Design of seawater temperature and depth measurement system based on AD7791

    In the process of developing and utilizing the ocean, temperature and depth play a very important role, which makes it the key observation object in ocean observation. Changes in ocean temperature and depth caused by long-term earth changes or other geographical factors are related to the development and utilization of offshore resources [1]. Therefore, this paper designs a large-capacity, high-precision and low-power temperature and depth measurement system to meet the measurement needs. The system can collect the temperature and pressure of environmental elements in different regions and water depths, and then obtain the ocean depth by analyzing the pressure and a series of calculations.
1 System Overall Design
    This system uses Atmel's ATmega64 microcontroller as the control chip, which is mainly divided into A/D data acquisition part, clock timing control part and data storage part. The overall structure of the system is shown in Figure 1. ATmega64 is a low-power 8-bit CMOS microcontroller based on the enhanced AVR RISC structure. Due to its advanced instruction set and single clock cycle instruction execution time, the data throughput of ATmega64 is as high as 1 MIPS/MHz, which can alleviate the contradiction between power consumption and processing speed of the system.

    The system is powered by a CR123A 3 V lithium battery and the crystal oscillator is 3.686 4 MHz, which makes the system power consumption very low. The 24-bit high-precision, low-power AD7791 analog/digital converter is used to convert the collected temperature and pressure analog quantities into digital quantities. The PCF8583 with an I2C bus interface and extremely low power consumption is used as the clock control chip. The sampling time, sampling interval, sampling frequency, etc. are written into the PCF8583 to control the microcontroller to work or sleep, which greatly reduces the power consumption. Since the measurement acquisition time is long and the amount of collected data is large, a large-capacity NAND memory with non-volatility and fast read and write speed is used [2].
2 System Circuit Analysis
    AD7791 is a high-resolution 24-bit analog/digital converter device launched by ADI in the United States. It has high precision and low power consumption. Its flexible serial interface allows AD7791 to be easily connected to a microprocessor or shift register. Since AD7791 uses the sum-difference (∑-Δ) conversion technology, it is not affected by the noise environment and is suitable for the measurement of wide dynamic range and low-frequency signals. Therefore, the A/D of the temperature acquisition circuit and the pressure acquisition circuit are both AD7791. The temperature acquisition circuit and the pressure acquisition circuit both adopt the idea of ​​the ratio method, which avoids the drift caused by the unstable power supply voltage and greatly improves the measurement accuracy of temperature and pressure.
2.1 Temperature acquisition circuit
    The temperature acquisition circuit is shown in Figure 2. U1 plays the role of a voltage follower in the temperature acquisition circuit, Rt uses an NTC thermistor, and the resistors R1 and Rt and U1 and AD7791 are powered by equal voltage. As shown in Figure 2, Vo=VCC×Rt/(R1+Rt), AD7791 input voltage AIN=Vo, and reference voltage VREF=VCC.

    The output code value is: Code = 2N × Rt (AIN / VREF), the temperature AD value calculated by the resistance ratio method is Code = 2N × Rt / (R1 / Rt), which is independent of the circuit power supply voltage, thereby eliminating the drift caused by unstable power supply voltage and improving the accuracy of the temperature measurement circuit. If the analog ground and the digital ground are directly connected over a large area, mutual interference will occur. R2 is 0? The redundant resistor is equivalent to a very narrow current path, which can effectively limit the loop current, suppress the noise, and improve the stability of the circuit [3].
    The temperature measurement circuit divides the voltage of the thermistor by the fixed resistor R1 and the thermistor Rt; then, through the voltage follower, the voltage of the thermistor collected by the AD is sent to the microcontroller for data analysis.
2.2 Pressure acquisition circuit
    The pressure sensor uses the Keller pressure sensor, which is a piezoresistive OEM pressure sensor. When pressure is generated on the pressure sensor, the resistor will deform, the resistance of the two relative resistors will increase, and the resistance of the other two relative resistors will decrease, and the voltage across the pressure sensor remains unchanged.
    Only when a pressure difference is generated in the middle of the bridge, the output voltage of the bridge will change, and the resulting voltage change is used as the output to reflect the pressure. Since the change voltage of the bridge output is very small, the pressure change value is amplified by the operational amplifier U6 and sent to AD7791, and then the obtained AD value is sent to the single-chip microcomputer. The pressure acquisition circuit is shown in Figure 3.

    R8 and U5 form a constant current source, and the current flowing through R8 is I=VCC/R8. Since the current flowing through the two branches of the bridge is equal, V32=V3-V2=0.5I×(R9-R10). U6 plays the role of differential amplifier, and its output voltage Vout=V32×G, G is the amplifier gain. After the signal is amplified, it is sent to AD7791 for analog/digital conversion. The output code value is: Code=2N×(AIN/VREF), the reference voltage VREF=VCC, and the AD7791 input voltage AIN=Vout.
    The calculation results in Code=2N-1×(R9-R10)×G/R8, that is, the AD value finally obtained by the ratio method is independent of the voltage, thereby eliminating the temperature drift caused by voltage changes, improving the stability of the system and the accuracy of pressure acquisition data.
2.3 Clock control circuit
    PCF8583 is a clock control chip with 256 B of static RAM, and uses the I2C bus interface to transmit addresses and data. PCF8583 has an internal 32.768 kHz crystal oscillator, and the first 8 B are used for clock, calendar and counting functions; the following 8 B can be used as an alarm register; the remaining 240 B is a free RAM area, which can be used to store data and other flags or constants such as sampling intervals.
    Y1 provides a stable 32.768 kHz crystal oscillator for PCF8583, and the circuit is shown in Figure 4. The microcontroller sets sampling parameters such as sampling time and sampling interval for PCF8583 through the I2C bus. When the PCF8583 timer count overflows, a timer overflow interrupt is generated to control the microcontroller to switch between the three states of sampling, standby and sleep, reducing power consumption.

2.4 Data storage
    Since the system is used to measure harsh environments such as the ocean, it is difficult to put it into service and recover it due to objective factors. In addition, in order to accurately observe the temperature and depth information, a large amount of data must be collected over a long period of time. Therefore, a large data storage capacity is required to meet the measurement needs. In view of this, this system uses Samsung's K9F1G08, which is a 128M×8 bit NAND Flash. K9F1G08 consists of 1 046 blocks, each block contains 64 pages, and one page is 2 KB+64 B=2 048 B+64 B=2 112 B. K9F1G08 uses pages as the basic unit for storage and blocks as the basic unit for erasure. It has a very fast writing and erasing speed and is a better storage device than a hard disk drive.
    The microcontroller first puts the collected data into its own RAM, and then stores it into K9F1G08 when the data reaches one page, thereby improving time utilization and reducing system power consumption.
3 System Software Design
    The system software mainly completes the functions of data sampling, data storage, external interrupt and timing interrupt control. The main program is controlled by interrupts to work in three states: sampling, sleep and standby, which greatly reduces the system power consumption. The program flow chart is shown in Figure 5.
     

    In October 2012, the measurement system was put into Qingdao Zhongyuan Wharf for calibration. The measurement system was put into the place about 8 m below the horizontal surface of the wharf for sampling. 24 hours after the deployment, the measurement system was taken out and connected to the computer. The collected A/D data was read out from the large-capacity memory NAND through the corresponding background software. The temperature, pressure and depth were calculated and plotted into a graph through the above calculation method. The temperature is shown in Figure 6, the pressure is shown in Figure 7, and the depth is shown in Figure 8.

    By analyzing the collected data, the temperature accuracy can reach ±0.002℃ (ITS-90 standard), and the depth sensor accuracy is 0.05% of the full scale, meeting the measurement requirements of high precision and low power consumption.
    This paper introduces the measurement method and working principle of temperature and depth. The whole measurement system has reliable performance, large storage capacity, low power consumption and high accuracy, so it has a very broad application prospect in the measurement of ocean temperature and depth.
References
[1] Bi Yongliang, Sun Yi, Huang Motao, et al. Research progress and prospects of ocean measurement technology [J]. Marine Surveying and Mapping, 2004, 24(3):65-7.
[2] Wang Shengan, Long Xiaomin, Chen Junchang, et al. Application of large-capacity flash memory in pressure wave and tide gauges [J]. Instrument Technology and Sensors, 2000(10):31-32
.
[3] Kang Huaguang. Basic simulation of electronic technology [M]. Beijing: Higher Education Press, 1999.
[4] FOFONOFF N, MILLARD R. Algorithms for computation of fundamental properties of seawater [J]. UNESCO Technical Papers in Marine Science, 1983(44):1-53.