- By Steve Radecky
- InTech
Summary
DAQ devices can accurately measure thermocouples in a Raspberry Pi environment. Here’s how.
Leading-edge Internet of Things (IoT) technology and advanced analytics are increasingly being used for process optimization and improved efficiency of industrial machinery because they enable predictive maintenance. The data being analyzed for this form of asset management often includes temperature measurements. And the compute power to perform those analyses is increasingly being provided by IoT devices based on Raspberry Pi.
Raspberry Pi is a series of small single-board computers developed in the U.K. by the Raspberry Pi Foundation in association with Broadcom. The Raspberry Pi project originally focused on teaching basic computer science in schools and in developing countries, but the growing base of Raspberry Pi means the computer boards are increasingly finding their way into industrial automation applications—particularly as IIoT devices. The use of open-source C/C++ and Python lets users develop applications on Linux.
Although thermocouples are a popular way to measure temperature, designing and building data acquisition (DAQ) devices that accurately measure thermocouples in a Raspberry Pi environment is challenging. This article explains the difficulties in making accurate thermocouple measurements, how the MCC 134 DAQ HAT accomplishes it, and how MCC 134 is being used in IIoT devices for machine health monitoring.
How thermocouples work
A thermocouple is a sensor used to measure temperature. It works by converting thermal gradients into electrical potential difference—a phenomenon known as the Seebeck effect. A thermocouple is made of two wires with dissimilar metals joined together at one end, creating a junction. Because two dissimilar metal wires create different electric potentials over a temperature gradient, a voltage that can be measured is induced in the circuit.
Different thermocouple types have different combinations of metal in the wires and are used to measure different temperature ranges. For example, J type thermocouples are made with iron and constantan (copper-nickel alloy) and are suited for measurements in the –210°C to 1200°C range, while T type thermocouples are made with copper and constantan and are suited for measurements in the –270°C to 400°C range.
The thermal gradient mentioned above is referred to as the temperature difference between the two junctions: the measurement, or hot junction, at the point of interest and the reference, or cold junction, at the measurement device connector block (figure 1). Note that the hot junction refers to the measurement junction and not its temperature; this junction might be hotter or colder than the reference or cold junction temperature.
Thermocouples produce a voltage relative to the temperature gradient—the difference between the hot and cold junction. The only way to determine the absolute temperature of the hot junction is to know the absolute temperature of the cold junction.
While older systems relied on ice baths to implement a known cold junction reference, modern thermocouple measurement devices use a sensor or multiple sensors to measure the terminal block (cold junction) where thermocouples connect to the measurement device.
Sources of thermocouple errors
Thermocouple measurement error comes from many sources, including noise, linearity, and offset error; the thermocouple itself; and measurement of the reference or cold junction temperature. In modern 24-bit measurement devices, high-accuracy ADCs are used, and design practices are implemented to minimize noise, linearity, and offset errors.
Accurately measuring the cold junction, where the thermocouples connect to the device, can be a challenge. In more expensive instruments like the DT MEASURpoint products, an isothermal metal plate is employed to keep the cold junction consistent and easy to measure with good accuracy.
In lower-cost devices, isothermal metal blocks are cost prohibitive, and without an isothermal block it is not possible to measure the temperature at the exact point of contact between the thermocouple and the copper connector. This fact makes the cold junction temperature measurement vulnerable to temporary error driven by quickly changing temperatures or power conditions near the cold junction.
Design challenges
To better understand the design challenges of the MCC 134, we can compare it to the design of MCC’s popular E-TC—a high-accuracy, Ethernet-connected thermocouple measurement device. The cold junction temperature of the E-TC is measured by Analog Devices’ ADT7310 IC temperature sensor.
The IC sensor design works well in the MCC E-TC because the measurement environment is controlled and consistent. The outer plastic case controls the airflow, and the electronic components and processors operate at a constant load. In the controlled environment of the E-TC, the IC sensor does an excellent job of measuring the cold junction temperature accurately.
However, when the MCC 134 was first designed with an IC sensor to measure the cold junction temperature, the accuracy was insufficient. Because the IC sensor could not be placed close enough to the connector block, large and uncontrolled temperature gradients caused by the Raspberry Pi and the external environment led to poor measurement repeatability.
Although this added complexity to the design, the thermistors more accurately tracked the temperature changes of the cold junction, even during changes in processor load and environmental temperature. This design yields excellent results that are far less susceptible to the uncontrolled Raspberry Pi environment.
The MCC 134 should achieve results within the maximum thermocouple accuracy specifications when operating within the documented environmental conditions. Because certain factors still affect accuracy, users can improve measurement results by reducing quick changes in temperature gradients across the MCC 134 and following other best practices.
MCC 134 in Action: Thinaer Health Usage Monitoring System
Thinaer systems use Raspberry Pi nodes that communicate with smart sensors via Bluetooth Low Energy. These smart sensors, however, do not have the high-accuracy temperature or high-speed vibration data needed for better analysis.
The solution for Thinaer was to use the MCC 134 thermocouple measurement HAT (see box) to measure temperature (as well as the MCC 172 IEPE measurement HAT to measure vibration) and to collect the data needed to create accurate measurements, analyses, and strategy.
The stackable DAQ HATs also allow Thinaer to scale without having to change its platform or do any internal hardware development or assembly. The system was programmed using provided C and Python libraries for continuous, multi-HAT acquisition of data.
Using MCC technology saved Thinaer both time and labor. The MCC DAQ HATs easily fit into the existing system enclosure and the off-the-shelf design saved Thinaer from having to develop a custom, in-house solution.
Accurately measuring
Thermocouples provide a low-cost and flexible way to measure temperature, but measuring thermocouples accurately is difficult. Through innovative design and extensive testing, MCC overcame the challenge of measuring thermocouples accurately in the uncontrolled Raspberry Pi environment. The MCC 134 DAQ HAT provides the ability to use standard thermocouples with the fast growing, low-cost computing platform.
Raspberry Pi Thermocouple Measurement HAT
Up to eight MCC HATs can be stacked onto one Raspberry Pi. With the already available MCC 118, eight channel voltage measurement HAT, and the MCC 152 voltage output and digital I/O HAT, users can configure multifunction, Pi-based solutions with analog input, output, and digital I/O.
Reader Feedback
We want to hear from you! Please send us your comments and questions about this topic to InTechmagazine@isa.org.