Application of TMS320VC5402 in Acceleration Wave Sensor

1 Introduction
Wave observation is an important part of ocean survey. Using a wave buoy equipped with an acceleration wave sensor is an effective way to measure waves. When the buoy equipped with a wave sensor moves with the wave surface, the sensor in the buoy outputs a signal reflecting the change of the acceleration of the wave surface. After a secondary integration process, a signal proportional to the change of the wave surface height can be obtained. This signal is then processed to obtain the wave height and wave period data. The acceleration signal integration uses an analog integration circuit or a numerical integration method. Usually the wave period is 2 to 30 s. The analog integration circuit uses a large integral capacitor value, which makes the sensor volume larger. In addition, the analog circuit is easily affected by external factors such as temperature and humidity, which is not easy to debug. The use of numerical integration can effectively overcome these problems.
Numerical integration requires a large number of multiplication and addition operations. DSP is a microprocessor suitable for digital signal processing operations, which can be used to implement various real-time and fast digital signal processing algorithms. The TMS320C54x series DSP is a 16-bit fixed-point digital signal processor designed by TI to achieve low-power, high-speed real-time signal processing. It has high operational flexibility and running speed and is suitable for embedded applications. Therefore, this design selects TMS30VC5402 DSP as the data processor.

2 System hardware circuit design
Figure 1 is a block diagram of the system composition of the acceleration wave sensor, which includes the acceleration sensor, anti-aliasing filter, A/D converter, digital signal processor, communication interface, power supply system and other parts.

The accelerometer is the core component of the wave sensor. Here, the MMA1260EG accelerometer produced by Freescale Semiconductor is selected. This device is a low-cost, small-size, silicon capacitive micromechanical accelerometer that uses signal conditioning, temperature compensation, and self-test technologies. The device has been processed with zero-g compensation and bipolar low-pass filtering, thereby simplifying the external circuit design. The operating voltage of the MMA1260EG is 5 V, the measurement range is ±1.5 g on the Z axis, and the sensitivity is 1,200 mV/g. Figure 2 shows the application circuit of the MMA1260EG.

The conventional wave period is in the range of 2 to 30 s. A low-pass filter is connected between the A/D converter acquisition as an anti-aliasing filter to remove high-frequency signal interference. The A/D converter uses TI's TLV2544 . TLV2544 is a high-performance, low-power, high-speed, 12-bit 4-channel serial CMOS A/D converter that operates on a single power supply with a voltage range of 2.7 to 5.5 V. The device can provide users with a serial port with 3 input terminals and 1 three-state output terminal, providing a convenient 4-wire interface for the microprocessor SPI serial port.

The digital signal processor TMS320VC5402 provides a high-speed, bidirectional, multi-channel buffered serial port McBSP, which can be directly connected to a serial A/D converter. Each BSP port works in SPI mode and I/O mode. In SPI mode, the BSP port is easy to connect to a serial device that follows the SPITM protocol. When the TMS320VC5402 interfaces with the TLV2544 , the device acts as an SPI master device to provide the TLV2544 with serial clock, command and chip select signals, achieving seamless connection without the need for additional logic circuits. The connection circuit is shown in Figure 3.

TMS320VC5402 is a 16-bit fixed-point digital signal processor (DSP) with a high cost-effectiveness produced by TI. Its operating speed can reach 100 MI/s. The internal resource configuration greatly facilitates users to construct systems. TMS320VC5402 is equipped with 4 K×16-bit on-chip shielded ROM (F000h~FFFFh) and 16 K×16-bit dual-access RAM (DARAM), of which the 4 K ROM contains the Bootloader program. When users design by themselves, if the program capacity does not exceed 16 K, they can use the internal resources of the device. The boot loading method is used to reduce the difficulty and cost of system design and speed up the design process. The basic hardware circuits of DSP include power supply circuit, reset circuit, clock circuit, etc. The power supply circuit is powered by dual power supplies, the core power supply CVDD uses 1.8 V, and the I/O power supply DVDD uses 3.3 V. The power supply circuit is implemented by TPS73HD318, as shown in Figure 4.

Figure 5 shows the reset circuit implemented by MAX706R. The clock circuit uses the internal oscillator of TMS320VC5402, and a crystal is connected between its X1 and X2/CLKIN pins to start the internal oscillator.

The communication interface is extended through the SPI bus, and the MAX3100 of Maxim is selected. The MAX3100 has a built-in simple UART, a baud rate generator with an SPI interface, and an interrupt generator. The baud rate, word length, checksum, 8-byte receive FIFO, and general UART or Ir-DA are set through the "write structure register", and the shutdown state and 4 interrupt tasks are controlled. Figure 6 shows the UART circuit, and the MAX3221 in the figure is a level converter.

3 System software design
The system software design adopts the MATLAB-DSP system-level integrated environment, that is, the conceptual design, simulation/emulation, target code generation, operation and debugging are completed in the unified MATLAB environment. The use of the MATLAB-DSP system-level development environment greatly saves the time spent on programming and error correction, and speeds up the design process. The MATLAB-DSP integrated development environment completely changes the previous DSP design method. In this environment, operations on the target DSP can be completed, including accessing the DSP's memory and registers, etc., using MATLAB's powerful tools to analyze and visualize the data in the DSP memory, and directly generating DSP executable target code from the MATLAB program.

The acceleration data acquired through A/D acquisition is first transformed into frequency domain data through fast Fourier transform, and then filtered for 2 to 30 seconds after secondary integration in the frequency domain. Then, the data is inversely transformed through fast Fourier transform to obtain time domain data again. After scaling, the data is output through the serial port. The processing flow is shown in Figure 7.

Frequency domain integration is a very useful processing method. The numerical calculation formula of the frequency domain quadratic integral is:

Where, are the lower and upper cutoff frequencies, respectively; X(k) is the Fourier transform of x(r); △f is the frequency resolution.

4 Test results
The laboratory used a wave simulation calibration device to calibrate the accelerometer wave sensor with TMS320VC5402 as the processor. The calibrated sensor has a wave height measurement range of 0-20 m, a measurement error of ±(0.3+5%×measured value) m, and a wave period measurement range of 2-20 s, with a measurement error of ±0.5 s, which meets the requirements of the wave buoy industry standard. The accelerometer wave sensor with TMS320VC5402 as the processor was compared with the wave sensor using an analog integrator. Figure 8 shows a set of data obtained during the sea test of the wave buoy equipped with an analog integration wave sensor and a numerical integration wave sensor. From the waveform, the accelerometer using numerical integration (solid line) obtains data that is more consistent with the original sensor using an analog integrator (dashed line). Laboratory and field tests show that the design of the accelerometer wave sensor based on the frequency domain integration algorithm implemented by TMS320VC5402 is feasible.

5 Conclusions
This accelerometer wave sensor based on frequency domain numerical integration is simple to debug, has high stability and small size. It has been applied to wave buoys to replace the previous analog integral wave sensor to measure the wave height and wave period.