A High Performance LIA-Based Interface for Battery Powered Sensing Devices

Abstract

This paper proposes a battery-compatible electronic interface based on a general purpose lock-in amplifier (LIA) capable of recovering input signals up to the MHz range. The core is a novel ASIC fabricated in 1.8 V 0.18 µm CMOS technology, which contains a dual-phase analog lock-in amplifier consisting of carefully designed building blocks to allow configurability over a wide frequency range while maintaining low power consumption. It operates using square input signals. Hence, for battery-operated microcontrolled systems, where square reference and exciting signals can be generated by the embedded microcontroller, the system benefits from intrinsic advantages such as simplicity, versatility and reduction in power and size. Experimental results confirm the signal recovery capability with signal-to-noise power ratios down to -39 dB with relative errors below 0.07% up to 1 MHz. Furthermore, the system has been successfully tested measuring the response of a microcantilever-based resonant sensor, achieving similar results with better power-bandwidth trade-off compared to other LIAs based on commercial off-the-shelf (COTS) components and commercial LIA equipment.

Keywords: CMOS analog integrated circuits; gas sensing; low-voltage low power; portable lock-in amplifiers.

Figure 1

Electronic interface for portable sensing applications.

Figure 2

Block diagram of the signal conditioning stage.

Figure 3

LIA scheme, detailing the instrumentation amplifier architecture: a precision differential architecture design with input buffering.

Figure 4

Schematic of the designed 1.8 V class AB rail-to-rail two stage OpAmp. The auxiliary control circuit for rail-to-rail operation is depicted in grey.

Figure 5

Integrated lock-in amplifier (detail) for the prototype, with an active area of 306 × 114 μm². It includes the switches and instrumentation amplifiers (IA).

Figure 6

Photograph of the proposed portable electronic interface, with dimensions 85 × 99 mm².

Figure 7

LIA signals for an input signal of 150 mV_rms at an operating frequency of 80 kHz. Square quadrature signals V_r, (green), V_r90 (blue), sensor signal OUTPUT (purple) and mixers signals V_{mixer_x} (black) and V_{mixer_x} (light blue).

Figure 8

LIA experimental recovered input amplitude vs. input signal with different amplitude values for a 500 kHz square input.

Figure 9

Recovery error vs. input amplitude (square signal) at different frequencies (G = 1). Grey area corresponds to the area between the lower and upper limits of the recovery errors.

Figure 10

LIA experimental recovered input amplitude vs. input signal with amplitude values from (a) 1 μV to 200 μV and (b) 1 μV to 20 μV for a 10 kHz square input using a system gain of G = 100.

Figure 11

Relative error vs. SNR for a 10 mV_rms square input signal buried in white noise. Grey area corresponds to the area between the lower and upper limits of the recovery errors.

Figure 12

Amplitude recovery for the test signal of 10 mV_{rms_no} at an operating range of 10 kHz (a–c) and 100 kHz (d–f) in 5 Hz steps buried in −20 dB of sinusoidal interference whose frequency is around the main frequency components of the square data signals.

Figure 13

Amplitude recovery for the test signal of 10 mV_rms at the maximum operating frequency limit of 1 MHz buried in −20 dB of sinusoidal interference.

Figure 14

Normalized amplitude (a) and phase (b) values obtained for a frequency sweep around fundamental frequency of a microcantilever using three different recovery devices using a frequency step of 1 Hz.