Temperature detection is widely used in our daily lives and industrial settings, making the accuracy of temperature measurement circuits increasingly important. How can we improve the accuracy of temperature measurement circuits? This article will take a resistance temperature detector (RTD) as an example, starting with the selection parameters of RTDs, to briefly explain the directions for improving temperature measurement accuracy.
Platinum resistance thermometers (RTDs) possess excellent long-term stability and accuracy, making them commonly used industrial temperature sensing elements. In recent years, thin-film printing processes have reduced the amount of precious metal platinum used, significantly lowering the cost of PTDs and leading to their wider application. When using PTDs in conjunction with subsequent circuits, focusing on three basic parameters—nominal resistance, temperature coefficient, and accuracy class—helps determine the appropriate PTD selection. Understanding temperature-resistance conversion characteristics, measuring current, and wiring methods helps minimize the introduction of additional circuit errors and build accurate temperature measurement circuits.
1. Nominal resistance
The nominal resistance is the resistance value of a platinum resistance thermometer at its freezing point of 0°C. The most common nominal resistance is PT100 with a nominal resistance of 100Ω, but there are also PT200, PT500, and PT1000 with nominal resistances of 200Ω, 500Ω, and 1000Ω.
2. Temperature coefficient
The temperature coefficient (TCR) is the average change in resistance per unit temperature of a platinum resistance thermometer between the freezing and boiling points of water. Different organizations use different temperature coefficients as their standards. The European IEC 60751 and Chinese GB/T 30121 use a temperature coefficient of 0.003851, while the American ASTM E1137 uses a temperature coefficient of 0.003902. 0.003851 is currently the industry standard recognized in China and most other countries.
The calculation process for the temperature coefficient is as follows, taking PT100 as an example.
The resistance at the boiling point of 100℃ is R100=138.51Ω, and the resistance at the freezing point of 0℃ is R0=100Ω. Divide the difference of 38.51 by the nominal resistance and then by 100℃ to get the average temperature coefficient.
3. Accuracy level
IEC 60751 specifies the accuracy classes and permissible errors for platinum resistance thermometers. Taking a Class A platinum resistance thermometer as an example, the maximum temperature error consists of two parts: a fixed error of 0.15℃ caused by the deviation of the nominal resistance value at 0℃, plus an error of 0.002×|T| introduced by temperature coefficient drift. Here, T is the actual temperature measurement range. If T does not exceed the application temperature range of -30 to +300℃ in the accuracy class table, the platinum resistance thermometer will not exceed the permissible error of its accuracy class.
When the measured temperature is 100℃, the total error of a Class A platinum resistance thermometer is 0.15 + 0.002 × 100 = 0.35℃. When selecting a platinum resistance thermometer, its nominal resistance, temperature coefficient standard, accuracy class, and application temperature range are the criteria for our selection.
4. Temperature resistance conversion characteristics
The temperature-resistance conversion relationship of platinum resistance thermometers is described by the following formula, which distinguishes between two cases: below 0℃ and above 0℃.
When T≤0℃: RT=R0?(1+A?T+B?T^2+C?(T-100℃)?T3)
When T≥0℃: RT=R0?(1+A?T+B?T^2)
Where RT is the resistance at temperature T, R0 is the resistance at 0℃; A, B, and C are three constants specified in IEC60751.
The measured temperature T can be solved by directly substituting the resistance value RT into the formula, but it requires solving a cubic equation, which is computationally complex.
To simplify calculations, the temperature resistance curve of PT100 in the range of -200 to +850°C is output using the formula, as shown in the figure below. The resistance of PT100 changes approximately linearly with temperature within the range of 18 to 400Ω.
If we perform two-point linear calibration directly using the two endpoints of -200°C and +850°C to simplify the calculation, the temperature resistance curve within the temperature range is shown in the figure below. At this point, the maximum nonlinear error exceeds 16Ω, which is relatively large.
Generating a table of temperature resistance values based on formulas, and then performing linear interpolation within a small range in the lookup table, is a method that is both simple to calculate and can achieve accurate approximation. IEC 60751 includes a lookup table of temperature resistance values in 1°C intervals.
5. Measuring current
Platinum resistance thermometers almost always use direct current excitation for measurement. The measurement current inevitably generates heat in the resistor, introducing self-heating error. Platinum resistance thermometer manuals specify two parameters: measurement current and self-heating coefficient. The typical measurement current I is 0.3~1mA, and the self-heating coefficient S is about 0.015℃/mW.
The temperature error introduced by the measuring current can be calculated based on the self-heating coefficient, according to the following formula.
For example, given a current of 1mA, with a maximum resistance of 400Ω for the PT100, the self-heating temperature is approximately 0.01℃, in which case the error is almost negligible. When the self-heating coefficient of the platinum resistance thermometer is unaffected, the measuring current should be set to its maximum value. If the current is too low, the output voltage amplitude decreases, and the signal-to-noise ratio drops. 1mA is a commonly used measuring current value.
6. Wiring method
Platinum resistance thermometers (RTTs) have two-wire, three-wire, and four-wire output lead configurations. Among these, the error introduced by the lead resistance in a two-wire configuration cannot be eliminated; a four-wire configuration has no lead resistance error, but requires the most leads; a three-wire configuration is the most commonly used method because, under the same physical dimensions, the three leads have equal lead resistance values, and the lead error can be eliminated through calculation after two resistance measurements.
7. Summary
The accuracy of the temperature measurement circuit depends not only on the initial selection of the RTD, but also on the subsequent hardware design and software algorithm optimization. ZLG Zhiyuan Electronics provides the PT100 interface module TPS02 with a three-wire interface for platinum RTD measurement. It features a high-stability measurement circuit including a built-in excitation current source, a 24-bit ADC, a resistance-temperature value linearization algorithm, and 2500V electrical isolation. By connecting to the platinum RTD, the temperature value can be read through the IIC digital interface.
