive information of visual, vocal, and high rate data. In this paper, by computer simulation we analyzed the number of optical detector array of optical satellite communication downlink in case of using channel coding and no channel coding for BISDN between geo-satelites and earth stations under clear weather condition. It was supposed that 1 watt semiconductor laser was used and as modulation method, the binary FSK was used. The data rate of 10Gbps was used for B-ISDN. Also, hardly affected by atmospheric absorption 1.55$\mu$m wave-length was used to reduce influence of dispersion and chirp generated at a high speed transmission. We analyzed the received power, SNR and BER. The number of optical detector array was determined to satisfy for the BER less than 10$^{-7}$. Also, we ananlyzed the possibility of reducting the number of optical detector array in case of using channel coding. the number of optical detector array is one in the region where the elevation nangle is between 38$^{\circ}$ and 90$^{\circ}$ and two where the elevation angle is between 33$^{\circ}$ and 37$^{\circ}$ and three where the elevation angle is between 30$^{\circ}$ and 32$^{\circ}$ and increases per one as the elevation angle decreases per 1.deg.. So in the region where the elevation angle is 25$^{\circ}$, the number of optical detector arrays is eight. In case of using channel coding, the number of optical detector arrays decreases to five in the region where the elevation angle is 25$^{\circ}$. Therefore, we remaark the advantage of the channelcoding to decrease the size of received antenna and the number ob optical detector arrays.

this thesis proposes a new echo canceller that can be used in a fulll-duplex digital subscriber loopmodem. The modem suffers from nonlinear distortion such as transmitted pulse asymmetry, saturation in transformers, and nonlinearity of data converters. The proposed nonlinear echo canceller can compensate the nolinear distortion by using a nonlinear digital filter based on canonical pieceewise-linear (CPWL) function. Numerical results based on computer simulation are given in this paper. It is shown that the convergence characteristics depend on the initial values of weights of linear filters with absoluters and that the nonlinearity in digital-to-analog(D/A) converter can be compensated by a relatively small number linear filters with absoluters. It is also shown that the proposed algorithm has a faster convergence rate in comparison with Voterra algorithm.

The phenomenological loss equivalence method incorporated into the wideband lossy transmission line model is applied to the characterization of high desity digital transmission lines. The pulse propagation characteristics are analyzed using the calculated frequency characteristics and the discrete Fourier transformation. This approach has been verified by comparing the calculated frequency characteristics with the FEM and the esperimental results. This method is very suitable for computer-aided analysis of high density/high speed interconnection circuits because of the simple calculation as well as the calculation accuracy. We have found that pulse ftransmission speed and dispersion of hgih density digital transmission lines can be optimized by managing the conductor and dielectric losses in addition to the impedance matching.

The scattered electric field from a multilayered circular dielectric cylinder is caculated. Approximate boundary condition used in on-surface radiation boundary condition(OSRC) method has been applied to all the boundary surface of N-layered dielectric cylinder. It was assumed that scattered electric field at inner boundary surface in one region transmitted to the adjacent region at outer boundary surface. In the whole region, the unknown coefficients of electric field are acquired by the given incident electric field with ease. Electric field distribution at each boundary surface and the scattered electric field in free space are taken with the calculated unknown coefficients. the results obtainted were compared with those results that were used by regular surface boundary condition.

We derive analytic expressions for the spontaneous emission spectrum of a device using DFB structures including the effects of both facet reflections and the phase of a grating. In solution of coupled mode equations. Equations for the reations between the spontaneous emission spectrum and structure parameters of the device are derived to investigate the effect of structure parameters on the spontaneous emission spectrum and operating characteristics of possibility of evaluating the structure parameters of the device from its spontaneous emission spectrum.

In this paper, we present a new HPEJTC system which is capable of real-time multi-target recognition and tracking with better discrimination by extracting the phase signal of reference function from the JTPS of the conventional optical JTC retaining the amplitude signal of the input function. In order to test the correlation discrimination performance of the HPEJTC system, some experiments are carried out on the scenarios susceptible to the false alarms and missing in which many similar targets are periodically loacted. And, the proposed HPEJTC is analyzed to be the real function version of the POF and finally the possibility of the real-time implementation of the POF is suggested, because it can be implemented by using spatial light modulator, CCD detector and some other optical components.

Characteristics of electron cyclotron resonance (ECR) plasma thermal oxide grown at low-temperature have been investigated. The effects of several process parameters such as substrate temperature, microwave power, gas flow rate, and process pressure on the growth rate of the oxide have been also investigated. It was found that the plasma density, reactive ion species, is strongly related to the growth rate of ECR plasma oxied. It was also found that the plasma density increases with microwave power while it decreases with decreasing O2 flow rate. The oxidation time dependence of the oxide thichness showed parabolic characteristics. Considering ECR plasma thermal oxidation at low-temperature, the linear as well as parabolic rate constants calculated from fitting data by using the Deal-Grove model was very large in comparison with conventional thermal oxidation. The ECR plasma oxide grown on (100) crystalline-Si wafer exhibited good electrical characteristics which are comparable to those of thermal oxide: fixed oxide charge(N$_{ff}$)= 7${\times}10^{10}cm^{-2}$, interface state density(N$_{it}$)=4${\times}10^[10}cm^{-2}eV^{-1}$, and breakdown field > 8MV/cm.

In this paper, we ecamined the effect that MCM-L technique influencess on the design and fabrication of multichip memory modules in increasing the packing desity of memory capacity and maximizing its electrical characteristics. For that purpose, we examined the effective methods of reducing the area of module layout and the wiring length with the variation of chip allocation and the number of wiring layers. We fabricated a 256K${\times}$8bit SRAM module with eight 32K${\times}$8bit SRAM chips. The routing experiment showed that we could optimize the area of module layout and wiring length by placing chips in a row, arranging module I/O pads parallel to chip I/O pads, and equalizing the number of terminal sides of module I/O's to that of chip I/O's. The routing was optimized when we used three wire layers in case of one sided chip mounting or five wire layers in case of double sided chip mounting. The fabricated modules showed 18.9 cm/cm$^{2}$ in wiring density, 65 % in substrate occupancy efficiency, and module substrate and functionally tested to find out the module working perfectly.

A new Hydrodynamic model for the high energy tail electrons(Tail Electron Hydrodynamic Model : TEHD) is developed using the moment method. The Monte Carlo method is applied to a $n^{+}-n^{-}-n^{+}$ device to calibrate the TEHD equations. the discretization method and numerical procedures are explained. New models for the impact ionization and injection into the gate oxide using the tail electron density are proposed. The simulated results of the impact ionization rate for a $n^{+}-n^{-}-n^{+}$ device and MOSFET devices, and the gate injection experiment are shown to give good agreement with the Monte Carlo simulation and the measurements.

New formulas for the minimum noise figure and the optimum source impedance of microwave FETs are derived using the noise equivalent circuits obtained from the steady-state Nyquist theorem for multi-terminal semiconductor devices. The derived formulas manifest the relationships between the noise sources and the physical parameters of a noise equivalent circuit. Furthermore the formulas can explain the effect of gate leakage current on the minimum noise figure and the optimum source impedance. comparisons with the published experimental data confirm the validity and usability of our formula.

We have successfully fabricated a super low noise pseudomorphic HEMT(PHEMT) device with AlGaAs/InGaAs/GaAs sturcture by using improved T-Gate which have increased a large gaet cross-sectional area about two times in comparision with those of conventional T-gate processes. The PHEMSTs with 0.15$\mu$m-long and 140$\mu$m-wide gates have eshibited a super low noise characteristics, the noise figure of 0.45dB with associated gain of 10.87dB at 12GHz. The cut-off rewuqncy of the device is 94gHz with a transconductance of 418mS/mm.

A theoretical self-consistent thermoelectric model has been developed for optimal thermal design in the self-signal processing infraed detectors. The model is achived by employing the coupled thermoelectric equation which allows which allows the simultaneous investigation of the termal and electrical aspects of device behavior. The thermal limitation of detectivity and responsivity are determined by the enegy gap, carrier concentration, lifetime, and mobility as a function of the temperature. The calculated results indicate that the detectivity is decreased at bias fields above about 50 V/cm, because the performence is limiting by temperature when the bias voltage reached the level associated with Joule heating. It has been also found that the improvement in the mid-band modulation transfer function(MTF) may be restricted by increasing the bias fields. Further, the important paramerers in the thermal optimization of SPIR detector, such as temperature in the device, ambipolar velocity, element thickness and length, are also considered. The analytical study provides a mathematical basis for optimal design of such a photoconductive IR detector and the agreement between the experimental and theoretical results are seen to be good.

We discovered that each distinct shape of the roof-shaped peaks of (111) facets, which are generated on (110) cross-section of the directly bonded (100) silicon wafer pairs after KOH etching, can be mapped to one of three conditions of the interfacial oxide existing at the bonding interface as follows. That is, thick solid line can be mapped to stabilization, thin solid line to disintegration, and thin broken line to spheroidization. also we confirmed that most of the interfacial oxides of a well-aligned wafer pairs were disintegrated and spheroidized through high-temperature annealing process above 900$^{\circ}$C while the oxide was stabilized persistently when two wafers are bonded rotationally around their common axis perpendicular to the wafer planes.

The electrochemical impulse oscillations of the cathodic currents at the platinum group (Pt, Pd) electrode/(0.05M KHC$_{8}H_{4}O_{4}$) buffer solution interfaces have been studied using voltammetric, chronoamperometric, and electrochemical impedance methods. The periodic impulses of the cathodic currents are the activation controlled currents due to the hydrogen evolution reaction, and depend on the fractional surface coverage of the adsorbed hydrogen intermediate and potential. The oscillatory mechanism of the cathodic current impulses is connected with the unstable steady state of negative differential resistance. The widths and periods of the cathodic current impulses are 4ms or 5ms and 152.5ms or 305ms, respectively. The H$^{+}$ discharge reaction step is 38 or 61 times faster thatn the recombination reaction steps and the H$^{+}$ mass transport processes. The atom-atom recombination reaction step is twice faster thatn the atom-ion recombination reaction step. The two kinds of active sites corresponding to the atom-atom and atom-ion recombination reaction steps exist on the platinum group electrode surfaces.

Packet switching is an important aspect of ISDN. In the TDX-10 switching system, which is being developed as an ISDN exchange in Korea, the packet switching function is implemented on the basis of Case B scenario of X.31. After implementing the packet switching function the performance test must be executed to check whether the design objectives of performance are satisfied or not. However, it is not easy to set up test environment for performance measurement under full load conditions. This paper presents the design of the traffic generator which enables us to do performance test of the TDX-10 packet handler under full load conditions. To generate packet traffic we change only software programs without any change of the hardware system of the packet handler.

Measurement factors for the dielectric properties of low dielectric loss materials (tan${\dalta}{\le}10^{-4}$) were investigated using the dielectric rod resonator method. It was shown that the relative conductivity (${\sigma}_{r}$) should be controlled within a 5% to obtain the standard deviations of less than 0.07 for permittivity .epsilon.r and 0.06${\times}10^{-4}$ for tan.delta.respectively. Surface resistivity (R$_s$) could be reduced when the surface roughness of parallelled conducting plate was less than 0.07 .mu.m. Measurement error for the permittivity was $\pm$0.02% independent of probe loop size, whereas the error in Q value was reduced with the decrease in probe loop size and also with the increase in the absolute values of Q. Reliable Q values were determined with the probe loop size of less than 4mm. The accurate for the distance between the measuring probe loop and the sample could be obtained when the insertion loss of resonant frequency ranged -15dB - -30dB.