Dartmouth Events

Optional ZOOM LINK
Meeting ID: 963 0043 6329
Passcode: 314133

Abstract: Applications including micro-robotics and haptics have motivated extensive exploration of dielectric actuators due to their high bandwidth and high efficiency at cm- and mm-scale. To maximize their benefits, dielectric actuators—including piezoelectric and electrostatic actuators—typically require high driving voltages, ranging from 100 V to several kV. Generating such high driving voltages presents a challenge for drive circuits, particularly when powered by a small battery. Dielectric actuators are predominantly capacitive in most circumstances, necessitating the drive circuits to deliver and recover reactive energy for efficient operation. In some applications, such as micro-robotics, the drive circuits may need to have a weight significantly less than 1 g and a volume less than 1 cm³.

This dissertation explores topologies and integration strategies for drive circuits for small-scale, high-voltage dielectric actuators. Specifically, we outline a pseudo-soft switching series-parallel switched-capacitor (SC) converter that reduces hard-switching loss by stepping the load voltage with small increments and recovers the energy stored in the load to high-energy-density flying capacitors. The operation and performance benefits of the circuit are validated with several integrated circuit prototypes. A first prototype uses on-chip photovoltaic cells as a power source and achieves a drive voltage over 100 V with up to 14x reduction in power compared to a conventional hard switching driver. A second prototype uses a decentralized daisy-chain control scheme and chip-chip stacking to extend drive voltages to the kilovolt range, beyond the buried oxide limit of conventional high-voltage SOI CMOS processes. A final prototype extends the pseudo-soft switching concept to gate drivers for silicon and GaN power semiconductor devices. This is used to show that the pseudo-soft switching concept can be extended to MHz switching regimes with sub-ns waveform tuning, while reducing gate drive power by 5x to 7x compared to conventional gate drivers.

Thesis Committee: Prof. Jason T. Stauth (Chair), Prof. Charles R. Sullivan, Prof. John X.J. Zhang, Prof. Juan Rivas-Davila (Stanford University)