The diagram above illustrates the error introduced during calibration prior to the validation studies. This time the thermocouples are placed in a reference bath, and the vessel is not running. Since the vessel is not active, the temperature on the inside (T2) and outside (T1) of the vessel wall are relatively uniform, therefore the error due to the junction box is negligible . The user assumes all the errors have been compensated for in the calibration procedure. Next the user goes ahead with the validation studies. Since the vessel is now active the temperatures on the vessel wall have significantly changed. The temperature on the inside of the vessel wall (T2) is now calculated at 110°C and the outside of the vessel wall (T1) is 30°C . If we now calculate the error, we can see that the error introduced by the gradient across the juction box is approximately 4°C .
Error introduced using connectors or junction boxes can be mathematically calculated. In this example quality thermocouple wire (SW), has been connected from the measuring system (MS) to the junction box on the outside of the vessel and then from the inside of the vessel to the process. Although the junction box utilises matching material on the connectors the purity of the metals (SC) does not match the thermocouple (SW). The error introduced is equal to the difference in the output of the materials, multiplied by the temperature gradient across the junction box.
After completion of the studies the User performs a Post calibration to verify that the thermocouples have not varied since initial calibration. Since the vessel is now not running the gradient across the junction box is eliminated and the calibration matches that of the initial calibration and the study is assumed to be successful.
Since these errors are random and dependent on the composition of the junction and the gradient temperatures across them, they should be avoided. The only way to ensure the elimination of these possible errors is by utilising one continuous length of thermocouple wire from the measuring system to the inside of the vessel.
The sensor and circuit errors discussed can seriously impact the test data validity, however with good practices you can eliminate these errors and achieve accurate results.