of these applications involve profiling the temperature of hundreds of devices under test (DUTs) simultaneously. For example, testing laboratories often need hundreds of channels to monitor temperatures of heat-generating consumer devices such as ovens, dryers, automobile seats, and stoves; or component level devices such as precision thermoelectric coolers (TECs) and amplifier modules. Fortunately, today’s test system builders have a variety of sensor and instrumentation options from which to choose when configuring temperature measurement applications.
The most widely used temperature sensor options include thermocouples (T/Cs), resistance temperature detectors (RTDs), thermistors, solid state monolithic (IC) temperature sensors, and infrared (IR) sensors, plus a few more exotic choices such as temperature-sensitive paint. If implemented correctly and matched appropriately with the application, any of these will produce useful results. This article offers guidelines that can help ensure the reliability of temperature measurement results.
The first factor to consider when selecting a temperature sensor is the range of temperatures to be measured. Other factors that should be taken into account include the level of accuracy required, the ease of sensor replacement, the types of instruments required to interface with the sensors, and the physical environment (for example, some RTDs are too fragile to withstand much shock or vibration). This article will consider the advantages and disadvantages of some common sensors, as well as the instrumentation required to convert sensor signals into measurement results. When deciding which type of sensor to use, be aware that the thermocouple is the most versatile and useful for significant distances between the sensor and the instrument, the thermistor is the most sensitive (which means the thermistor’s resistance will change much more in response to temperature changes), the four-wire RTD is the most stable, and the three-wire RTD minimises the number of conductors per sensor.
Thermocouples, the most widely used type of temperature sensor, are based on the thermoelectric or Seebeck effect, named after physicist Thomas Seebeck. When two wires made up of dissimilar metals are joined together, a voltage is generated. The level of generated voltage is a function of temperature. As temperature changes, the voltage changes, so the thermocouple voltage equates to a temperature reading. This is the basic operating principle of the thermocouple.
The linearity of a thermocouple’s output varies depending on thermocouple type and temperature range. Although thermocouples are rugged and can cover a very wide range of temperatures (some as high as 2300 °C), they produce very small output voltages, so the instrumentation used with them must offer sufficient resolution to discern and distinguish among small voltage changes.
Some sources of temperature measurement error are specific to the use of thermocouples; the CJC (Cold Junction Compensation) may not be configured or compensated — thermocouples measure the difference in temperature between the hot junction (used for the actual measurement) and the cold or reference junction (at the instrument). If the temperature of the cold junction is not known or not compensated for (by means of cold junction compensation), the temperature readings will be inaccurate.
Connections between the thermocouple and the instrument should be made with the same kind of wire used for the thermocouple. In theory, one could use copper wire, but it would require controlling the temperature of all the wire, which is generally impractical.
Thermocouple outputs are just microvolts, so the instrument used must have sufficient resolution to measure those low voltages accurately. Many modern digital multimeters (DMMs) have built-in support capabilities that make them particularly appropriate for temperature measurement applications with thermocouples. For example, Keithley’s Model 3706 System Switch/Multimeter supports thermocouple types J, K, N, T, E, R, S, and B. Depending on which type of thermocouple is used, the Model 3706 can measure temperatures ranging from -200°C to 1820°C at resolution levels from 0.001°C (with type J) to 0.1°C (with type B).
When a thermocouple is connected directly to the input of a DMM, at least one of those connections will be a junction made up of two dissimilar metals, introducing a thermoelectrical voltage that will be algebraically added to the thermocouple voltage, producing an erroneous temperature measurement. To cancel the effects of this unwanted thermoelectrical voltage where dissimilar wire connections must be made, the thermocouple circuit requires a reference junction that is held at a stable, known temperature.
For example, as long as the temperature of this reference (cold) junction is known, the Model 3706 can factor in the reference temperature to calculate the actual temperature reading at the thermocouple. Although the standard reference temperature used as the fundamental reference for NIST’s voltage-to-temperature conversion tables is the ice point (0°C), other known temperatures can be used. The Model 3706 can acquire the cold junction temperature by measuring the cold junction using a thermistor or four-wire RTD, or the user can enter the known temperature value.
The most accurate thermocouple measurements are achieved by using a simulated reference junction using an ice point reference (Figure 1). The copper wire to thermocouple wire connections are immersed (but electrically isolated) in the ice bath, and the user enters the 0°C simulated reference temperature into the instrument. The simulated reference temperature for the Model 3706 can be set from 0° to 65°C.
Figure 1: A simulated reference junction using an ice point reference.
Long lengths of thermocouple wire can have a significant amount of capacitance that is seen at the input of the DMM. If an intermittent open occurs in the thermocouple circuit, this capacitance can produce an erroneous on-scale reading. To prevent these errors, the Model 3706 provides an open thermocouple detection circuit, which, when enabled, applies a 100A pulse to the thermocouple before the start of each temperature measurement.
Resistance temperature detectors
RTDs measure temperature by monitoring changes in the electrical resistance of metal wires or films. The wire used is usually platinum, although RTDs are also constructed of other metals, including nickel, a nickel/iron alloy, and copper. RTDs exhibit the greatest stability and linearity of all the temperature transducers described in this article.
Just as is true with thermocouples, several sources of temperature measurement error are specific to RTDs.
An RTD must conduct a current in order to read temperature, and this current will cause the RTD to heat up. The level of current through a Pt100 RTD (so-called because it has a resistance of 100 ohms at 0°C), for example, must be kept to less than 1mA to avoid self-heating problems; the current must be even lower for nickel or nickel/iron alloy RTDs.
The resistance of the test leads can have a great effect on the accuracy of an RTD temperature measurement; when leads are longer than a few inches, it is usually better to use a three-wire or four-wire (Kelvin) connection.
Although a platinum wire RTD can be used from –240°C to 649°C (–400°F to 1200°F), and a platinum thin-film RTD from –196°C to 538°C (–320°F to 1000°F), a nickel RTD, with a larger output than a Pt unit, is limited to –350°C to 316°C (–212°F to 600°F). A nickel/iron alloy RTD, although it has a much larger output than a platinum RTD, is limited to –73°C to 204°C (–100°F to 400°F).
Although an RTD can be exceedingly accurate, its output is small; for example, a 1°C temperature change will cause a resistance change of only 0.385 ohms in a standard Pt100 RTD, so the measuring instrument must be able to measure small changes in resistance accurately.
Platinum RTDs are either wire-wound or metal film resistors, with the latter exhibiting the faster response time. Because a Pt100 sensor is basically a resistor, its value can be measured with an ohmmeter. However, the low resistance of the sensor and its low sensitivity (0.385 Ω/°C) make accurate measurements difficult due to lead resistance. A 1 ohm resistance in each lead connecting the Pt100 to the meter will cause an error of more than 5°C. To avoid the problem of lead resistance errors, most Pt100 measurements are made using a four-wire configuration, in which two of the wires are used to provide an excitation current and the other two connect a voltmeter over the RTD. Provided the impedance of the voltmeter is high, then a few ohms of resistance in the cables will not cause an error.
The Model 3706 offers an RTD temperature measurement range of -200°C to 630°C (0.01°C resolution). It supports a number of RTD types including PT100, D100, F100, PT385, and PT3916.
By default, a Model 3706 performs four-wire measurements using offset-compensated ohms, which provides the most accurate way to measure the low resistance of the RTD. For faster RTD measurements in situations when the application doesn’t require the most accurate measurements, offset-compensation may be disabled.
The word thermistor is short for temperature sensitive resistor. A thermistor’s resistance changes nonlinearly with changes in temperature. For thermistors with a negative temperature coefficient (NTC), as temperature increases, their resistance decreases; for positive temperature coefficient (PTC) thermistors, an increase in temperature produces an increase in resistance. Although thermistors produce larger outputs than RTDs, their temperature range is more limited, their accuracy is less than that of RTDs, their interchangeability is poor to fair, and their long-term stability is poor. Although they are a good choice when the application requires measuring slight changes in temperature, the downside of their increased sensitivity is loss of linearity. Because they are especially non-linear at high temperatures, it is best to limit their use to measurements of temperatures less than 100°C.
Thermistors can also contribute to temperature measurement errors. A thermistor, like an RTD, must conduct a current in order to read temperature, and this current will increase its temperature. It’s important to limit the excitation current to the smallest amount that will produce a satisfactory output.
Although a thermistor has a much larger output than an RTD or thermocouple, its operating temperature is much more limited.
After a DMM measures the resistance of the thermistor, it employs various equations to calculate the temperature reading. The equations used to calculate thermistor temperature employ curve fitting constants. Be aware that the thermistor manufacturer’s specified curve fitting may not be exactly the same as that used by the DMM manufacturer.
Common sources of errors
For any temperature sensor type, if the sensor is connected with fairly thick wires, enough heat may be conducted along those wires to change the temperature reading. It’s generally best to ensure the leads near the sensor are close to the temperature being measured.
If the sensor makes poor thermal contact with the object to be measured, it is likely to be at a different temperature. When measuring the temperature of a surface, fasten the sensor solidly to it. If the temperature is moderate, some thermal (heat sink) compound can also be used to improve thermal contact.
When measuring fluids, ensure the temperature sensor is reading the temperature of the main body of the fluid, not the temperature in a spot where the fluid is not circulating.
Although temperature sensors such as thermocouples, RTDs, and thermistors are compatible with many types of measurement instruments, digital multimeters are typically the most common choice for temperature applications and are often used in conjunction with switching hardware if the application in question requires monitoring the temperature at multiple points. For example, the Model 3706 System Switch/Multimeter provides numerous other features that make it suitable for a variety of temperature monitoring and control applications:
Up to 360 thermocouple channels in a single 2U chassis.
Automatic cold junction compensation (CJC) on the compatible Model 3720, 3721, and 3724 Multiplexer Cards with a screw terminal accessory for thermocouple-type temperature measurements.
Built-in support for measuring temperature with three thermistor types: 2.2k , 5k , and 10k.
LXI/Ethernet connection for simplified temperature monitoring in remote locations.
Option to expand to additional temperature monitoring channels in additional Series 3700 chassis via the built-in TSP-Link interface.
14 programmable digital I/O lines allow controlling external devices, such as component handlers or other instruments, or sending alarm indications if a critical temperature parameter exceeds tolerance.
Ensuring temperature measurement accuracy is not particularly difficult if sensors and instrumentation appropriate to the application are chosen and properly configured. The first, critical step is to understand the strengths and limitations of the options available and find the correct combination of functions and accuracy.
Dale Cigoy is an Applications Engineer with Keithley Instruments, Inc.