Effects of the external optical feedback on the optical transmission characteristics were evaluated for 3 kinds of commercialized DFB-LD modules for 2.5Gbps optical communication systems. As a result, intensity noise induced by external optical feedback via generating the external cavity modes deteriorated the transmission performances of sthe DFB-LD modules as much as 1.5dB, 1.2dB, and 1.0dB for the maximum feedback ratio of -8.8dB which corresponds to the worst reflection along the fiber transmission line. Additional penalty of 0.5dB was occurrred when state of polarization of feedback light was adjusted. From the experiments evaluating the isolation ratio of the single polarization optical isolator, decrease of the polarization dependent isolation ratio as much as 10dB from that of specification and the effective isolation ratio as much as 20dB from that of specification was mjor cause of high sensitivity of DFB-LD modules to external feedback. It was suggested that polarization independent optical isolator or single polarization optical isolator with isolation ratio specification of more than 40dB is required for 2.5Gbps DFB-LD modules with complete suppression of external feedback effects.

In the soliton transmission system for a long-length and high bit rate data transmission systme using the nonlinear/dispersive optical fibers, the improtant problem is the loss characteristics and is the limited transmission rate by interaction forces. In this paper, It is explained the reasons of moved time position for the soliton peak value due to interaction force sof adjacent solitons. And for the analysis of interaction force affection level in the losslessmedia, we define the percent parameter of error rate due to the interaction forces and propose the optimum time distance of adjacent solitons by peak value tracking methods. With the results, initial percent of error is approximately 50% when itme difference between adjacent solitons is 5 times of funddametnal soliton pulse width. And it is confirmed that maximum transmission length of th esolitons is approximately 38 times of fundamental soliton period, which the maximum allowable percoent of error is 50%.

This paper deals with the design of a recursively-indexed binary code for facsimile soruces and its performance. Sources used here are run-lengths of white pixels form higher-resolution facsimile. The modified huffman code used for G.3 facsimile is chosen for the performance comparison. Experiments confirm the fact that recursive indexing preserves the entropy of a memoryless geometric source: the entropy of recursively-indexed physical surce iwth roughly geometric distributin remains within 2% of the empirical source entropy. The designed recursively-indexed binary codes consist of a code applied to text-type documents and to graphics - type documents is compared iwth that of the modified huffman code. Numerical resutls show that the modified huffman code performs well for text-type documents and not equally well for graphics-tyep documents. On the other hand, recursively-indexed binary codes have shown a better performance for graphics-type documents whose distribution are similar to a geometric distribution. Specifically, the code rates of recursively-indexed binary codes with 60 codewords are from 8% to 20% of the empirical source entropy smaller than that of th emodified huffman code with 91 codewords.

In this paper a simpel algorithm for all-digial high speed symbol synchronization is proposed. The proposed algorithm has a structure based on BECM (band-edge component maximization). The algorithm requires only tow samples per symbol for its operation. We analyze and evaluate performance of the proposed algorithm. Simulation results reveal that the proposed algorithm has better performance than the gardner algorithm in narrowband.

In DS/SS systems it is necessary to synchronize the locally generated despreading signal with the received spreading signal to demodulate the received signal. The synch process between the two signals is usually accomplished in two steps : first acquisition then tracking. In this paper, an acquisition system aided by the neural network is proposed for the rapid and exact acquisition in DS/SS. the neural netowrk is composed o fthree-layered perpecptrons and trained by the backpropagation algorithm. The performance of the proposed system is analyzed and compared with ones of conventional systems using the sequential estimation technique under an additive while gaussian noisy channel. In all of th econsidered simulations, the proposed system outperforms conventional systems.

In this paper, a multiple-access algorithm for PCN is proposed. The proposed algorithm provides th integrated service of information soruces and can be operated stably in high load state. Given bandwidth is efficiently used for it in the microcell environment. And system performance can be improved through the statistical-multiplexing technque. In order to process the speech signal usually requiring real-tiem processing, we adopt a random access of AlOHA type for th ebasic protocol sturcture and assume the form of ALOHA-reservation. We have analyzed the performance of the proposed algorithm through system throughput and packet delay in the microcell environment.

The probability distributio function (pdf) of transmitted symbols should be asymmectric for recovering the received data in the 3rd order blind equalizer system. In this paper, we hav edesigned the blind equalizer using symmetric-asymmetric (SA) and asymmetric-symmetric (AS) transforms for less computational complexities and robustness in the noisy environments. The method of SA and AS transform was peformed using natural logarithmic operation. This paper proves that the proposed in this paper can be performed as real time operation. This paper proves that the compression factor k has no effet on transmitted symbols. Also 3rd order equalization method proposed in this paper can be performed as real time operation. As a result of computer simulatio, the computational complexity of proposed algorithm is reduced to be an half of 4th order method and MSE is enhanced as 10dB at the case of 4-PAM and 15dB at the case of 8-PAM respectivley. Conclusively, we have found that 3rd order blind equalizer can be implemented when the pdf of transmitter is asymmetric.

In this paper, we proposed the new medium access control protocol for the virtual token bus netowrk. The network is applied to inter-processor communication network of large capacity digital switching system and digital mobile system with distributed control architecture. in the virtual token bus netowrk, the existing medium access control protocols hav ea switchove rtime overhead when traffic load is light or asymmetric according ot arbitration address of node that has message to send. The proposed protocol optimized average message delay using cyclic bus access chain to exclude switchover time of node that do not have message to send. Therefore it enhanced bus tuilization and average message delay that degrades the performance of real time communication netowrks. It showed that the proposed protocol is more enhacned than virtual token medium access control protocol and virtual token medium access control protocol iwth reservation through performance analysis.

parallel bonding wires separaated with a screeing bonding wire are proposed and characterized in order to redue mutual coupling and parasitics of high-speed and high-density device packaging. The mehtod of moments (MoM) with the incorporation of the ohmic loss has been used in a wide range of frequencies. From the calculated results, we have found that the screening bonding wire effectively reduces inductive and capacitive crostalk levels more than 3dB. the parasitic self inductance is also reduced more than 12% by the screening effect. Therefore, for a general VLSi package, the packaging density can be increased more than 30% using the screening bonding wire. This screeing bonding wire and the analysis can be effectively used to reduce crosstalk and increase packaging density of high density devices.

In this paper, the analysis o fthe electric field distribution for TEM cell which is matched iwth 50 is performed, and the relations of variables for characteristic impedance are derived. Quasi-static approximations are used to calculate the fiedl strength of the internal field of TEM cell. The results of the improved method for analysis of the electric field is compared with that of R.J. Spigel. and the improved method for characteristic impedance and the results of numerical analysis are shown.

In this paper, we proposed an inductance loaded small type open-loop travelling wave antenna model for the suitable VHF/UHF frequency bands moisture measurement systems and analyzed its operating characteristics by galerkin's solution method. This antenna model enables to change the oeprating rsonance frequency from the frequency range 150 Mhz to 1.0 Ghz without changing the antenna design dimensions by only varing the load inductance values and it showed the similar operating charcteristics and narrowband resonance characteristics at each resonance frequencies. Also this model showed the uniformly distriubte dpower radiation characteristics above the loop axis and it can effectively suppress the spurious radiations from its narrowband resonance characteristics.

In this paper an MMIC VCO which can be used in a double conversion TV tuner is designed, fabricated and measured. The VCO is designed using the small signal method and fabricated using ETRI GaAs MMIC foundry. The 3x200$\mu$m gate width MESFET with 1$\mu$m gate length is used for an active device and MIM capacitors, spiral inductors, thin film resitors are used as passive elements. The VCO has output power of 10.95dBm at 1955 MHz with 5V bias voltage and 4V tuning voltage. The oscillation frequency change form 1947 MHz to 1964 MHz is obtaine dby an external varactor diode connected to the gate with a tuning voltage from 0 V to 6V.

In this paper, we propose a new fabrication method for poly-Si TFTs with a self-aligned offset gated structure by employing a photoresist reflow process. Compared with the conventional poly-Si TFTs, the device is consist of two gate electrodes, of which one is the entitled main gate where the gate bias is employed and the other is the entitled subgate which is separate form both sides of the main gate. The poly-Si channel layer below the offset oxide is protected form the injected ion impurities for the source/drain implantation and acts as an offset region of the proposed device. The key feature of oru new device is the offset region due to the offset oxide. our experimental reuslts show that the offset region, due to the photoresist reflow process, has been sucessfully obtained in order to fabricate the offset gated poly-Si TFTs. The maximum ON/OFF ratio occurs at the L$_{off}$ of 1.1${\mu}$m and exceeds 1X10$^{6}$.

The epi-CoSi$_{2}$ layer was formed by alloying a Co(120$\AA$)/Ti(50$\AA$) bilayer. In addition, the ultra shallow p$^{+}$n junction of which depth is about not more than 40nm at the background concentration, 10$^{18}$atoms/cm$^{3}$ could be formed by annealing (RTA-II) the ion implanted epi-silicide. When the temperature of RTA-I is as low as possible and that of RTA-II is moderate, the p$^{+}$n junction that has low leakage current and stable epi-silicide layer could be obtained. That is, when th econdition of TRA-I was 900$^{\circ}C$/20sec and that of RTA-II was 900$^{\circ}C$/10sec, the reverse leakage current was as high as 11.3$\mu$A/cm$^{2}$ at -5V. The surface of CoSi$_{2}$ appeared considerably rough. However, when the conditon of RTA-I was 800$^{\circ}C$/20sec or 700$^{\circ}C$/20sec, the leakage currents were as low as 8.3nA/cm$^{2}$ and 9.3nA/cm$^{2}$, respectively and also the surfaces appeared very uniform.

Detailed analyses have been presentd for a basic single-electron-logic-cell consisting of two single-electron-transistors (SETs) in series. The interconnection between two SETs has been treated as a coulomb island and the joint probability density function of all three coulomb islands (two from the SETs and one form the interconnection) has been exactly calculated. The average number of electrons in each coulomb island and the steady-state ouptut voltage have been calculated and analyzed.

A new technique wasdeveloped which obtains selectively the htick fully recessed oxidized porous silicon layer (OPSL) with good dielectric property. The porous silicon layer was ocnverted to thick fully recessed oxide (FRO) with 3-step (1${\mu}$m, 1.5${\mu}$m, 1.8${\mu}$m) by multi-step thermal oxidation (after 400$^{\circ}$C, 1 hour by dry oxidation, 700$^{\circ}$C, 1 hour and then 1100$^{\circ}$C, 1 hour by wet oxidation). The breakdwon field of the FRO was about 2.5MV/cm and the leakage current was several pA ~ 100 pA in the range of 0 of 90 pF. The progress of oxidation of a porous silicon layer was studied by examining the infrared abosrption spectra. The refractive index (1.51) of the fRO, which was measured by ellipsometer, was comparable to that of the thermally grown silicon dioxide (1.46). The etching rate (1600${\AA}$/min) of the FRO was also almost equal to that of the thermal oxide.

Accurate equivalent circuit modeling using multi-circuit optimization has been perfomred for detemining small-signal model of AlGaAs/GaAs HBTs. Three equivalent circuits for a cutoff biasing and two active biasing at different curretns are optimized simultaneously to fit gheir S parameters under the physics-based constrain that current-dependent elements for one of active circuits are connected to those for another circit multiplied by the ratio of two currents. The cutoff mode circuit and the physical constrain give the advantage of extracting physically acceptable parameters, because the number of unknown variables. After this optimization, three ses of optimized model S-parameters agree well with their measured S-parameters from 0.045 GHz to 26.5GHz.

A breakdown model of gate oxides under constant current stresses is proposed. This model directly relates the oxide lifetime to the stress current density, and includes statistical nature of oxide breakdown using the concept of "effective oxide thinning". It is shown tha this model can reliably predict the TDDB characteristics for any current stress levels and oxide areas.

Gate insulator for Si field emitter is usually formed by e-beam evaporation. However, the evaported oxide requires densification for a stable process and a reduction of gate leakage which results from its Si-rich and nonstoicheiometric structure. In this study, we have developed the process technology able to densify the evaporated oxide in H$_{2}$O ambient. Using this process, we have fabricted thefield emitter array with 625 emitters per pixel, of which gate hole diameter is 1.4.mu.m, for the pixel, anode current of 14.3.mu.A was extracted at a gate bias of 100V and gate leakage was about 0.27% of the total emission current.

To overcome the drawbacks of conventional LDMOSFETs, we propose RESURF EDMOSFETs which can be adapted in varous circuit applications, be driven without charge pumping circuity and thowe threshold voltage can be adjusted. The devices have the diffused drift region formed by a high tmperature process before the gate oxidaton. After the polysilicon gate electrode formation, a fraction of the drift region around the gate edge is opened for supplemental self-aligned ion implantation to obtain self-aligned drift region. This leads to a shorter gate length and desirable drift region junction contour under the gate edge for minimum specific-on-resistance. In additon, a and maximize the breakdown voltage. Also, by biasing the metal field plate, we can reduce the specific-on-resistance further. The devices are optimized by using the TSUPREM-4 process simulator and the MEDICI device simulator. The optimized devices have the breakdwon voltage and the specific-on-resistance of 101.5V and 1.14m${\Omega}{\cdot}cm^{2}$, respectively for n-channel RESURF EDMOSFET, and 98V and 2.75m.ohm..cm$^{2}$ respectively for p-channel RESURF EDMOSFET. To check the validity of the simulations, we fabricated n-channel EDMOSFETs and confirmed the measured breakdown voltage of 97V and the specific-on-resistance of 1.28m${\Omega}{\cdot}cm^{2}$. These results are superior to those of any other reported power devices for smart power IC applications.

We have adopted the strain compensated PBH(planar buried heterostructure) - LD in which the MQW active layer consisted of 1.4% compressively strained GainAsP (E$_{g}$ = 0.905eV) wells and 0.7% tensile strained GaInAsP(E$_{g}$ = 1.107eV) barriers grown by metal organic vapor phase epitaxy (MOVPE). We hav einvestigated effects of number of wells and the structure of the separate confinement heterostructure (SCH) layer in the strain-compensated MQW-PBH-LD. The threshold current, the external quantum efficiency, the transparency current density J$_{o}$, and the gain constant .beta. have been evaluated for uncoated MQW-PBH-LD. As the number of wells increases, the internal quantum efficiency and the transparency current density decreases, whereas the gain contant increases. The small width of the SCH layer shows the large internal quantum efficiency. The small internal loss and the large gain constant have been obtained by inserting the large bandgap SCH layer.

Electron beam lithography is one of the importnat technologies which can delineate deep submicron patterns. REcently, electron beam lithography is being applied in delineating the critical layers of semiconductor device fabrication. In this paper, we present a development simulation program for electron beam lithography and study the development profiles of resist when resist is exposed by the electron beam. Experimentally, the development parameter of positive and negative resists are measured and the data is applied to input parameter of the simulation program. Also simulation results are compared of the process results in the view of resist profiles. As a result, for PMMA and SAL 601 resist, the trend of simulation to the values of process parameters agree with real process results very well, so that the process results can be predicted by the simulation.

We derive the analytic equations for the transmission gain spectrum of a phase shift controlled (PSC) DFB filters with complexed coupled gratings considering both facet reflections and the phase of gratings using the transfer matrix method. The number of parameters of the equations is reduced by using the parameter of effective phase shift defined by the sum of the phase shift in a PSC region the effect of both facets reflections and the effective phase shift on the transmission gain spectrum and verify the validity of those equations from the computer simulation results. Computer simulation results show the PSC DFB filter with a pure index coupled grating has the widest tunable range and that with a pure gain grating has the largest side mode suppression ratio.

Since many typical 2.5 Gbps optical transmitter modules use a 50$\Omega$ characteristic impedance, they require relatively high voltage and high power sources compared to the 25$\Omega$ module. However, simple replacement of the 50$\Omega$ internal matching impedance with 25$\Omega$ results in bandwidth reduction and consequent problem of data transmitter module is proposed in order to expand the modulator bandwidth. From the calculated resutls based on accurate 3-dimensional inductance analysis, we have found that the series parasitic inductance is a dominant element limiting the bandwidth and the insertion of a 2.5pF capacitor in parallel to the 20$\Omega$ matching resiter can increase the 3 dB bandwidth about 1.4GHz wider. The time-domain results show the rise time (140 psec) without the compensation is greatly improved to 63 psec with the compensation. This capacitive ocmpensation can be implemented easily and be compatible with common manufacturing process of the optical transmitter module.

a simpel coordinate transformaton method is suggested that converts Schr$\"{o}$dinger's equation involving a position-dependent effective mass in a heterostructure to an equation involving a positon-independent effective mass. This method enables the conceptual study of the effect of the positon-dependent effective mass inserted between the divergence operator and the gradient operator in Schr$\"{o}$dinger's equation. It is also shown that the characteristics such as a transmission coefficient in various heterostructures involving a position-dependent effective masses can be obtained iwth ease by the suggested method.

We derive analytic expressions for the spontaneous emission spectrum (SES) of a multisection distributed feedback (DFB) structure device employing complex coupled gratings including the effects of both facets reflections. The multisection DFB structure device used in the analysis is a general model which allows the independent phase of a grating in each section, and the sections without gratings. The expressions are the same as those derived by makino and glinski in case the gratings are index coupled and the phase of a grating in each section, ${\varphi}_k$ is '0' which means the phase of gratings in the device is ocntinuous. The expressions for the SES of a phase-shift-controlled (PSC) DFB structure device using tunable devices are derived from the general expressions. The number of parametes of the expressions is reduced by using the parameter of effective phase shift defined by the sum of the phase shift in a PSC region and the difference of the phase of a grating in each active region. Equations showing the effect of both facets reflections and the effective phase shift on the SES are derived. The validaty of the equations is verified by computer simultions. Computer simulation results also show the possibility of evaluating the structure parametes of the device from its SES.

$\lambda$/4 phase-shifted DFB lasers with nonideal grating structure have been studied by using an effective-index transfer matrix method where material dispersion, vwaveguide dispersion and waveguide-structuredependent loss are taken into consideration. Nonideal grating structure in the center phase-shift region does not incur serous degradation of laser characteristics. Phase-shift error from the ideal shift of $\pi$ causes a decrease in the threshold gain difference and lasing wavelength shift and should be less than $\pi$/4 when residual facet reflectivity is 0.7%. also, positional error of the phase-shift should be less than 9% of the cavity length in order for the threshold gain difference to be decreased less than 10%.

In this paper, a simple scheme of two-stage free-space photonic swiching system for nonblocking optical interconnections has been investigated using holgorams. The multistage system requires a smaller number of interconnections for a given number of inpuut-output nonblocking ports than the single stage system does. Here hologram elements were used to change interconnection beam paths. In order to increase the idffraction efficiency of the hologram elements in photographic plates, a bleaching technique was used, which converts the amplitude hologram to the phase hologram. To show sthe feasibility of our optical interconnection system, it was implemented using the bleached hologram arrays and an LCTV spatial light modulator, and the sytem as a photonic switching system was demonstrated.

Hardware-software codesign becomes improtant to effectively sagisfy perfomrance goals, because designers can trade-off in the way hardware and software components work teogether to exhibit a specified behavior. In this paper, a hardware-software pratitioning algorithm is presetned, in which the system behavioral description containing a mixture of hardware and software components is partitioned into hardware part and software part. The partitioning algorithm tries to minimize the given cost function under constraints on hardware resources or latency. Recursive moving of operations between the hardware and software parts is used to find a near optimum partition and the list scheduling approach is used to estimate the hardware area and latency. Since memory may take substantial protion of the hardware part, memory cost is included in sthe hardware cost. Experimental resutls show that our algorithm is effective.

This paper has been presented a design of a POCSAG decoder in RT-level VHDL and implemented in a FPGA chip, and tested. In a single clock of 76.8KHz, the decoder receives all the data of the rate of 512/1200/2400bps and has maximum 2-own frames for service enhancement. To improve decoder performance, the decoder uses a preamble detection cosidering 9% frequency tolerance, a SCW detction and a ICW detection at each 4 bit. The decoder also corrects a address data and a message data up to 2 bits and proposes the PF (preamble frequency) error for saving battery. The decoder increases a battery life owing to turn off RF circuits when the preamble signal is detected with nises. The chip has been designed in RT-level VHdL, synthesized into logic gates using power view$^{TM}$ of viewlogic software. The chip has been implemented in an ALTERA EPF81188GC232-3 FPGA chip with 98% usability, and fully tested in shield room and field room. The chip has been proved that the wrong detection numbers of preamble of noises are significantly reduced in the pager system using PDI 2400 through the real field test. The receiving performance is improved by 20% of aaverage, compared with other existing systems.