ABSTRACT
Temperature, as well as pressure, acceleration, and so on, is a variable that can be measured to acquire information about a physical process to be scientifically described and mastered in an engineering application. To do this, one must interact with the process through a measuring system embedding a physical phenomenon capable of translating the process variable into an ‘‘indicated’’ suitable signal, usually some electrical variable such as voltage, current, capacitance, etc. The indicated signal should be of electrical nature because further processing can be accomplished through analog circuits and digital microprocessors that are basically electronic devices. Possibly, in a near future, signal processing will be accomplished through photonic devices and our transducers will be based on physical phenomena through which the process variable modulates some light-related variable as, for example, the effect of temperature on a fiber Bragg grating sensor. Anyway, the fundamental and frequently overlooked concept here is that the measured or indicated variable is the response to the stimulus imposed by the process and, as such, it contains transformed rather than original information about the process. Thus, any measurement problem is actually an inverse problem (in the mathematical sense of the term) because one wants to recover the original information from the transformed information, that is, the process signal from the indicated signal. The question of if and how this is possible constitutes an important new research area. One important type of sensor is the resistance temperature detector or resistive thermal
device (RTD) whose working principle is based on the change in electrical resistance of
as iron copper. However, platinum is the most common resistance thermometer (PRT) because of its linearity with temperature and chemical stability. RTDs are gradually becoming predominant in industrial applications, particularly in applications under 6008C, due to their higher accuracy and repeatability, in addition to the simplicity of its conditioning electronics compared with thermocouples or other types of thermal sensors. More specifically, the RTD being essentially a resistor element, one can take advantage of a great number of standard electronic measurement techniques and integrated components suitable for measuring under myriads of practical condition. Specifying the most adequate RTD to a particular application can be a difficult task, as it
can be for all other types of sensor. To ensure the desired performance, one must consider a number of aspects such as thermochemical compatibility and materials, dimensions and size, temperature range and dynamics, accuracy, precision and errors, effects of lead wiring configuration, conditioning electronics, and nominal resistance and temperature coefficients. Some of these aspects are informed by the manufacturers of the sensor, electronic components, etc., and others are dictated by the specificities of the application. Platinum RTD (PRT) standards help defining a general frame of reference within which these issues can be addressed. The European standard DIN=IEC 60751, one of the most commonly adopted worldwide, requires that the RTD’s electrical resistance has to be of 100.00V at 08Cwith a temperature coefficient of resistance of 0.00385V=V=8C between 08C and 1008C. In DIN=IEC 60751, there are two classes of resistance tolerances: Class A¼ 100.00 0.06 V @ 08C and Class B¼ 100.00 0.12 V @ 08C. The combination of resistance tolerance and temperature coefficient defines the resistance=temperature characteristics of the sensor and, ultimately, an envelope around the nominal transduction equation within which lies the actual calibration curve of each particular sensor. (This point will be elaborated in the section dedicated to error analysis.) Consequently, the greater the sensor’s resistance tolerance the more the calibration curve will deviate from the generalized curve and more variation there will be from sensor to sensor. Interchangeability is an important issue in applications where the sensor is expected to be replaced from time to time, particularly if the RTD’s information is used for billing purposes, such as in custody transfer in the petroleum industry. As mentioned above, some aspects to be considered when specifying an RTD are
intrinsic to the sensor, and others are application dependent. Among the intrinsic aspects, probably the most important one is the necessary conditioning electronics and lead wiring. An RTD is intrinsically a two-wire resistance that must be connected to its conditioning electronics through lead wires, which introduce stray impedances to the circuit. Therefore, most applications are developed based on three-or four-wire circuitry to compensate for these stray effects producing a truer indication of the measured temperature. Figure 4.1 shows the corresponding diagrams. The three-wire circuit is based on the assumption that the lead wires have the same impedance that can be cancelled out by adding a third resistance to one of the adjoining arms. Due to its simplicity and the availability of highquality connection cables, this is a very common choice in industrial applications in which the distance between the sensor and the conditioning electronics is less than 500 m. The four-wire Kelvin connection uses separate pairs of current-carrying and voltage-sensing cables, providing virtually full cancellation of stray impedances of up to 15 V cables. Due to its complexity, this configuration is commonly restricted to laboratory applications where very high accuracies are required.
