This paper presents the designing/manufacturing scheme of the diplexer for satellite transponder which is composed of 2-channel bandpass filters, coupled with E-plane T-junction, having symmetrical waveguide E-plane step discontinuity structures. The 2-channel bandpass filters of the diplexer are designed by the scheme of connecting ${\lambda}_{g}$/2 shunt stub to the waveguide E-plane, playing the role of heat sinks without attaching a special heat sinks and to give the profitable productivity and allow the low-cost manufacturing at mass production, tuning screws are eliminated which have been used to compensate the operating characteristics of manufactured filters. The 2 channel bandpass filters of manufactured 14/12 GHz diplexer for domestic Ku-band satellite transponder have the return loss of 25 dB, insertion loss of 0.3 dB in passband, and below-50 dB of stopband characteristics (T$_{x}$ : f$_{o}{\pm}$250 MHz, R$_{x}$ : f$_{o}{\pm}$220 MHz) at room temperature.

The complex permittivity and the linewidth of microwave ferrites are measured using the well known cavity perturbation method. The cavity perturbation method has been widely used for the measurement of complex permittivity and conductivity of low loss magnetic and dielectric materials at microwave frequencies. TT1-2000, TT73-2200, and G-113 samples ordered from Trans-Tech Inc, are tested. TE$_{103}$ and TE$_{106}$ rectangular waveguide cavities are fabricated and the ferrite sample of cylidrical rod and sphere shapes are prepared. The error between the measurement values and the supplier's data is less than 1 percent in case of $\varepsilon$’ and about 10 percent in case of linewidth ($\Delta$H). Worst case error analysis shows that our measured results are well within the error bound calculated from the accuracy specification of the measuring instruments.

The E-polarized scattering problem by a resistive strip grating is analyzed by the method of moments in the spectral domain. For an E-polarized scattering, the induced current density is expected to blow up at both edges. Then the induced surface current density on the strip is expanded in a series of the multiplication of Ultraspherical polynomials with zeroth order and functions with appropriate edge condition. The expansion coefficients are calculated numerically in the spectral domain. When the resistivity of the strip becomes 0, the numerical results agree with those of a perfectly conducting strip srating. And the convergence of the moment-method solutions is very much improved using the basis functions proposed in this paper, compared with that of the existing exponential functions.

In the international BISDN used satellite, the laser that has large BW has to be used as a carrier for transmitting a lot of visual, vocal, and data information. Interoptical satellite communication has now developed in theoretical and practical aspects. But the optical communication, between satellite and earth station, is hindered by atmospheric absorption, scattering, and turbulence. In this paper, it was supposed that 1Gbps information was transmitted by binary FSK and 50mW AlGaAs semiconductor laser was used as a optical source in the satellite communication link between geosynchronous orbit satellite and earth station. We analyzed the BER and the entire diameter of the noncoherently combined optical heterodyne receiver as el evation angle, and determined the number of the optical heterodyne rece ivers, which is necessary for the BER of the receiver to be less than 10$^{-9}$ by computer simulation under the clear weather condition. It is shown that the BER and the number of the optical heterodyne receivers decrease as the elevation angle increases. In the region used the same number of the optical heterodyne receivers, it is shown that the entire diameter of the receiver increases but the BER decreases as the elevation angle increases.

As an approach to develope the intelligent control scheme, this paper will propose an adaptive neural-fuzzy control scheme. The proposed neural-fuzzy control system, which consists of the Fuzzy-Neural Controller(FNC) and Model Neural Network(MNN), has two important characteristics of adaptation and learning. The error back propagation algorithm has been adopted as a learning technique.

In this paper, MODFET devices with various gate length are simulated using the Monte-Carlo method. The number of superparticle is 5000 and the Poisson equation is solved to obtain field distribution. The structure of MODFET is n-AlGaAs/i-AlGaAs/iGaAs and doping concentration of n-AlGaAs layer is 1${\times}10^{17}/cm^{3}$ and the thickness is 500.angs., and the thickness of i-AlGaAs is 50$\AA$. The devices with gate length 0.2$\mu$m, 0.5$\mu$m, 1.0$\mu$m respctively are simulated and the current-voltage curves and transport characteristics of that devices are obtained. Occupancy of each subband and electron energy distribution and conduction energy band in channel have been analyzed to obtain transport characteristics, and particles transposed from source to drain have been analyzed to current-voltage curves. Current level is highest for the device of Lg=0.2$\mu$m and transconductance of this device is 310mS/mm.