This thesis describes the outcome of study to investigate methods of broadband matching to photonic devices such as lasers and high speed detectors. The thesis is divided into two areas of interest relating to the design of broadband fiber optic links. The first area is the application of numerical methods and commensurate line methods to the design of compact equalisers which allow an improved transducer power gain over a wide band. It is shown that physically small equalisers can yield an improvement of 4 dB over a 2 GHz bandwidth. The second area considered is the distortion inherent in a laser diode. Detailed measurements of the second order and intermodulation products are given. A small signal perturbation analysis is presented which helps to explain the observed distortion products. The results of numerical simulation of the distortion using a state variable implementation of the full rate equations and related first, second and intermodulation equations is presented and possible methods of reducing the distortion are explored. It is shown that in principle the distortion could be reduced by pre-generating the distortion and adding an inverted form of the distortion to the transmitted signal. The distortion can then be corrected in the fiber and simulation studies suggest that an improvement of 13 dB optical or 26 dB electrical may be possible.

This letter presents the design and characterization results of a multi-octave power amplifier fabricated in a 0.13μm SiGe-BiCMOS technology. The single stage power amplifier is implemented as the stack of a cascode amplifier combining broadband input matching network with resistive feedback, and a common-base amplifier with base capacitive feedback. Measurement results show that the design delivers a peak saturated output power level of 20.2 dBm, with output 1 dB compression at 19.4 dBm. The measured 3 dB power bandwidth is from 4 GHz to 32 GHz, covering three octaves. The corresponding power fractional bandwidth is 156 %. The measured peak power added efficiency is 20.6 %, and peak small signal gain is 18.6 dB. The fabricated integrated circuit occupies an area of 0.71mm2. To compare state-of-the-art multi-octave power amplifiers, the power amplifier figure of merit defined by the international technology roadmap for semiconductors is modified to include power fractional bandwidth and area. To the knowledge of the authors, the presented design achieves the highest figure of merit among multi-octave power amplifiers in a silicon based integrated circuit technology reported in literature.

Le problème de l'adaptation d'impédance en ingénierie des hyper fréquences et en électronique en général consiste à minimiser la réflexion de la puissance qui doit être transmise, par un générateur, à une charge donnée dans une bande de fréquence. Les exigences d'adaptation et de filtrage dans les systèmes de communication classiques sont généralement satisfaites en utilisant un circuit d'adaptation suivi d'un filtre. Nous proposons ici de concevoir des filtres d'adaptation qui intègrent à la fois les exigences d'adaptation et de filtrage dans un seul appareil et augmentent ainsi l'efficacité globale et la compacité du système. Dans ce travail, le problème d'adaptation est formulé en introduisant un problème d'optimisation convexe dans le cadre établi par la théorie de d'adaptation de Fano et Youla. De ce contexte, au moyen de techniques modernes de programmation semi-définies non linéaires, un problème convexe, et donc avec une optimalité garantie, est obtenu. Enfin, pour démontrer les avantages fournis par la théorie développée au-delà de la synthèse de filtres avec des charges complexes variables en fréquence, nous examinons deux applications pratiques récurrentes dans la conception de ce type de dispositifs. Ces applications correspondent, d'une part, à l'adaptation d'un réseau d'antennes dans le but de maximiser l'efficacité du rayonnement, et, d'autre part, à la synthèse de multiplexeurs où chacun des filtres de canal est adapté au reste du dispositif, notamment les filtres correspondant aux autres canaux. / The problem of impedance matching in electronics and particularly in RF engineering consists on minimising the reflection of the power that is to be transmitted, by a generator, to a given load within a frequency band. The matching and filtering requirements in classical communication systems are usually satisfied by using a matching circuit followed by a filter. We propose here to design matching filters that integrate both, matching and filtering requirements, in a single device and thereby increase the overall efficiency and compactness of the system. In this work, the matching problem is formulated by introducing convex optimisation on the framework established by the matching theory of Fano and Youla. As a result, by means of modern non-linear semi-definite programming techniques, a convex problem, and therefore with guaranteed optimality, is achieved. Finally, to demonstrate the advantages provided by the developed theory beyond the synthesis of filters with frequency varying loads, we consider two practical applications which are recurrent in the design of communication devices. These applications are, on the one hand, the matching of an array of antennas with the objective of maximizing the radiation efficiency, and on the other hand the synthesis of multiplexers where each of the channel filters is matched to the rest of the device, including the filters corresponding to the other channels.