How to Select Temperature Measurement Components
Some applications use a modular approach to temperature sensing while others must be built from scratch. Fortunately, there is no shortage of temperature sensor options to solve even the most complex problems.
Most people who purchase personal computers buy fully configured and assembled units. This is especially true for computer buyers who plan to use them for everyday tasks like word processing or surfing the web. Some buyers, however, prefer to design their own systems, selecting individual boards, cases, drives, power supplies and so on. Technically demanding applications like gaming or digital video editing demand a computer built with certain requirements in mind.
Likewise, an engineer working on a new temperature measurement application has a similar choice. It is possible to buy an integrated instrument with everything in one package. Or, the application may call for a different type of solution that must be assembled from individual components. System selection and complexity depend upon what is needed to deliver the necessary performance.
This article looks at how temperature-measurement applications can be analyzed to help users select the appropriate products for the process application. For brevity’s sake, it will concentrate on permanently mounted temperature sensors capable of measuring liquids or gases rather than optical systems that use infrared approaches.
Start with the Sensor
An electronic temperature sensor creates an analog signal in response to the temperature it detects. Resistance temperature detectors (RTD) and thermistors change their resistance in proportion to changes in temperature. Thermocouples (TC) produces an electrical voltage that changes in proportion to the difference in temperature detected at two different points. A thermocouple determines a temperature difference, so it is necessary to have a known reference. As a result, a thermocouple application needs two sensors to determine the actual process temperature.
At some point, the raw analog signal from an RTD or a thermocouple has to be turned into a meaningful temperature value. This can be as simple as using a multi-tester to take an ohm reading. When using a PT100 RTD, a reading of 113.61 ohms corresponds to 95°F. This figure can be found on an RTD Chart showing the range of resistance readings and their temperatures. Similar charts are available for the different thermocouple types. This might be useful for a science fair project or perhaps in a troubleshooting situation, but a real-world application calls for something more practical.
FIGURE 1. Temperature transmitters encompass many functional capabilities and form factors whose use depends on the application.
Simple Displays and Controllers
The most basic signal-processing solution is a single panel meter temperature display. These are typically off-the-shelf units designed to receive a signal from a single sensor. Most have the ability to work with a variety of RTD and thermocouple types, depending on the specific terminal arrangement or software configuration step. Those able to work with thermocouples will have an internal sensor — usually an RTD or thermistor — to provide the reference temperature.
Typically, a panel meter provides a local visual display in degrees Fahrenheit or Celsius. It may have the ability to send a corresponding signal — a 4 to 20 mA value, for instance — to a programmable logic controller (PLC) or larger automation system. Panel meters work well for applications where the distance to the sensor is relatively short and no higher degree of functionality is necessary.
More functionality can be provided by a small controller or PLC, often with a human-machine interface (HMI) screen. These usually have an input card of some type to interface with a single or multiple temperature sensors. They can be configured to display or make calculations from the temperature readings. The way in which these units display or record the readings depends on the setup and application.
For both these approaches, the sensor is connected directly to the host device. With an RTD, this usually requires three or four wires; for a thermocouple, it requires two. If extension cabling is required, it must be the right type for the sensor.
FIGURE 2. Basic sensor/transmitter combinations include a thermowell, sensor and transmitter designed to ease installation.
Moving to a higher degree of sophistication requires a temperature transmitter (figure 1). A transmitter takes the raw signal from a sensor and converts it to a 4 to 20 mA signal with some other type. Depending on the type of transmitters used, the signal might be converted to HART; or to a fieldbus signal such as Foundation Fieldbus or Profibus PA; or to a wireless signal such as WirelessHART.
Transmitters incorporate three subsystems:
- Input subsystems convert the analog sensor measurement into a digital signal for internal processing.
- Signal-conditioning subsystems take the digital signal and perform various functions to produce a representation of the temperature measurement along with any diagnostic routines.
- Output subsystems convert the preliminary digital signal to the desired output format, which may be analog or digital.
FIGURE 3. Sensors configurations run the gamut from rigid sheaths to flexible to spring-loaded designs. Many designs are available in to satisfy virtually any industrial process application requirements.
The signal-conditioning functions have different capabilities depending on the sophistication of the transmitter. The default capabilities available from virtually any transmitter will include basic functions such as providing a reference for thermocouples, converting to appropriate engineering units, damping and ranging. Smart transmitters can add diagnostic functions to monitor the sensor itself, look for problems that might be developing and support redundant sensor arrangements.
Prepackaged Temperature Instruments
Selecting a preconfigured sensor and transmitter assembly is like buying a new laptop. There are modules designed to fit together easily while still offering a range of capabilities to match applications (figure 2). The specifier generally will have to check boxes for various options such as:
- Thermowell length.
- Sensor type, to match the measurement range.
- Transmitter type, to match the desired signal output format.
- Housing configuration.
The assemblies usually consist of several components:
- A thermowell designed to mount to the process vessel or piping.
- A sensor with the correct threads and sheath length to mate with the thermowell.
- A transmitter in an enclosure mounting on the other end of the sensor assembly.
The lead wires from the sensor extend into the enclosure, where they mount on the transmitter terminals.
Having a transmitter at the head keeps the leads short to minimize the possibility of the signal being corrupted by electrical interference. It also makes it possible to change the type of sensor if this proves necessary. If a local display is needed, most suppliers have an option to add it to the configuration.
FIGURE 4. Transmitters have flexibility on inputs but normally only one output. This example is 4-20 mA with HART.
The connection to the host system can use conventional instrumentation cabling because the transmitter has already converted it to a more robust signal format. There is no need for special temperature sensor cabling.
Prepackaged temperature instruments are convenient and use the same types of components as custom configurations. Accommodating more complex applications such as those designed for hygienic or hazardous environments usually can be accomplished with an appropriate transmitter housing. Setting up a bill of materials for each assembly helps ensures the same components are available when it is necessary to replace one of the items.
FIGURE 5. Adding a local display usually requires a larger housing.
More complex applications may require different types of components to solve a measurement challenge. Fortunately, there is no shortage of options when looking for the ideal items.
Sensor selection can take advantage of all the different types of RTDs and thermocouples (figure 3), which are offered in sizes and shapes to match different thermowell configurations. They can be rigid sheaths, flexible, spring loaded and so on.
A transmitter can be selected from a range of transmitter types and packages. Different transmitter types offer capabilities from basic signal processing to sophisticated electronics that can handle redundant sensors, perform diagnostic routines, validate internal measurements, provide sensor drift alarms, retain calibration information, log data internally and perform other functions. Some of these functions may be performed by the host system, but the ability to do them at the transmitter relieves the larger automation system. This frees up processing power to allow the host system to perform more pressing tasks.
The given plant environment is far more likely to influence sensor selection than the communication-protocol-to- host system. Therefore, while most transmitters can handle different sensors such as RTDs and various thermocouple types, they have only one output capability, such as 4 to 20 mA, with HART, Foundation Fieldbus or WirelessHART (figure 4).
The transmitter form factor reflects the application needs based on where the transmitter will be located and its degree of sophistication. The three basic approaches are head mount, field mount and rail mount.
Head Mount. This approach places the transmitter directly adjacent to the sensor at the point of insertion. Mechanical approaches vary, but the housing is attached to the thermowell or sensor. The sensor uses a DIN A or DIN B round-puck configuration, which has a hole in the center for the sensor leads. Connection terminals then are arranged near the perimeter of the puck so the leads spread like spokes on a wheel. DIN A is the larger of the two (67 millimeter diameter versus 45 mm); therefore, it tends to be used for transmitters with more sophisticated capabilities. If a local display is desired, one can be added at the head although a larger housing may be necessary (figure 5).
FIGURE 6. Rail-mount transmitters can be designed for external mounting or placement inside a control panel enclosure.
As mentioned earlier, head mounting keeps the sensor leads as short as possible, minimizing the potential for picking up interference. There can be situations, however, where it is not practical to mount the transmitter at the head. This is frequently due to accessibility constraints. Extended lead-wire options place the transmitter at a specified distance away from the thermowell. In most cases, this is tolerable and solves the access problem.
Field Mount. This approach is a variation on the conventional head mount. If the plant environment is particularly hostile due to dust, moisture, corrosive fumes or other hazards, the transmitter can be enclosed in a dual-compartment housing. The transmitter itself with the sensor lead connections is isolated in one half of the housing while the terminal block with the host connections is in the other. If maintenance personnel need to check the terminals, this can be done without exposing the transmitter itself. This maximizes protection from the physical hazards and electrical interference.
Rail Mount. In environments where a head-mount or field-mount approach is not necessary, transmitters also are available in DIN rail-mountable enclosures (figure 6). These are suitable for insertion into a control panel or mounting directly on a larger piece of machinery. Many plant environments do not pose the hostile challenges of refineries and chemical processors, so sealed and explosionproof housings would be overkill. If a rail- or wall-mountable transmitter is available with an appropriate IP rating, it can be placed in the open when no other enclosure is convenient.
The head-, field- and rail-mount devices are the primary categories, there are other transmitter types. One example is the multi-channel version able to capture data from up to eight individual transmitters. These typically send the individual readings sequentially via one signal, and the host system sorts out the individual values.
Adding WirelessHART capability allows users to avoid cabling costs. A single internal power module can drive both the sensor transmitter and communication transmitter. Depending on the frequency of updates, these can operate for eight to 10 years without replacement.
In conclusion, the effectiveness of a self-designed temperature setup depends on understanding the function of each component and its position in the chain from sensor to host system. Accuracy depends on the sensor itself, but it also on the transmitter, lead wires and even the skill of the individuals making the terminations. The range of components available from manufacturers can cover virtually any imaginable application, delivering whatever performance is required.