This engineering brief demonstrates the operation of a thermocouple using the system shown in Figure 1. Specifically, it allows the student to interact with the hot junction, each of the cold junction terminals, and the cold junction temperature sensor. The article begins with the hands-on experiment. It also includes background information for those readers who need to brush up on their thermocouple vocabulary.
This analysis is based on:
- MIKROE-5767 featuring the Analog Devices MAX31855KASA+T thermocouple converter IC.
- Omega SA3-K-SRTC Thermocouple
- Click-compatible platform of your choice. In this case an Arduino shield was used on top of an Arduino UNO R4 Minima.
Key Takeaways
- Cold junction temperature compensation is only accurate if the cold junction and associated temperature sensor are at the same temperature.
- Some devices such as the MAX31855 include the cold junction temperature sensor on its silicon die. Others such as the LTC2983 allow external cold junction sensing devices.
Figure 1: Image of the MIKROE Thermo click installed on an Arduino adapter.
Physical Observations
A jet of hot air may be used to explore the three sensitive areas of the Figure 1 thermocouple system. In this case I set my workbench Hakko 851-2 air station to a low setting.
Figure 2 shows the results when each section is subjected to a short burst of air.
-
Hot junction heated: At 60 s, the air is directed toward the thermocouple measurement junction. The measured temperature jumps to 118 °F.
-
Cold junction heated: At 160 s, the air is directed toward the Alumel side of the socket near the PCB connection. The measured temperature drops to 50 °F. This is a temperature-gradient error. In this case, the MAX31855 is performing the temperature compensation without recognizing that the cold junction temperature has significantly increased.
-
MAX31855 heated: At 360 s, the air is directed directly to the MAX31855. The reported temperature jumps to 113 °F. At the same time, the internally sensed temperature jumps to 123 °F. Again, this is another temperature-gradient error. In this case the cold junction temperature remains steady, but the internally sensed temperature has detected an increase in temperature.
Figure 2: Thermocouple Response to Localized Heating of the Hot Junction, Cold Junction, and MAX31855.
Thermal Compensation Gotcha
We are making an assumption that the cold junction and the associated MAX31855’s internal silicon thermometer are at the same temperature. The system accuracy depends on this assumption.
Any temperature difference between the cold junction and the silicon results in a measurement error according to the equation included in the background section. Recognize that the MAX31855 cannot correct for a thermal gradient. This is noted in the MAX31855 datasheet:
Optimal performance from the device is achieved when the thermocouple cold junction and the device are at the same temperature. Avoid placing heat-generating devices or components near the MAX31855 because this could produce cold-junction-related errors.
Improve Cold-Junction Accuracy with an External Temperature Sensor
For improved accuracy, consider the ADI LTC2983. This IC has provisions for an external sensor. For example, a suitably packaged TO-92 diode, RTD, or thermistor could be placed into the clamp shown in Figure 3. The key design solution is to make the measurement system track the cold junction temperature as closely as possible.
Figure 3: Image of an advanced Omega thermocouple socket with provisions for a TO-92 packaged temperature sensor.
Thermocouple Design Implications
The design implications are clear. The temperature gradient between the Omega socket and the MAX31855 must be minimized. A temperature difference will yield a measurement error.
Let’s look at a poor design. Let’s assume:
- The socket is physically mounted to allow a thermocouple connection outside of the equipment enclosure.
- An elevated temperature inside the equipment enclosure and a colder exterior.
Given these conditions we can expect a thermal gradient between the socket and the IC. While there are many design considerations, we should expect some measurement error. In this case the cold junction will be colder than the IC. Based on the previous observation, we would expect the system to measure hotter on the coldest days.
The solution is to minimize the temperature difference between the cold junction and the junction sensor. We could:
- Move the socket deeper inside the enclosure.
- Move to a system with an externally compensated cold junction such as the previously mentioned LTC2983 coupled with an isothermal thermocouple socket. This allows the sensing device to be mounted as close as possible to the cold junction region.
Thermocouple Background
The thermocouple is a misunderstood technology. People often assume the temperature is measured directly as a voltage drop across two dissimilar metals joined at the probe’s tip. This explanation is incomplete. Instead of visualizing a tiny voltage source located at the welded joint, we need to think in terms of a voltage developed across a thermal gradient.
The actual operation is summarized by NIST:
Because thermocouples are sensitive only to differences between two temperatures, one circuit component must be kept at a well-characterized reference point (for example, the ice melting point) while the other component is at the temperature to be measured.
The two relevant junction points are known as the hot and cold junctions:
- Hot Junction: The measurement junction is also known as the hot junction. In this case, the Omega thermocouple has an adhesive backing used to affix the sensor to the item being measured.
- Cold Junction: The reference junction is also known as the cold junction. It gets its name from the physics textbooks where it is often implemented as an ice bath.
Always remember that the thermocouple system performs a relative measurement. We can only determine the temperature at the hot junction if we know the temperature of the cold junction.
Tech Tip: Note that the Chromel and Alumel alloys extend from the wire to the cable connector. Close inspection shows that the blades are marked AL and CH. The cold junction region is not an attribute of the cable. Instead, the region is associated with the matching cable socket. Specifically, the PCB electrical connections taps the Chromel and Alumel junctions.
Temperature Compensation
We do not need an ice bath for the cold junction. Instead, the MAX31855 performs the temperature compensation using a semiconductor sensor incorporated into its silicon die as shown in Figure 4.
On its own, this internal thermometer sensor is described as a 12-bit temperature system with a measurement range of -55 to 125 °C. The critical observation is that the sensor measures the die temperature of the MAX31855.
The MAX31855 datasheet states:
For a K-type thermocouple, the voltage changes by about 41µV/°C, which approximates the thermocouple characteristic with the following linear equation:
VOUT = (41.276µV/°C) x (TR - TAMB)
where VOUT is the thermocouple output voltage (µV), TR is the temperature of the remote thermocouple junction (°C), and TAMB is the temperature of the device (°C).
Note that the MAX31855 assumes a linear relationship between temperature and voltage.
Figure 4: Block diagram of the MAX31855.
About this Author
Aaron Dahlen, LCDR USCG (Ret.), is a Senior Applications Engineer at DigiKey in Thief River Falls. His background in electronics and industrial automation was shaped by a 27-year military career as both technician and engineer, followed by over a decade of teaching.
Dahlen holds an MSEE from Minnesota State University, Mankato. He has taught in an ABET-accredited electrical engineering program, served as coordinator of an electronic engineering technology program, and instructed military technicians in component-level repair.
Today, he has returned to his home in northern Minnesota, completing a decades-long journey that began with a search for capacitors. Read his story here.