Charge-Domain Analog/Mixed-Signal Circuits and Applications

Abstract

As IoT is increasingly integrated into our everyday life through applications such as health monitoring, surveillance, activity detection, the demand for resource efficient circuit design is rising. This trend is driving circuit research in highly resource aware designs and calls for re-examination of basic building blocks to increase their intrinsic efficiency. Charge-domain circuit techniques are attractive for their inherent focus on charge as an energy-centric mode of operation, which has the potential to achieve a level of efficiency that is higher than that obtained by using conventional current or voltage mode of operation. Also, it provides interesting features that cannot be achieved with a conventional domain of design. However, to date, there has been only few studies that exploits their unique resource saving qualities. In this thesis, I introduce novel charge-domain techniques and discuss their properties that enable highly efficient design. The enhanced efficiency from charge-domain techniques is demonstrated through applications to analog/mixed-signal circuits, which include a motion-triggered image sensor that achieves best-in-class ADC FoM, and a charge-domain pre-amplifier that achieves the state-of-the-art discrete-time equivalent NEF.