Synthesis of dissipative systems using quadratic differentialforms: part II

For pt.1 see ibid., p.53-69 (2002). The authors discuss several important special cases of the problem solved in Part I. These are: disturbance attenuation and passivation, the full information case, the filtering problem, and the case that the to-be-controlled plant is given in input-state-output representation. An interesting aspect is the notion of full information, which we define in terms of the observability of the to-be-controlled variables from the control variables. When the system is given in state space form, we obtain conditions for the existence of a controller that renders a system dissipative in terms of two coupled algebraic Riccati inequalities. The controller turns out to be a feedback system with a transfer function that is proper, but, in general, not strictly proper. Another issue that we study in this paper is feedback implementability. We find conditions under which, in the context of synthesis of dissipative systems, a controlled behavior can implemented by a feedback controller     

Transient response of a microstrip line circuit excited by anexternal electromagnetic source

The transient voltages and currents induced by an external electromagnetic field along a microstrip line interconnecting active and/or passive components are studied by using a distributed-source transmission-line model. The influence of the angle of incidence as well as that of the microstrip geometrical and electrical parameters on the line response is analyzed. The instantaneous voltage and power induced on the loads by various types of pulse excitations are computed. Numerical results are obtained both for dispersive and nondispersive lines. A line feeding a transistor microwave amplifier is also considered. The results show the effects of the multiple reflections of the field inside the dielectric substrate and of the signal at the ends of the line. Indications of how to reduce the coupling between the external field and the considered structure are given     

Characterization of charge trapping in submicrometer NMOSFET's bygate capacitance measurements

Trapping of net positive charge at low gate stress voltage, and of net negative charge at high gate stress voltage, is observed through changes in the gate-to-drain capacitance of the stressed junction. These observations can be explained in terms of electron trapping, hole trapping, and generation of acceptor-like interface states located in the upper half of the bandgap. Channel shortening is also observed and found to exhibit a logarithmic time dependence     

Comparison of two models for a high temperature reactor coupled to a gas turbine

This paper presents two independent dynamic models of a nuclear gas turbine power plant. Both the high temperature nuclear reactor (HTR) and its energy conversion system (ECS) based on a direct Brayton cycle have been modelled. One model utilises RELAP5 for the ECS, the other Aspen Custom Modeler (ACM). The reactor model used in both models is a point kinetic model derived from a detailed reactor model. The ECS model is described and compared componentwise, with an emphasis on the turbomachinery. The total plant models are compared with each other by calculating two representative transients: one load rejection transient and one transient with the system at part load.     

An electrooptic intensity modulator with improved linearity

A highly linear electrooptic modulator has been designed, fabricated, and evaluated. The design of this modulator consists of only a simple modification to the directional coupler. Two-tone testing has demonstrated that, for an optical modulation depth of 30% per channel, the third-order intermodulation distortion is more than 30 dB lower than that of the conventional directional coupler or Mach-Zehnder modulators. This improvement was not observed to be accompanied by any increase in second harmonic distortion. Also included are results of two-tone computer simulations which predict the improvement in linearity of this device for a range of modulation depths.<>     

A least mean square error synthesis with pattern constraints atmultiple frequencies

The pattern synthesis techinque described allows the power pattern to be optimized in a very useful sense over the desired operating band of the antenna array. This is an improvement over previous techniques involving constraints on such artificial parameters as source norm and superdirectivity ratio, which do not, in fact, ensure a useful pattern bandwidth. The procedure described provides minimum mean square error over an arbitrary number of discrete frequencies