Understanding Infrared Thermometry

John R. Gyorki
Editorial Director

Infrared (IR) thermometry appears to be rather straight forward: point, press the button, and read the temperature. However, measurement results will be quite disappointing without a thorough understanding of the instruments’ principle of operation and specifications.

Temperature measurement instruments can be divided into contact and noncontact types. Sensors used in contact-type instruments include thermocouples, resistance temperature detectors (RTDs), thermistors, and semiconductor temperature sensors. Since contact sensors measure their own temperature they require physical contact with the measured object to bring the sensor body to the object’s temperature.

In some applications this contact creates problems: The measured object or media may be located at a distance or in a hazardous environment with no easy access. Measurements of moving objects are also difficult. A small object’s temperature may be altered when a relatively large sensor touches it and acts as a heat sink.

Noncontact infrared (IR) thermometers, if used properly, offer convenient solutions for these and many other measurement applications. However, you should select the measuring instrument and measurement techniques to be compatible with the application.

How IR Thermometry works
Heat is transferred from one body to another through conduction, convection, or radiation. Radiation is a process where heat energy in a form of electromagnetic waves is emitted by a hot object and absorbed by a colder object. Most of this radiation is in the infrared (IR) region of the electromagnetic spectrum, but some also spreads into the visible light band. The IR wavelength band stretches from 0.7 to 1000 microns, however practical IR measurement systems use only certain wavelength bands between 0.7 and 14 microns because the radiation is the strongest in this range.

If an object is exposed to IR energy radiated by a heat source, such as an electric heater, light bulb, sun, or other source, the energy reaching the object is called incident energy.  Part of this energy is reflected off the object surface. Theoretically, the object’s coefficient of reflectivity can vary from 0 (no reflection) to 1.0 (100% reflection). Rough, matt surfaces have low reflectivity. Polished and glossy surfaces, especially metals, have high reflectivity.

Depending on the object material, thickness, and the radiation wavelength, part of the radiation can go through the object or be transmitted. The coefficient of transmission can vary from 0 (no energy transmitted through object) to 1.0 (100% energy transmitted through object). High transmittance examples include glass, quartz, plastic film, and various gasses. Materials opaque in the IR spectrum have close to zero transmission coefficients.

The remaining energy is absorbed by the object and raises its temperature. A hypothetical body that has no reflection or transmission and absorbs all incident energy across the entire spectrum has a coefficient of absorption equal to 1.0 and is called a blackbody. Real-life objects, referred to as gray bodies, have coefficients of absorption that fall between 0 and 1.0.

Incident energy, W_I, is defined as:

W_I = W_R + W_T + W_A

Where:
W_I =  incident energy received by the object, W
W_R = energy reflected off the object’s surface, W
W_T = energy transmitted by the object, W
W_A = energy absorbed by the object, W

As the object absorbs energy and heats, it also emits energy.  When an object is in a state of thermal equilibrium, the amount of energy it absorbs (W_A) equals the amount of energy it emits (W_E): W_A = W_E. When an object absorbs more energy and its temperature increases, the amount of radiation it emits also increases.

IR thermometry is based on the fact that any body (solid, liquid, or gaseous) that has a temperature above absolute zero (0^oK or -273^oC) emits radiant energy. This energy is proportional to the forth power of the body temperature, and the body’s ability to absorb and emit IR energy is called emissivity. Energy radiated by a body can be expressed as follows:

W = E σ T4 A

Where:

W = energy, W
E = emissivity
σ = Stefan-Boltzmann Constant = 5.6703 10^-8, W/m²K⁴
T = absolute temperature, ^oK
A = emitting area, m²

Emissivity can range from 0 to 1 for various bodies. A hypothetical blackbody emits and absorbs all energy and thus has an emissivity equal to 1. Real-life objects have an emissivity between 0 and 1.

When an IR thermometer measures an object’s temperature, consider the energy that actually enters the lens. That is, in addition to emitting energy related to its own temperature, the object may reflect energy coming from another source, or transmit energy passing through it from a source behind it. For accurate measurements, survey the surrounding area for possible sources of extraneous IR radiation and choose the thermometer position and aiming angle to minimize the effects of those sources.

IR Thermometers
Infrared temperature measurement instrument design varies from simple hand-held thermometers that can be purchased for less than a hundred dollars to complex special-purpose instruments that cost hundreds and even thousands of dollars. However, some building blocks are common for most designs.

A typical infrared thermometer consists of optical components, IR detector, electronics, and a display or interface output stage. Optical parts focus radiation energy onto the IR detector and filter out radiation outside the desired wavelength band.  These components include collecting optics, lenses, fiber optics, and spectral optical filters.

IR Detectors
The majority of IR detectors are either single-wavelength (also called single-color), or dual-wavelength (also called two-color) type. The single-wavelength detectors measure IR energy within a certain wavelength band, and the instrument calculates object temperature based on the detector output and the preset emissivity. Some thermometers have adjustable emissivity, and most simple units have fixed emissivity.

Dual-wavelength detectors measure energy at two different wavelength bands, and the instrument calculates temperature based on the ratio of the two readings. If emissivity or the energy changes by the same amount at both bands, the measurement accuracy is not affected. Emissivity or the amount of radiated energy may change due to object change or movement, lens contamination or misalignment, or view obstruction.  The dual-wavelength detector’s drawback is higher cost and lower accuracy under certain conditions.

Emissivity of many materials and surfaces remains relatively constant over the IR wavelength range, and measuring energy in any narrower band will be acceptable. Other materials have wavelength bands with higher and lower emissivity due to high reflectivity or transmission, and require narrow band detectors tuned to high emissivity wavelengths.

Another factor is the atmosphere. Its transmission coefficient vs. wavelength curve has many peaks and valleys, which swing from almost 1.0 to near zero and block the IR energy transmission. Most general-purpose IR thermometers use the largest high-transmission band between 7 and 14 microns to minimize atmospheric attenuation.

To measure temperature of objects with emissivity that varies greatly over the IR wavelength spectrum and objects obscured by glass, smoke, steam, or other barriers, engineers need to use narrow band IR detectors. For example, short-wavelength detectors handle variable emissivity objects, lens contamination, and measurements through glass windows. Long-wavelength detectors are more prone to errors due to emissivity changes, but have a wide temperature range.

Special applications, such as measuring the temperature of glass, crystal, flame, gas, and thin film require detectors with specific narrow bands.  For example, detectors with a narrow band centered on 5 microns give the best results when measuring glass temperature.  Metals and metallic foils usually require 1 micron detectors where they have the highest level of radiation.

Based on the principle of operation, IR detectors fall into one of two categories: thermal detectors and photo detectors (photodiodes). Thermal IR detectors absorb the incident energy, raise the sensing element temperature, and change the detector’s electrical properties: thermopiles generate thermoelectric voltage, bolometers change resistance, and pyroelectric devices change their polarization. In general, they are slower than photo detectors.

A thermopile is made by connecting several thermocouples in series and placing their hot junctions in contact with a black body that absorbs the incident IR energy and heats the hot junctions. The cold junctions are placed in the area of the detector with adequate heat sinking. These detectors have fast response, broad band, large dynamic range, and are frequently used in general-purpose, automotive, air conditioning, and human-body thermometers.

Bolometers use a slab of material that changes its resistance in response to a change of temperature. The circuit converts resistance change to a voltage change, which is further processed by the instrument. Bolometers are frequently used for measuring low-level IR energy, often as an attachment to a telescope.

Pyroelectric devices become electrically charged when their body temperature changes. To produce a usable signal, the incident IR energy has to “pulse”. The output peak-to-peak AC signal is proportional to the pulse energy. Since energy emitted by measured objects is usually steady, thermometers that use pyroelectric detectors have a mechanical or optical chopper in front of the sensor. These sensors are used in many home security systems.

Photo detectors are built on a silicon substrate with an IR sensitive area that releases free electrons when impacted by the photons. The flow of electrons produces electrical signals proportional to the incident energy. These detectors are often used as arrays in thermal imaging systems.

A detector needs protection from the environment, and the selected window material must allow the correct wavelength band to pass through with minimum attenuation. A zinc sulfide or germanium window is best for the long-wavelength detectors, glass is suitable for short-wavelength detectors, and quartz for the mid-wavelength spectrum. Some instruments use a fiber-optic light guide to direct the radiation to the detector.

Since all types of IR detectors produce signals in the microvolt range, a high-gain amplifier should follow the detector.  Detector output vs. temperature curves are not linear and fluctuate greatly with a change in ambient temperature. To remedy this a signal-conditioning circuit stabilizes the temperature and linearizes the signal. Many applications require an analog-to-digital converter (ADC) to convert the temperature reading to a digital format.

Hand-held and many other instrument types have a built-in display, while other devices connect to a computer, data acquisition system, or temperature control system via an RS232 or RS-485 cable.  Some instruments simulate a thermocouple output, others have a 0 – 20 mA or 4 – 20 mA current loop, or voltage output.

The key specifications and considerations for any IR thermometer application are field of view (FOV) and distance; spectral band; response time; accuracy and repeatability; emissivity of the object or media being measured; media between the object and IR thermometer, such as vacuum, air, steam, gas, glass, or other; object temperature range; mounted or hand-held application; and type of output signal or display.

The FOV characterizes the diameter of a circle (target) that the IR detector will “see” at a certain distance from the measured surface. However, there is always a minimum target diameter that depends on the optical system and detector size. The detector measures and averages temperatures of all objects within the target area.  FOV is commonly called distance-to-spot size ratio and is a ratio of the distance between the meter and the target to the diameter of the target.

For example, a 10:1 distance-to-spot size ratio means that if a measured surface is located 10 inches from the thermometer it will measure and average the temperature of a circle with 1 in.diameter. Move the thermometer away to 20 in. and the target will increase to 2 in., and so on. A thermometer with a 1:1 ratio will measure within a one-foot diameter circle when held one foot away from the target.

Thermometers designed for measurements of small areas have a very narrow FOV and measure temperatures of objects less than a tenth of an inch. For example, such a thermometer held near a component on a pc board will measure the temperature of just that component and ignore the
components around it.

Other optical systems allow accurate temperature measurement of a spot several inches in diameter at a distance of tens of feet. However, such measurements require accurate pointing. Though notches on top of the instrument provide some help, aiming lights and built-in laser pointers prove to be most helpful.

Unfortunately, a laser pointer may occasionally lead to erroneous measurements if the user is not familiar with the IR thermometer operation and the FOV concept. Some first-time users mistakenly think that the laser beam that they see has something to do with the process of measuring temperature.  They presume that the instrument displays the temperature of the tiny spot where the laser beam meets the surface. Such measurements will not yield satisfactory results.

Practical considerations

Here are some helpful tips:
• Avoid degrading measurement accuracy by environmental elements, such as dirt, dust, smoke, steam, other vapors, extremely high or low ambient temperatures, and electromagnetic interference from other devices.
• Select an IR thermometer with a wavelength band compatible with the measured object (especially high reflectivity objects) and with the media between the thermometer and measured object (especially glass, smoke, or steam).
• Select an instrument with a temperature range not much greater than the maximum application temperature. Wider than needed temperature ranges lead to lower accuracy or higher instrument cost.
• An IR thermometer averages the temperature of all objects within its field of view: Select the instrument with an appropriate FOV, and calculate the proper distance so that only the desired area is measured.
• Avoid hot objects near the measured object. They radiate energy that can be reflected or transmitted by the measured object into the thermometer FOV.

For more information go to:

www.omega.com

www.watlow.com

www.gesensing.com

www.raytek.com

www.murata.com

Designing with RTD temperature sensors

By John R. Gyorki
Editorial Director

Analyze the RTD’s strengths and weaknesses with respect to the application before making a selection.  The application ultimately determines the RTD’s specifications.

Watlow’s RTD sensors are designed to ensure precise and repeatable measurements as well as meet environmental requirements for each application. A high signal-to-noise ratio output increases the accuracy of data transmission and permits greater distances between the sensor and the measuring equipment. These resistance-wire RTDs have a positive temperature coefficient; the resistance change is proportional to the temperature measured.

Resistance Temperature Detectors (RTD) typically operate within a broad temperature range of -200°C to +850°C, are fairly linear, and have excellent long-term stability. Unlike thermocouples, cold junction compensation is not needed, and their temperature range and linearity are superior to both thermistors and thermocouples. When applied correctly, RTDs exhibit extremely low drift, so they do not require recalibration. On the other hand, their comparative weaknesses include lower sensitivity, slower response time, and susceptibility to self-heating. All these qualities make them application-specific.

Principle of operation
RTDs are passive components that require an excitation current to produce an output signal. Similar to thermistors, their resistance “varies” in direct proportion to changes in temperature. The temperature-sensitive element is made of metal or a metal alloy, which gives them the positive temperature coefficient.

Platinum, gold, silver, tungsten, nickel, and copper have been successfully used in different RTD devices, however, platinum is superior to the other metals. It has the highest resistivity, 59 ohms per circular mil foot (Ω/cmf), a wide temperature range, good linearity, and low long-term drift. Nickel or nickel alloy RTDs are more economical, but have a narrower temperature range, poorer linearity, and a greater long-term drift.

There are two major types of RTDs: wire-wound and thin film and each has some advantages and disadvantages. For example, a typical wire-wound RTD uses platinum wire wound around a ceramic or glass bobbin in one of two configurations: birdcage and helix. The birdcage winding construction keeps the platinum wire loose and lets it expand and contract freely with a change in temperature. This minimizes long-term stress-induced resistance change, but has very poor resistance to vibration and is primarily limited to lab use. In a sealed helix-constructed, wire-wound RTD, the bifilar winding is wound around the bobbin and then sealed with molten glass, ceramic cement, or another high-temperature, non-conductive coating. This construction helps protect the wire from vibration, but it is prone to long-term stress induced resistance change when the bobbin and platinum wire have different temperature coefficients of expansion.

In the simplest wire-wound RTD construction (top), thin platinum wire is wound around an insulator bobbin. The wire ends are spot welded or high-temperature soldered to the lead wires. A non-conductive protection coat with good thermal transfer properties covers the whole RTD element assembly. In the thin-film type, the sensing element is formed by depositing a thin layer of platinum onto a ceramic substrate and attaching the leads to the connecting pads. A glass coat (not shown) encapsulates the element.

Newer thin-film type RTDs are made by depositing platinum or another metal alloy film onto a substrate, etching the shape of the resistive element, and then sealing the sensor. The thin film devices are smaller, faster, and considerably less expensive than the wire-wound parts. Thin-film platinum RTDs have virtually linear resistance versus temperature curves and provide a low cost alternative to high-accuracy, wire-wound devices. The drawbacks of film-type RTDs include poor long-term stability and narrower temperature range.

The RTD temperature coefficient represents the sensors’ sensitivity to temperature change. The larger the temperature coefficient (α), the larger the resistance change (ΔR) in response to an ambient temperature change (ΔT):

ΔR = αR_o ΔT,

Where:

α = temperature coefficient, Ω/Ω/°C
R_o = nominal sensor resistance at 0°C, Ω
ΔT = temperature change from 0°C, °C

According to the DIN 43760 standard, the resistance-temperature coefficient of platinum wire typically used in RTD manufacturing is 0.00385 Ω/Ω/°C at 0°C. Another frequently mentioned value, 0.00392 Ω/Ω/°C at 0°C, is the resistance-temperature coefficient of chemically pure platinum wire used for standards. To illustrate using the equation above, consider an ideal 100 Ω RTD that has a resistance of 100.000 Ω at 0°C. Therefore, at +1°C the RTD resistance will be:

R_T = [R_o + (αR_oΔT)] = 100 + (0.00385)(100)(1) = 100.385 Ω.

A problem here is that the RTD temperature coefficient changes over the temperature range, so to obtain an accurate value at any given temperature, a curve-fitting process is required. Use the Callender-Van Dusen equation to calculate the RTD resistance over the entire temperature range:

R_T = R_o + R_o α  [T – δ(T/100 – 1) (T/100) – β(T/100 – 1)(T/100)3]

Where:

R_T = resistance at temperature T, Ω
R_o = nominal RTD resistance at 0oC, Ω
α = temperature coefficient, Ω/Ω/oC
δ = 1.49 for pure platinum
β = 0 if T > 0
β = 0.11 if T < 0

Excitation current vs. resistance
Because RTDs are resistors, they need an excitation current to produce an output voltage. At a given current value, an RTD with higher resistance will have higher voltage resulting in lower required amplification and higher signal-to-noise ratio. However, increasing the RTD’s resistance considerably slows its response, which might be unacceptable for many measurements.

An RTD does not produce any voltage by itself. A source of voltage and an excitation resistor, Rex are needed to make the RTD work. The excitation resistor and the RTD form a voltage divider. Voltage drop across the RTD is proportional to its resistance, so when RTD resistance changes with temperature, the change in voltage represents the change in temperature.

Selecting a low-resistance RTD for remote sensor installations can also present a problem. Long lead wires can add significant error to the temperature measurement. For example, if an RTD with a nominal resistance of 100 Ω is installed at a distance of 200 ft from the signal conditioning circuit, and the two lead wires are made of 24 AWG stranded (7 strands) tinned copper wire, the output voltage seen by the voltmeter will be the sum of the voltage drops across the RTD and both lead wires.

The wire selected for this example has a resistance of 0.023Ω/ft, so the total wire resistance is calculated as:  R_wire = (0.023)(200 + 200) = 9.2 Ω. The total RTD resistance plus the wire resistance is 109.2 Ω, which adds a 9% error to the measurement. The problem stems from the fact that the same two wires are used to supply the excitation current and make the measurement. To make matters worse, copper resistance changes with temperature, making compensation difficult. Therefore, to obtain an accurate temperature reading, the excitation current wires and the measuring connection wires are separated in a 4-wire, Kelvin connection.  One pair of wires supplies constant excitation current, while the other pair connects the voltmeter (or signal conditioning circuit) directly across the RTD. Here, the wire resistance does not affect the voltage drop across the RTD, because the excitation current is constant, and the resistance of the measurement wires has no effect on accuracy since it is negligible compared to the high input impedance of the voltmeter.

One pair of wires in a two-wire connection supplies both the excitation current and measures the RTD output, so the voltage drop due to the resistance of the excitation current lead wires can add a significant error to the output signal. On the other hand, the four-wire or Kelvin connection separates the excitation and measurement wires. Here, the excitation current flows only through the RTD and excitation wires and the voltage drop does not appear in the measured variable. This makes the measurement more
accurate, even in remote installations.

A compromise between the 2-wire and the 4-wire connection is the most widely used 3-wire connection.  One end of the RTD connects to one wire, and the other end connects to two wires: a wire for power and a wire for signal.

Another frequently used circuit that avoids the effect of lead-wire resistance on the accuracy of a measurement is a four-resistor Wheatstone bridge. Two lead wires apply excitation power to the bridge. The bridge output is connected to a voltmeter, an operational or instrumentation amplifier, or a high-resolution analog to a digital converter. Ideally, the three bridge resistors should have a zero temperature coefficient, so only the RTD resistance depends on temperature. The output voltage depends only on the bridge resistance unbalance — it is not affected by the resistance of the lead wires.

The Wheatstone Bridge circuit using an RTD is similar to the circuit used with thermistors. The values and temperature coefficients of resistors R1, R2, and R3 should be selected in accordance with the accuracy needed for the application.

When selecting the RTD resistance and the excitation current, maintain a balance between the resolution and response time. It might be tempting to select a low resistance RTD for faster response, but since V_out = (R_RTD)(I_ex), a lower resistance requires higher excitation current to maintain the same output voltage and high system resolution. Higher excitation current generates more heat and raises the sensor temperature above the temperature of the object being measured, which produces a significant error. As a rule, the excitation current should be kept as low as possible to reduce the self-heating error. Typically, self-heating errors can be kept below 0.5°C, which is considered acceptable.

Another measurement error, thermal shunting, might creep in when measuring the temperature of a small object. Due to its relatively large size, the RTD might act as a heat sink and alter the temperature of an object that is similar in size or smaller than the sensor itself.

Maintenance
Interchangeability is critical when replacing worn or failed RTD elements. Knowing the allowed variance of readings between two sensors, allows equipment maintenance without recalibration. The American Society for Testing and Materials (ASTM), American Scientific Apparatus Manufacturers Association (SAMA), International Electrotechnical Commission (IEC), and Japanese Standard (JIS) have all developed several standards for platinum RTD elements. These standards guarantee element interchangeability when used within the specified temperature range.

For example, European standards IEC751 and DIN 43760 contain identical accuracy and tolerance parameters. They are collectively referred to as DIN IEC 60751, or just IEC751 and specify resistance values at various temperatures for platinum RTD sensors. Standard compliant devices have a resistance of 100.00Ω at 0°C and a temperature coefficient of resistance of 0.00385 Ω/Ω/°C from 0°C to 100°C. Specified overall temperature range and tolerance depend on the class.

The most accurate, DIN Class A has a ±0.06% tolerance at 0°C that spreads to ±0.24% at -200°C and ±0.46% at 650°C. DIN Class B has a ±0.12% tolerance at 0°C with a wider spread of ±0.56% at -200°C and ±1.34% at 850°C. Manufacturers frequently express tolerance specifications in Ω or in °C instead of %.

Most American and European manufacturers produce Class A and B IEC751-compatible elements. In addition, some RTD manufacturers offer less accurate class C and D elements with a ±0.2% and a ±0.5% tolerance at 0°C, respectively.

Most Japanese and some American manufacturers use Japanese Standard JIS C 1604. This standard specifies the same base resistance of 100.00Ω at 0°C, a temperature range of –200 to +650°C, and tolerances as IEC751, except the temperature coefficient of resistance is 0.003916 Ω/Ω/°C.  JIS C 1604 J and K standard tolerance classes correspond to the A and B classes of the European standard.

The environment
Many industrial and some lab applications require the RTD element to be protected from the environment. Moisture, corrosive environments, mechanical impact, and vibration can quickly degrade the sensor if not taken into consideration. Selecting the type of RTD or RTD probe appropriate for the application and compatible with the environment is the key for reliable service.

RTDs for industrial applications are typically built into a probe with a stainless steel or Inconel sheath protecting the sensing element from the environment and mechanical impact. This protection allows the measuring end of the probe to be inserted directly into the measurement area. Termination wires or a connector installed on the opposite end connect probes to the measuring instrument.

Low temperature range, thin-film RTD probes are typically safe to use in the -40°C to +200°C range.  Low temperature range wire-wound probes extend further into the -200°C to +200°C range. Inside the probe, the RTD element uses silver-plated copper lead wires with plastic insulation, such as Teflon® rated at 260 °C. By comparison, fiberglass insulation is usable up to 480 °C, while PVC insulated wires are limited to only 105 °C. Other considerations include:

• Welding — the preferred method of wire termination.
• Empty space inside the probe is typically packed with aluminum oxide powder that has good heat-transfer characteristics and acts as a shock and vibration absorber.
• To protect the element from moisture, the probe is sealed on the lead-wire side using epoxy or other potting compound.

High-temperature probes designed to work in the -200°C to +600°C range typically have internal lead wires made of nickel held in place with magnesium oxide insulators. Empty space inside the probe is typically packed with magnesium oxide powder, and the lead-wire end is sealed with epoxy. In both designs, the lead wires are connected in a 2-wire, 3-wire, or 4-wire configuration and are brought out with several options of insulated wires, terminated wires, and male and female connectors.

Surface, gas, and liquid measurements require different sensor configurations. Probes most suitable for liquid measurements are typically encased in a stainless steel sheath that has excellent corrosion resistance. Inconel sheaths provide superior protection against corrosion and oxidation at high temperatures. Outside diameters are typically 1/8 in., 3/16 in., ¼ in., and 3/8 in.

Surface sensors vary greatly depending on application and method of attachment. They can be touched, screwed, bolted, or glued to the measured surface. Gas and air measurements require free access of the gas to the RTD element to facilitate the heat transfer.

Certain precautions should be observed during installation and operation. RTDs, especially the wire-wound type, are susceptible to mechanical damage and should be installed with care and protected during use. In addition, minimizing mechanical and thermal stresses is essential for long service life. Electromagnetic interference can be a serious problem because the signal level is quite low. Use proper shielding and twisted-pair wires to keep electrical noise at an acceptable level. Increasing the gauge of lead wires minimizes their resistance, which is especially important for minimizing measurement error in two-wire RTD connections.

For more information:

Omega Engineering, Inc.
http://www.omega.com

Thermo Sensors Corp. (TSC)
http://www.thermosensors.com

U.S. Sensor Corp.
http://www.ussensor.com

JUMO Process Control, Inc.
http://www.jumoplus.com

Watlow Electric Manufacturing Co.
http://www.watlow.com

Pyromation, Inc.
http://www.pyromation.com

Minco
http://www.minco.com