The problem
A thermocouple does not measure absolute temperature: it generates a voltage that depends on the temperature difference between the measuring junction and the reference junction, known as the cold junction. Standardized tables and polynomials assume that the reference junction is at zero degrees Celsius. In practice, the ends of the thermocouple reach an instrument that is at ambient temperature. Without compensation, the reading is wrong by an amount tied to the temperature of that connection.
The principle of compensation
The temperature of the reference junction is measured, the voltage that a thermocouple of that type would generate between that temperature and zero degrees is calculated, and it is added to the measured voltage. That corrected voltage is then used to enter the table or polynomial. The important point is that voltages are added, not temperatures, because the relationship between the two is not linear.
Methods
- Ice bath. The reference junction is physically held at zero degrees with ice and water. It is very accurate and is used in the laboratory, but it is impractical in a plant.
- Electronic compensation. A temperature sensor located on the instrument’s terminal block measures the temperature of the reference junction, and the device makes the correction. This is the norm in indicators, transmitters and recorders.
- Extension or compensating cable. It carries the reference junction to a more stable or more convenient location. It must be of the corresponding type and the polarity must be respected.
An example
A thermocouple measures a hot process and runs to an indicator whose terminals are at room temperature. The indicator measures the temperature of the terminals, calculates the equivalent voltage and adds it. If the terminals had an internal temperature gradient, or were heated by a nearby source, the compensation sensor would not see the actual temperature of the connections and an error would appear.
Sources of error
- Terminals with a thermal gradient, caused by air currents or heat from components.
- Use of ordinary copper instead of extension cable along part of the run.
- An extension cable type that does not match the thermocouple.
- An instrument configured for a thermocouple type different from the actual one.
How to interpret the result
A reading that changes when the enclosure is opened, or with ambient temperature, points to a compensation problem. Short-circuiting the instrument input should give the temperature of its own terminals, which serves as a quick check.