Design of a net-zero charge neural stimulator with feedback control

Abstract

This paper presents a high efficiency, net-zero charge neural stimulator. A new stimulation strategy is proposed to reduce the charge error that originates from the irreversible charge diffusion, which is a common issue in traditional current matching stimulator designs. In addition, an arbitrary channel configuration of the working and counter electrodes is achieved. Two methodologies are applied to the proposed design to increase the stimulation efficiency: i) feedback control of an adaptive driving voltage, which enables a constant low operating voltage for the entire active circuits; ii) charge recycling, which “recycles” the accumulated charges on the blocking capacitor. An improved current mode DAC and a digital feed-forward error compensation comparator are integrated in the output stage to suppress the process variation, and minimize the charge error in continuous stimulation pulse trains. Performance characterization and invivo experimental result of a prototype chip fabricated in standard 180nm CMOS technology are presented. An efficiency improvement of 51% is measured in the experiment.