A VCO-Based CTDSM With Integrated Phase Error Correction for Neural Interface

Neural recording front-ends with voltage controlled-oscillator (VCO) based continuous-time
$\rm \Delta \Sigma $
modulators (CTDSM) enable direct analogue to digital conversion with minimized aliasing and hardware overhead, however, their performance suffers from phase glitches due to sampling meta-stability and large interference from input or power line. This brief presents a self-correction scheme for multi-phase VCO-based CTDSM. The VCO’s phase pattern is processed in both continuous-time (CT) and discrete-time (DT) domain with real-time error detection and correction to guarantee robust operation. Based on the this scheme, a phase counter can be implemented to assist the XOR-based phase detector (PD) with minimized error, extending the phase detection range, therefore leading to improved quantization resolution and system linearity even with glitches. Fabricated in a 180-nm CMOS process, the prototype achieves 70.7 dB SNDR, 71.0 dB SNR, 83.1 dB SFDR and 71.2 dB DR within 10 kHz bandwidth. The measured input referred noise (IRN) is 5.94
$\mu \text{V}\rm _{rms}$
over 0.5 Hz-10 kHz bandwidth. The performance of this self-correction scheme is verified by injecting phase error with different frequency. A maximum of 16.2 dB improvement of SNDR can be observed for 625 Hz injected interference.