Design and Implementation of a CMOS-Based High-Speed Data Acquisition System for Industrial Sensors

Abstract

The paper presents the structure, construction and experimental analysis of a CMOS based high-speed data acquisition system (DAS) that is specifically designed towards industrial applications in the form of sensors that need high signal accuracy and speed. It combines a rigorously designed low-noise front-end amplifier, a 14-bit successive approximation register (SAR) analog-to-digital converter (ADC) and a high-throughput general purpose digital interface all built on a well-proven 180 nm CMOS process. The high resolution, fast speed and power consumption represented by the industrial environments are balanced in the proposed DAS which operates at a sampling rate of 50 megasamples per second (MS/s). The front-end amplifier features chopper stabilized differential topology with active common-mode feedback that dramatically mitigates flicker noise and makes it more immune to industrial sources of electromagnetic interference (EMI). The SAR ADC is optimized with a bootstrapped sampling switch structure and split-array capacitive digital-to-analog with the aim to maximize linearity and reduce mismatch error, and an on chip reference buffer helps keep voltage references stable during high speed conversion. The average power consumption is less than 50 milliwatts, using power-saving mechanisms, like dynamic bias scaling, fine-grain clock gating. This feature allows the system to be deployed in energy-constrained or battery powered industrial applications. The measured performance parameters of the fabricated prototype show that it has good performance as it boasts of an effective number of bits of (ENOB) 13.8, a signal-to-noise-ratio (SNR) of 84 dB, and minimal and differential such as INL and DTL of less than or equal to 1.2 LSB and 0.6 LSB, respectively. Extensive verification of the ability to effectively monitor in real-time using a wide range of industrial sensors including vibration, pressure, and temperature transducers demonstrates (in drug manufacturing, predictive maintenance, structural health monitoring, and process automation applications, to name but a few) the commodity of the overall system. They show a 25 percent increase on energy efficiency of similar DAS architecture through benchmark comparison. Future work planned is aimed at integrating on-chip calibration procedures and designing the architecture to support multi-channel applications so as to address the increased applications of scalable industrial sensing networks.