Methods of Measuring Temperature

Methods of Measuring Temperature

Temperature is the most common physical parameter that is measured and converted to electrical form. Several types of temperature transducers respond to temperature and produce a corresponding indication by a change or alteration in a physical characteristic that can be detected by an electronic circuit. Common types of temperature transducers are thermocouples, resistance temperature detectors (RTDs), and thermistors.

The Thermocouple

A thermocouple is formed by joining two dissimilar metals. A small voltage, called the Seebeck voltage, is produced across the junction of the two metals when heated, as illustrated in Figure 1.

Fig. 1. A voltage proportional to temperature
is generated when a thermocouple is

The amount of voltage produced is dependent on the types of metals and is directly proportional to the temperature of the junction (positive temperature coefficient); however, this voltage is generally much less than 100 mV.

The voltage versus temperature characteristic of thermocouples is somewhat nonlinear, but the amount of nonlinearity is predictable. Thermocouples are widely used in certain industries because they have a wide temperature range and can be used to measure very high temperatures.

Some common metal combinations used in commercial thermocouples are chromel-alumel (chromel is a nickel-chromium alloy and alumel is a nickel-aluminum alloy), iron-constantan (constantan is a copper-nickel alloy), chromel-constantan, tungsten-rhenium alloys, and platinum10% Rh/Pt.

Fig. 2. Output of some common thermo-
couples with 0 C as the reference

Each of these types of thermocouple has a different temperature range, coefficient, and voltage characteristic and is designated by the letter K, J, E, W, and S, respectively. The overall temperature range covered by thermocouples is from to -250oC to 2000oC. Each type covers a different portion of this range, as shown in Figure 2.

Thermocouple-to-Electronics Interface

When a thermocouple is connected to a signal-conditioning circuit, as illustrated in Figure 3, an unwanted thermocouple is effectively created at the point(s) where one or both of the thermocouple wires connect to the circuit terminals made of a dissimilar metal.

methods of measuring temperature
Fig. 3. Creation of an unwanted thermocouple in a thermocouple-to-electronics

The unwanted thermocouple junction is sometimes referred to as a cold junction in some references because it is normally at a significantly lower temperature than that being measured by the measuring thermocouple.

These unwanted thermocouples can have an unpredictable effect on the overall voltage that is sensed by the circuit because the voltage produced by the unwanted thermocouple opposes the measured thermocouple voltage and its value depends on ambient temperature.

Example of Thermocouple-to-Electronics Interface

As shown in Figure 4, a copper/constantan thermocouple (known as type T) is used, in this case, to measure the temperature in an industrial temperature chamber. The copper thermocouple wire is connected to a copper terminal on the circuit board, and the constantan wire is also connected to a copper terminal on the circuit board.

temperature measurement methods
Fig. 4

The copper-to-copper connection is no problem because the metals are the same. The constantan-to-copper connection acts as an unwanted thermocouple that will produce a voltage, Vx, in opposition to the thermocouple voltage because the metals are dissimilar.

Since the unwanted thermocouple connection is not at a fixed temperature, its effects are unpredictable and it will introduce inaccuracy into the measured temperature. One method for eliminating an unwanted thermocouple effect is to add a reference thermocouple at a constant known temperature (usually 0oC).

Fig. 5. Using a reference thermocouple in a
temperature-measuring circuit.

Figure 5 shows that by using a reference thermocouple that is held at a constant known temperature, the unwanted thermocouple at the circuit terminal is eliminated because both contacts to the circuit terminals are now copper-to-copper. The voltage produced by the reference thermocouple is a known constant value and can be compensated for in the op-amp circuitry.

Compensation: It is bulky and expensive to maintain a reference thermocouple at a fixed temperature (usually an ice bath is required). Another approach is to compensate for the unwanted thermocouple effect by adding a compensation circuit as shown in Figure 6. This is sometimes referred to as cold-junction compensation.

Fig. 6. Simplified circuit for compensation
of unwanted thermocouple effect.

The compensation circuit consists of a resistor and an integrated circuit temperature sensor with a temperature coefficient that matches that of the unwanted thermocouple. The current source in the temperature sensor produces a current that creates a voltage drop, Vc, across the compensation resistor, Rc. The resistance is adjusted so that this voltage drop is equal and opposite the voltage produced by the unwanted thermocouple at a given temperature.

When the ambient temperature changes, the current changes proportionally, so that the voltage across the compensation resistor is always approximately equal to the unwanted thermocouple voltage. Since the compensation voltage, Vc, is opposite in polarity to the unwanted thermocouple voltage, the unwanted voltage is effectively cancelled.

Most thermocouple applications require the cold-junction compensation illustrated in Figure 6 as well as other signal conditioning such as isolation and linearization. These functions are available in special integrated circuits and hybrid modules called thermocouple signal conditioners.

The 1B51 is an example of a thermocouple signal conditioner. It includes a transformer isolation unit and is designed with very high common-mode rejection because thermocouples have a small signal, which is susceptible to interference, particularly in industrial environments.

A built-in 3 Hz low-pass filter helps reject interfering signals such as power line pickup. The cutoff frequency is set to be very low because the thermocouple cannot respond immediately to temperature changes. The 1B51 generates its own input side power to provide channel-to-channel isolation when multiple thermocouples are used.

Signal conditioning is widely used in industrial environments, so sophisticated controllers with multiple isolated inputs and outputs have been developed for use with thermocouples and other sensors. Outputs from these controllers include wireless transmitters, USB, Ethernet, PCI, IEEE-488 and other computer interfaces, making it simple to log data on a computer or dedicated data logger.

Methods of Measuring Temperature

Resistance Temperature Detectors (RTDs)

A second major type of temperature transducer is the RTD (resistance temperature detector). The RTD is a resistive device in which the resistance changes directly with temperature (positive temperature coefficient). The RTD is more nearly linear than the thermocouple.

RTDs are constructed in either a wire-wound configuration or by a metal-film technique. The most common RTDs are made of either platinum, nickel, or nickel alloys. Generally, RTDs are used to sense temperature in two basic ways.

Fig. 7(a)

First, as shown in Figure 7(a), the RTD is driven by a current source and, since the current is constant, the change in voltage across it is proportional (by Ohm’s law) to the change in its resistance with temperature.

Second, as shown in Figure 7(b), the RTD is connected in a 3-wire bridge circuit; and the bridge output voltage is used to sense the change in the RTD resistance and, thus, the temperature.

Fig. 7(b)

As in the case of thermocouples, special signal-conditioning integrated circuits are available for RTDs. An example is the 1B41, which uses the same pinout as the 1B51, described previously for thermocouples. The 1B41 has an internal transformer-coupled isolation amplifier with high common-mode rejection. It generates its own input side power and includes special linearization circuits designed especially for RTDs.

The 3-Wire Bridge: To avoid subjecting the three bridge resistors to the same temperature that the RTD is sensing, the RTD is usually remotely located to the point where temperature variations are to be measured and connected to the rest of the bridge by long wires. The resistance of the three bridge resistors must remain constant. The long extension wires to the RTD have resistance that can affect the accurate operation of the bridge.

Figure 8(a) shows the RTD connected in the bridge with a 2-wire configuration. Notice that the resistance of the long connecting wires is shown as “lumped” resistors RA and RB in Figure 8. These resistors are in the same leg of the bridge as the RTD. Recall that Vout = 0 and the bridge is balanced when RRTD = R3 if R1 = R2.

Fig. 8(a)

The wire resistances will throw the bridge off balance when and will cause an error in the output voltage for any value of the RTD resistance because they are in series with the RTD in the same leg of the bridge.

The 3-wire configuration in Figure 8(b) overcomes the wire resistance problem. By connecting a third wire to one end of the RTD as shown, the resistance of wire A is now placed in the same leg of the bridge as R3 and the resistance of wire B is placed in the same leg of the bridge as the RTD.

Fig. 8(b)

Because the wire resistances are now in opposite legs of the bridge, their effects will cancel if both wire resistances are the same (equal lengths of same type of wire). The resistance of the third wire has no effect; essentially no current goes through it because the output terminals of the bridge are open or are connected across a very high impedance.

The balance condition is expressed as

RRTD + RB = R3 + RA

If RA = RB, then they cancel in the equation and the balance condition is completely independent of the wire resistances.

RRTD + = R3

The method described here is important in many measurements that use a sensitive transducer and a bridge. It is often used in strain gauge measurements.

Basic RTD Temperature-Sensing Circuits

Two simplified RTD measurement circuits are shown in Figure 9. The circuit in part (a) is one implementation of an RTD driven by a constant current. The operation is as follows.

Fig. 9(a)

From your study of basic op-amp circuits, recall that the input current and the current through the feedback path are essentially equal because the input impedance of the op-amp is ideally infinite.

Therefore, the constant current through the RTD is set by the constant input voltage, VIN, and the input resistance, R1, because the inverting input is at virtual ground.

The RTD is in the feedback path and, therefore, the output voltage of the op-amp is equal to the voltage across the RTD. As the resistance of the RTD changes with temperature, the voltage across the RTD also changes because the current is constant.

Fig. 9(b)

The circuit in Figure 9(b) shows a basic circuit in which an instrumentation amplifier is used to amplify the voltage across the 3-wire bridge circuit. The RTD forms one leg of the bridge; and as its resistance changes with temperature, the bridge output voltage also changes proportionally. The bridge is adjusted for balance (Vout = 0) at some reference temperature, say 0oC. This means that R3 is selected to equal the resistance of the RTD at this reference temperature.


A third major type of temperature transducer is the thermistor, which is a resistive device made from a semiconductive material such as nickel oxide or cobalt oxide. The resistance of most thermistors changes inversely with temperature (negative temperature coefficient).

The temperature characteristic for thermistors is more nonlinear than that for thermocouples or RTDs; in fact, a thermistor’s temperature characteristic is essentially logarithmic. Also, like the RTD, the temperature range of a thermistor is more limited than that of a thermocouple.

Thermistors have the advantage of a greater sensitivity than either thermocouples or RTDs and are generally less expensive. This means that their change in resistance per degree change in temperature is greater. Since they are both variable-resistance devices, the thermistor and the RTD can be used in similar circuits.

Thermistors have a relatively narrow range of temperatures over which they can respond (about -40oC to 160oC) and are highly nonlinear; however, compensation circuits can make up for nonlinearity.

Despite these shortcomings, thermistors have important advantages over other types of sensors when working within their temperature range. They are small, inexpensive, very sensitive, and can be highly accurate within their temperature range. Because they are small and have a small thermal mass, they can respond quickly to temperature change.

Like the RTD, thermistors can be used in a constant-current configuration or a Wheatstone bridge. The constant-current configuration forces a constant current through the thermistor and the voltage drop produced is measured.

Applications for Thermistors

Thermistors are used in many applications for which temperature monitoring is important. Heating systems use thermistors in thermostats to sense the temperature in a heated or cooled space and also to sense duct temperatures. The signals from these thermistors are used by a controller for turning on or off heat and fans.

Automobiles use thermistors in engine and power-train management and control, inside temperature control including duct temperatures, and overheating sensors for monitoring coolant temperature. Medical applications include highly accurate patient thermometers and infant monitoring, temperature baths, and respiratory probes.

Zener Based IC Precision Temperature Sensor

Various types of solid-state temperature sensors exploit various temperature dependencies in semiconductors, which include resistivity effects and the change in the base-emitter voltage with temperature of a forward-biased transistor.

One type uses the zener diode breakdown voltage as a temperature indicator. The breakdown of a zener diode is directly proportional to the absolute (Kelvin) temperature. The zener breakdown voltage equals + 10 mV/K. The absolute temperature in kelvins (K) is equal to 273 + oC

Unlike thermistors, the output of the zener temperature sensor is linear. This type of sensor has the advantage of being small, accurate, and linear; however, these sensors are limited in the temperature range over which they can operate (about -40oC to 150oC).

The LM135, LM235, and LM335 are examples of zener-based temperature sensors and differ only in the range of temperatures over which they can operate. This type of sensor has an adjustment input for calibration purposes. It is not necessary to use the adjustment input unless high accuracy is required.

The LM135/LM235/LM335 is an integrated circuit available in different package styles. The circuit is much more complex than a simple zener diode but, in effect, displays a very precise zener characteristic and, most importantly, an output (reverse voltage) that varies linearly with temperature as previously mentioned.

The adjustment input allows you to calibrate the device at one temperature if high accuracy is required. By calibrating the output correctly at one temperature, the output at all temperatures is correct because of the linearity.

Leave a Comment

Your email address will not be published. Required fields are marked *