Design of High Precision Temperature Sensor based on Platinum Resistance

Abstract. for Pt100 platinum resistance temperature measurement system has low accuracy， duplicate hardware circuit design and other issues， the articles design a high-precision temperature measurement system with Pt100 platinum resistance. With Pt100 temperature sensor， article takes STM32F103 as the control center through filtering， amplification and other signal conditioning circuit and a method of combining software look-up table to compensate the nonlinear compensation， thus achieving high-precision temperature measurement. Research results show that the system has a measurement accuracy high stability， scalability， and other characteristics.

Key words： Pt100； high precision temperature measurement； STM32F103

1 Introductions

In the industrial process temperature measurement is not only important physical parameters， detection and control its temperature directly affect the safety， product quality， production efficiency， energy conservation and other major technical and economic indicators.

For example： water temperature is the most important parameter sea state， and which changes the distribution of the main factors directly influence the various naval equipment. Temperature measurement is already a very mature technology， temperature sensitive components both traditional thermal resistance， thermocouple， thermistor temperature sensors， there are modern integrated temperature sensor， digital temperature sensor， as well as ultra-high temperature optical temperature sensor， wherein the thermal resistance， especially Pt1000 platinum resistance temperature measurement with high precision， high sensitivity and other characteristics in the low-temperature measurements plays an important role.

The program of improved Pt1000 platinum resistance temperature measurement accuracy has three below：

（1） using a four-wire connection to compensate measurement errors. The use of platinum resistance wiring generally have three kinds： namely， two-wire connection， three-wire and four-wire connection， two-wire connection which does not take into account the lead resistance can only be applied to temperature relatively close distance occasions； three-wire connection method to eliminate some of the lead resistance； four-wire system is the thermal resistance at the root ends of the two wires for each connection， which provides a constant current I two leads as thermal resistance， the R is converted into a voltage signal U， then through another two to lead the U cited measuring end. This leads the way to completely eliminate the effects of lead resistance， but higher costs， mainly for high-precision temperature sensing.

（2） the sampled signal is filtered by reducing random error.

Due to factors outside interference or some unpredictable， subject to the interference for analog， after the results of A/D conversion deviates from the true value， there may be some random error， if only sampled once， unable to determine whether the results credible， It must be a data sequence A/D conversion through multiple sampling in order to get a high confidence results after processing through software algorithms. This method is a digital filter. A filter is a frequency signal that can be useful while containing （or significantly attenuate） the electronic apparatus unwanted frequency signals， can be divided into analog and digital filters. The analog filter is used mainly R， L， C and other passive components filter circuit or by the op amp and R， C consisting of active filter. The digital filtering algorithm is the use of filtering software. Premise digital filtering is performed multiple times on the same data sample.

（3） software algorithms to adjust Pt1000 of nonlinearity.

We know the characteristics of Pt1000 that linearity of Pt1000 is better， but because of its non-linear temperature - resistance function， using SCM calculation method is resource-intensive and time than more. Typically researchers use look-up table and algorithms scale conversion method to calculate the temperature， not only computing faster， less occupied MCU internal resources， but also to a certain extent， we adjust Pt1000 linearization to achieve very precise temperature measurement results.

2 Principles of composition and temperature

measurement system

Multi-channel temperature sensor measurement system consists of a constant current source circuit， analog electronic switch group， a differential amplifier circuit， A / D converter circuit， system block diagram shown in Figure 1.

2.1 Constant current source driver

When a group of STM32F103RBT6 strobe analog switch， constant current source produces a current of 1 mA， the role of the Pt100 platinum resistance temperature sensor and 100 Ω standard resistor （one thousandth of accuracy）， and the temperature signal into a differential signal. Pressure signal generated is small millivolt signal. Therefore， the study by the differential amplifier circuit for the small signal input 100 times magnification， and finally through 12 A / D converter converts the analog signal into a digital signal to the microprocessor for processing.

r1，r2，r3-wire resistance， when the lead length of the same， the same size wire resistance； RX-Pt100 resistance；R-100 Ω standard resistance

2.2 Design of the differential amplifier circuit

The system of operational amplifier is ISL28233， which uses chopper-stabilized structure technology to achieve a very low input offset voltage （8μV）， temperature drift voltage （0.05μV/℃） and noise voltage， quiescent current of 17μA， CMRR of 125 dB， which is a high stability， high reliability of rail-to-rail operational amplifiers. The differential amplifier circuit shown in the right part of FIG 3. Xout， Yout end Pt100 temperature within the range of 0～100℃， the maximum input pressure of 0～38.51 mV. The differential amplifier circuit may be a weak voltage signal conditioner 100 times magnification obtained after 0～3.851 V voltage from the buffer again by half a voltage follower， via a low-pass RC filter， a voltage reflecting the current temperature value， the subsequent transmission to the a / D converter for processing.

2.3 The principle of AD sampling

STM32 integrates very powerful peripheral ADC. Enhanced products in the STM32F103xC， STM32F103xD and STM32F103xE， with 3 12 bit ADC， each ADC share up to 21 external channel， can be single or multiple scan conversion. If the article is used in STM32F103VET6， which belongs to the enhanced CPU， it has 18 channels， can measure 16 external and 2 internal signal source. Each channel of the A/D transformations can perform single， continuous or discontinuous mode， scanning.

1） resolution： 12 bit resolution. Can not directly measure the negative voltage， so there is no sign bit， namely the minimum quantization unit LSB = Vref+ / 212.

2） conversion time： conversion time is programmable. Sampling time of at least 14 ADC clock cycle， and the clock frequency of ADC up to 14MHz， that is to say， the sampling time it's short for 1us. It can completed the sampling frequency of the digital oscilloscope.

4 Conclusions

In this paper， combined with the practical requirements of engineering， design development of the temperature measuring system based on platinum resistance transducers， and analyze system testing and error， reaches the design requirements of wide measurement range， high precision， reliability and stability analysis of the system design， and path consistency， so that the entire temperature detection system have the stable and reliable working.

Reference

[1] Fang Yixi，Lei Kaizhuo，Qu Jiankang.High precision temperature measurement system based on PT1000[J].Electronic design engineering， 2010， 18 （10）： 79-82.

[2] Zhang Yu， Zhang Shengwei.Design of high precision temperature measuring system based on platinum resistance transducers[J]. Journal of transducer technology， 2010， 23 （3）： 311-314.

[3] Chen Runming， Qin Hui， Han Chao.A high precision temperature detection scheme （J）.Communication of science and technology， 2011 （7）： 222-227.

[4] Li Zhibin.A new type of thermocouple， thermal resistance automatic verification system[J].Micro computer information， 2007， 5 （1）： 191-192204.

[5] Zhang canlan.A method of precision temperature measurement optimization [J].Electrical and Mechanical Engineering， 2009， 26 （9）： 57-59， 80