Energy Efficient Circuit Design Techniques and Beyond CMOS Exploration for Internet of Things (IoT)

Location: FXB 1012

The Internet-of-Things (IoT) has long been viewed as the next major computing class that has significant societal benefits on people’s lives. Sensor nodes have become ubiquitous in a broader range of fields such as healthcare, smart infrastructures, agriculture, environmental data collection, and industrial monitoring. With the rapid shrinkage and added intelligence of these systems, they are becoming increasingly power hungry. This thesis presentation introduces new circuit design techniques supporting ultra-low power operation with innovative power management solutions and exploration of energy efficient circuit architectures for beyond CMOS technologies.

The first work presents the sample and average common mode feedback technique for capacitively coupled amplifiers. It provides a method to implement a resistance in the order of 100 TΩ reliably across process corner and temperature variations enabling lower noise and hence lower power. The second work delves into circuit design using the Magneto-Electric Spin Orbital (MESO) technology which has the potential to be an alternative to CMOS in the future. This work introduces new architectures for logic, memory and in-memory computation using the MESO technology targeted to improve energy efficiency and latency. Finally, the third work proposes a nano-ampere load current switched capacitive high voltage (up to 200V) DC-DC up-down converter using a single small off-chip capacitor. This unique topology enables both up and down conversion of voltage with the same switched capacitive structure.

Chair: Professor David T. Blaauw