A Scalable Low-Power Time-Encoded Interface for Unbiased Capacitive Microphones

This paper presents a novel interface for MEMS capacitive microphones that eliminates the need for typical DC biasing (i.e., charge pump and high-ohmic resistance) and power-hungry analog blocks (e.g., preamplifiers, opamps, etc.). The proposed interface utilizes a switched-capacitor network (SC) to charge and discharge the MEMS sensor, with an average current that depends on the sensing capacitance and switching frequency. The architecture operates in a closed-loop configuration, ensuring that the average current flowing through the SC network is constant and is fixed by a current reference. As a result, any variation in the sensing capacitance produces a corresponding variation in the switching frequency, which is generated by a ring oscillator within the loop. The instantaneous oscillation frequency is then processed by digital electronics, performing a frequency-to-digital conversion, and a highly scalable ring oscillator-based ΣΔ analog-to-digital converter (ADC) is obtained. A proof-of-concept implementation of the novel ADC was fabricated in 130nm CMOS technology, with the analog part occupying an area of 0.02mm 2 . The measurement results demonstrate that the ADC achieves a signal-to-noise-and-distortion ratio (SNDR) of 31 dB in the voice bandwidth [300 Hz, 6.8 kHz] at a sound pressure level of 94 dB SPL, with a total power consumption of 198 μW. Although the SNDR measurement is penalized by the use of a conventional MEMS sensor designed for biased operation and the lack of back chamber optimization between the sensor and the lid, the novel architecture demonstrates the feasibility of the concept and opens the door to a new class of MEMS sensors optimized for low-bias voltage operation. The proposed ADC is suitable for use in wearable electronics audio applications, where a small chip area and low power consumption are key constraints.