Linear Design Method of PT100 Temperature Sensor/Platinum Resistor Temperature Measurement Circuit

Resistance Temperature Detector - A resistance made of a material that changes its value with the rise of temperature. If it increases with the rise of temperature and the value of resistance also increases, it is called positive resistance coefficient. If it decreases with the rise of temperature, it is called negative resistance coefficient. Most resistance temperature detectors are made of metal, of which the resistance temperature detectors made of Platinum (Pt) are the most stable - acid-base resistant, non-deteriorating, fairly linear.... Most used by industry.

The PT100 temperature sensor is a resistance temperature detector made of Platinum (Pt), which belongs to the positive resistance coefficient. The relationship between the resistance and the temperature change is as follows:
R=Ro(1+ α T) where α= 0.00392, Ro 100_ (resistance at 0 C), T at Celsius
Therefore, a resistance temperature detector made of platinum, also known as PT100.

1:Vo=2.55mA × 100(1+0.00392T)=0.255+T/1000.
2: When measuring Vo, do not separate any current, otherwise the measurement will be inaccurate.

circuit analysis

Since the power supply is usually noisy after more parts are supplied, we use the Ziner bipolar as the regulator part. Due to the function of 7.2V Ziner bipolar, the voltage sum of 1K resistance and 5K variable resistance is 6.5V. Adjustment of 5K variable resistance determines the radio (collector) polar current of the crystal. We have to adjust the collector current to 2.55mA so that the measured voltage V is 0.255+T/1000 as shown by the arrow.
Subsequently, the input resistance of the non-reverse amplifier is almost infinite, and at the same time it is amplified 10 times, resulting in an operational amplifier output of 2.55+T/100.
The 6V Zina bipolar acts as a 7.2V Zina bipolar, and we use it to call out 2.55V, so the output voltage V1 of the voltage follower is also 2.55V.
Subsequently, the output of the differential amplifier is Vo=10 (V2-V1)=10 (2.55+T/100-2.55)=T/10, and the output voltage is 2.5V if the room temperature is now 25 C.

Relevant articles: Linear design method of platinum resistance temperature measuring circuit

Abstract: A non-linear correction method for platinum resistance temperature measurement based on A/D conversion principle is introduced. The principle of platinum resistance linear temperature measurement is analyzed, and the interface circuit between A/D converter 7135 and single chip 89C51 and the test data are given.

Keywords: Platinum resistance, temperature measurement circuit design, non-linear correction of analog-digital conversion, data collection

I. Introduction

Platinum resistance temperature sensor is widely used because of its wide measurement range, good reproducibility and stability.

In the precision measurement system, the circuit structure diagram of the platinum resistance temperature measurement system is shown in Fig. 1. The platinum resistance signal is usually converted into a voltage signal through a bridge circuit, then amplified and A/D converted before being sent to a microprocessor for processing. In order to correct the non-linearity of platinum resistance temperature measurement, a high non-linearity correction scheme for platinum resistance temperature measurement is designed by using the principle of double integral A/D conversion. Practice has proved that this method not only has stable performance and simple structure, but also achieves accuracy of 0.15% FS (+) 4 words in the range of 0-200 C.

2. Principles of Nonlinear Correction

1. Principle of Nonlinear A/D Conversion

Because the platinum resistance is detected by the bridge, there is a functional relationship between the output voltage UM and the measured temperature q:

Formula A and B are constant coefficients.
If a function circuit can be constructed to have the same function form as the above:

With UM = UN, it is easy to get q = t (here, "q = t" has only mathematical meaning; in fact, their dimensions are different). Thus, on the premise of UM=UN, the measurement of temperature q is transformed into the measurement of time t.

The above is the design idea of realizing sensor non-linear correction in the form of variable transformation described in this paper. The dimension of t here is time, and its measurement is achieved by a double-integral A/D conversion. The double slope integral transformation is expressed as:

(1)

In: analog input voltage during Uin-A/D conversion,

The forward integration time during the T1-A/D conversion.

The inverse integral time during the T2-A/D conversion.

Reference input voltage for Uref-A/D conversion.

When Uref is fixed, Uin has a linear relationship with T2, so in this case the A/D output can be considered as:

T2 = T1Uin / Uref.

Assume that Uref(t) is a function of time t: Uref(t)=M+Nt(2)

M, N are the undetermined constant coefficients.
If the temperature nonlinearity of the platinum resistance can be fully compensated by the A/D converted output, the following are: Uin=aq+Bq2(3)

Therefore, the formulas (2) and (3) are substituted into the formulas (1).
Assume: AT1=M, BT1=N/2,
The values of T2 and Q are equal, that is, T2=q, which implements a linear conversion between temperature and number of platinum resistors.

It can be seen that during A/D conversion, the relationship between analog voltage input and digital output is not linear, and its functional relationship is just opposite to that of Rq-q. Linear q/T2 conversion can be obtained when their characteristics are fully compensated for each other. Obviously, the key to realizing the non-linear correction using the double integral A/D transformation is to satisfy the functional relationship represented by equation (3). This scheme uses RC loop to achieve this purpose very simply.

2. High A/D Converter ICL7135

The linearization design of platinum resistance temperature measurement circuit uses 4-bit semi-double integrator A/D converter ICL7135. Each conversion cycle of ICL7135 is divided into three phases: auto-zero phase, voltage integral phase under test, and reverse integral phase of reference voltage Uref. The working process of ICL7135 is analyzed below in conjunction with platinum resistance temperature measurements:

(1) Forward Integral Stage

The schematic diagram of the interface circuit between ICL7135 and 89C52 is shown in Figure 2. At this stage, ICL7135 integrates Uin at a fixed time of T1=10000T0 (T0 is the clock cycle). The output voltage of the integrator is:
(4)

At the same time, the reference capacitor C discharges the resistor R at this stage. External resistance R is set to correct the secondary nonlinearity of temperature characteristics of platinum resistors. When this phase is completed, the voltage at both ends of C is:

(5)

In formula, when UW is t = 0, the voltage values at both ends of capacitor C.

Expand the above formula according to Marklaurin's formula at t = T1 and simplify it to:

(6)

(2) Reverse integration stage:
At this stage, the voltage at both ends of the reference capacitor C is inversely integrated by the internal integral circuit. UC(t) is considered linear throughout the T2 stage, and the integrator output returns to zero at the end of the T2 stage, when: (7)

It can be concluded from Formula (4), Formula (6), Formula (7):

Substitute Formula (3) into Formula Upper to get:

If the constants on both sides of an equation are equal, then q=T2.

In T2 time, the A/D converter is clocked and output as a digital quantity, which reflects the measured temperature value quantitatively and realizes the digital measurement of the circuit.

3. New method of interface between ICL7135 and single-chip 89C52

In the past, 7135 was used to read A/D conversion results using its BCD code output with multiple dynamic scans, which is time-consuming and takes up more ports. In the measurement and control instrument, it is very important to occupy as few microprocessor I/O ports as possible, to complete as many tasks as possible and to minimize the number of original components. The simple method of interface between ICL7135 and single-chip computer introduced here is to use the "BUSY" end of 7135, which can send the A/D conversion value of ICL7135 into single-chip computer in the interrupt service program of more than ten microseconds by occupying only one I/O port of 89C51 and one internal timer. Practice has proved that this method has practical application value.

In Fig. 2, if the 89C51 clock uses 6MHz crystal oscillation, the ALE is a stable 1 MHz frequency without executing the MOVX command. The frequency of 500 kHz obtained by dividing the ALE into two frequencies can be supplied to the ICL7135 clock input. T0 is specified as Timing Mode 1, which meets the full range requirement of ICL7135 in 1999. The state output pin BUSY of ICL7135 is high during the A/D transition phase, indicating that the A/D converter is in the signal integration and inverse integration phase, which lasts until the end of the elimination integration phase. In the timer mode register TMOD, set the gate position GATE of T0 to 1, and use BUSY as the counter gate signal, the count of T0 will be controlled by BUSY. Control counters can only be counted at high BUSY level, then input signal: A/D conversion value = counter count value -10 001 during high BUSY level period

In Figure 2, the BUSY of ICL 7135 is used to connect the external interrupt of 89C52, POL is the signal polar output, and P1.7 of 89C52. The high and low levels indicate the positive and negative polarity of the measured signal.

IV. EXPERIMENTAL RESULTS AND ERROR ANALYSIS

In the linearized design scheme of platinum resistance temperature measurement circuit, the error originates from the error caused by the non-linearity of the discharge process of the reference capacitor on the one hand: when the RC value is satisfied, the error converted into temperature value can be less than 0.03 C. On the other hand, the error comes from the A/D conversion accuracy. When the 4-bit half A/D converter ICL7135 is selected, its accuracy is (+) 0.05%, the folding temperature error is 0.10 C, the two errors are relatively independent, and the overall temperature measurement error of the circuit is (+) 0.104 C. After the circuit is assembled, the actual performance test is carried out. The experimental data are shown in Table 1. From the test results, the error of the prototype is -0.18 C, which is similar to the analysis conclusion.

Table 1 (Platinum Resistance Scale Pt100) Standard temperature ((?)
Display temperature (C)
Error (C)

100

110

120

130

140

150

160

170

180

190

200
100.20

110.17

120.05

130.12

140.11

149.95

159.88

169.84

179.84

189.82

200.18
0.20

0.17

0.05.

0.12

0.11

-0.05

-0.12

-0.16

-0.16

-0.18

0.18

Reference

[1].PT100datasheethttp://www.dzsc.com/datasheet/PT100_1196170.html.
[2].89C51datasheethttp://www.dzsc.com/datasheet/89C51_105386.html.
[3].ICL7135 Datasheethttp://www.dzsc.com/datasheet/ICL7135 +_ 225194.html.
[4].89C52 Datasheethttp://www.dzsc.com/datasheet/89C52_105388.html.
[5].BCDdatasheethttp://www.dzsc.com/datasheet/BCD_1225719.html.
[6] Co-authored by R.E. Bedford, T.M. Doffery, H. Preston Thomas: Yuan Guangfu Translation, Temperature Measurement, Metrology Press, 1995
[7] Zhao Xuezheng, Detection and Sensing Technology, Harbin University of Technology Press, 1998.10
[8] Zheng Jianguo, a high platinum resistance temperature measurement scheme, automatic instrument, 1997.18(8)