Development of MSP430-Based Underwater Acoustic Recorder with Multi-MCU Framework

Abstract

Sound waves are highly conductive in the ocean; therefore, they are used in various aspects of underwater applications like, environment exploration, underwater target detection and signal transmission. For all these applications, underwater acoustic signal logging is one of the key elements of the system. In laboratory setup, data recorders are often configured on a single-board computer, industrial computer or personal computer, because the peripherals are easy to acquire and versatile in functions. However, for this approach the systems are bulky and expensive. More importantly, the power is assumed to be unlimited. Therefore this approach is not feasible for long duration operation in the field. In previous research, we already develop an underwater acoustic recorder with single low-power microcontroller. It adopts TI's low-power Micro Control Unit (MCU) MSP430F169 as the core of the system. MSP430 has built-in 12-bit A/D converter with 200 KHz sampling rate and Serial Peripherals Interface (SPI). The A/D converter digitizes the acoustic signal from the hydrophone, and throughput the data to a SD memory card via SPI. For its compactness, low-cost and low-power, the system can run long duration measurement with just eight AA alkaline cells for two to three days. However, the A/D converting thread is suspended while SPI is streaming data to SD card. In this work, we present a multi-MCU, master-slave, scalable scheme to resolve this limitation of the previous design. Multiple MCUs are used in this scheme: a master MCU, M, coordinates the sampling clock between the two units (denoted as SA and SB) which are identical MCU/SD design. As the system starts, the master MCU sends out square waves as the synchronization signal. SA is set to assume the task first. The element in the array takes turn to do the sampling at the rising edge of the clock until the buffer in each MCU is full. Then SA streams data to SD cards, and the sampling task is passed on to SB. Basically, the task is rotated between SA and SB. With this arrangement, the acoustic signal is sampled continuously but stored sequentially in MCU/SD units; and clips are stored in SA and SB alternatively. For the current setup, we implement two elements in the array. With the built-in 2K RAM in one MCU, this design is capable of streaming data at 100 KHz. One major advantage of this design is that the system is scalable in nature because the array can be augmented without modifying the circuit. For every additional element installed in the array, the data sampling and streaming rate can be increase by 50 KHz until the sampling clock limit is reached for the master unit.