SUB-THRESHOLD CMOS SENSOR INTERFACE FOR ULTRA–LOW-ENERGY WEARABLE HEALTH MONITORS

Wearable health monitors have required continuous physiological signal acquisition under severe energy constraints. Conventional sensor interfaces have consumed significant power due to super threshold biasing and frequent data conversion. Sub-threshold CMOS operation has emerged as a viable alternative for achieving ultra–low energy sensing while maintaining acceptable signal fidelity. Prior studies have highlighted feasibility, yet practical interface stability and noise robustness have remained limited. Existing wearable sensor interfaces have suffered from excessive energy consumption, reduced battery lifetime, and poor performance under low-voltage operation. Process variations, thermal noise, and leakage effects have degraded accuracy when circuits have operated in the sub-threshold region. These challenges have restricted long-term deployment in body-worn healthcare systems. This work has proposed a sub-threshold CMOS sensor interface that has targeted ultra–low-energy wearable applications. The interface architecture has included a low-noise transconductance amplifier, an adaptive bias generator that has regulated operating points, and an energy-efficient analog-to-digital conversion stage. A calibration mechanism that has compensated process variations has been integrated. Circuit blocks that which operate below threshold voltage have been optimized for robustness. The design has been validated using standard CMOS technology through post-layout simulations and measured prototypes. The proposed sub-threshold CMOS sensor interface demonstrates an average power consumption of 41 nW at a 0.4 V supply after 200 iterations, which represents a reduction of approximately 53% compared with event-driven sub-threshold designs. Input-referred noise remains below 4.6 µVrms, while the signal-to-noise ratio exceeds 58 dB. The energy per conversion reaches 16 pJ, and gain drift is limited to 3.0% under process and temperature variations. These results confirm that the interface supports stable and ultra–low-energy wearable health monitoring.