on the feed position; s 3.0, h s 0.762 mm, d s 6.0 cm, d s 5.196 r h cm, s 90Њ pimpedances are also presented, and the cavity-model solutions in general agree with the measured data. The deviation of the resonant frequency between the theory and experiment is within 0.6% in the case studied here. And from the results, it is seen that the input resistance level decreases with decreasing cylinder radius. To show the results more clearly, the input resistance at resonant frequency as a function of feed position for different cylinder radii is presented in Figure 4. The feed position is chosen from section AB of Ž . length d the height of the triangle; see Figure 1 . And z in h p this case also represents the distance between the triangle tip and the feed position. From the results, it is also seen that there exists a null input resistance at z rd ; 0.67. At this p h point, the excited electric field inside the cavity for the TM 10 mode is zero. In summary, cavity-model solutions of the cylindrical triangular microstrip antenna have been derived and analyzed. And since the computation of the cavity-model solution is w x usually more efficient than that of the full-wave solution 3 , the results presented in this paper can find applications in computer-aided design of triangular microstrip antennas mounted on cylindrical surfaces. ABSTRACT: A model for the terahertz optical asymmetric demultiplexer ( ) TOAD is presented, and its performance is assessed. A numerical ( ) model of a semiconductor laser amplifier SLA used as a nonlinear element is also gi¨en. Results of SLA carrier density, switching window, pulse switching, and a comparison of phase modulation and power gain for a frame rate higher than the recombination are presented. Also in¨estigated is the required optimum loop asymmetry for producing the desired switching window for channel selection, with reduced crosstalk, at the output port.