In an industrial and consumer electronic marketplace that is increasingly demanding greater real-world interactivity in portable and distributed devices, analog to digital converter efficiency and performance is being carefully examined. The successive approximation (SAR) analog to digital converter (ADC) architecture has become popular for its high efficiency at mid-speed and resolution requirements. This is due to the one core single bit quantizer, lack of residue amplification, and large digital domain processing allowing for easy process scaling. This work examines the traditional binary capacitive SAR ADC time and statistical information and proposes new structures that optimize ADC performance. The Ternary SAR (TSAR) uses the quantizer delay information to enhance accuracy, speed and power consumption of the overall SAR while providing multi-level redundancy. The early reset merged capacitor switching SAR (EMCS) identifies lost information in the SAR subtraction and optimizes a full binary quanitzer structure for a Ternary MCS DAC. Residue Shaping is demonstrated in SAR and pipeline configurations to allow for an extra bit of signal to noise quantization ratio (SQNR) due to multi-level redundancy. The feedback initialized ternary SAR (FITSAR) is proposed which splits a TSAR into separate binary and ternary sub-ADC structures for speed and power benefits with an inter-stage encoding that not only maintains residue shaping across the binary SAR, but allows for nearly optimally minimal energy consumption for capacitive ternary DACs. Finally, the ternary SAR ideas are applied to R2R DACs to reduce power consumption. These ideas are tested both in simulation and with prototype results. / Graduation date: 2013 / Access restricted to the OSU Community at author's request from Jan. 7, 2013 - Jan. 7, 2014

La douleur neuropathique centrale est une affection fréquente dont le traitement est complexe. En raison d’un important taux de résistance aux traitements pharmacologiques, des techniques de neuromodulation ont été développées. Parmi elles, on retrouve les stimulations du cortex moteur primaire (ou gyrus précentral), invasive (i.e. stimulation électrique épidurale, eMCS) et non-invasive (i.e. stimulation magnétique transcrânienne répétitive, rTMS). Ces techniques restent limitées par différents paramètres. La rTMS a principalement été étudiée à travers des séances uniques, et son efficacité comme moyen thérapeutique au long cours reste mal connue. La eMCS souffre d’un manque de prédicteurs individuels d’efficacité suffisamment robustes pour sélectionner à bon escient les candidats à la chirurgie. Enfin, le cortex moteur primaire est une cible de découverte empirique, et d’autres cibles sont à envisager pour améliorer les résultats de ces neuromodulations corticales. Notre travail avait pour objectif l’amélioration des connaissances vis à vis de ces différentes limites. Il s’est articulé autour de 3 axes principaux :- L’étude de la rTMS en séances répétées, au long cours, comme moyen thérapeutique à part entière. - L’étude de la rTMS en séances répétées comme moyen de prédiction de la réponse antalgique à la eMCS.- Le développement de méthodes permettant la localisation fiable et reproductible du cortex pré-frontal dorsolatéral comme cible alternative de stimulation. / Central neuropathic pain is a frequent and hard to treat condition. Because of a large amount of drug-refractoriness, neuromodulation techniques have been developed. Among them, the mostly used is motor cortex stimulation, which can be both invasive (epidural motor cortex stimulation (eMCS)) and non-invasive (repetitive magnetic transcranial stimulation (rTMS)). These techniques remain limited by different problems: On one side, rTMS has been mainly studied through unique session practice and its use for pain therapy in a long-term scale remains not well understood. On the other side, eMCS suffers from a lack of predictability: A great proportion of patients present an insufficient relief, making eMCS less and less used. Finally, the motor cortex target is a chance discovery, and some other targets could be intended to improve the results. This work had the increase of knowledge about cortical stimulations as a main goal, especially about their different limitations. This work concentrated on 3 aims: - The study of chronic, repeated sessions of rTMS, used as a long-term tool for pain therapy. - The study of repeated rTMS sessions to predict eMCS.- The development of reliable tool to help to localize others cortical targets.