You are probably wondering why accuracy was not the first topic covered, because RTD’s are generally known for their high degree of accuracy and it is typically one of the first specifications laid out. Well, the subject is not quite that simple, and it requires a bit of discussion. First, we must establish the difference between accuracy, precision, and repeatability. In the case of temperature, accuracy is commonly defined as how closely the sensor indicates the true temperature being measured, or in a more practical sense, how closely the resistance of the RTD matches the tabulated or calculated resistance of that type RTD at that given temperature.
Precision, on the other hand, is not concerned with how well the RTD’s resistance matches the resistance from a look-up table, but rather with how well it matches the resistance of other RTD’s subjected to that temperature. Precision generally refers to a group of sensors, and if the group has good precision at several temperatures, we can also say that they are well matched. This is important when interchangeability is a concern, as well as in the measurement of temperature gradients. Repeatability can best be described as the sensor’s ability to reproduce its previous readings at a given temperature.
Here’s an example. An ice point reading is done with an RTD that is then used to take readings at 100°C, 150°C, 37°C, and again at 0°C.
A comparison of the first and last ice point readings will give you an indication of the sensor’s repeatability under those conditions. A note of caution, however: an RTD’s repeatability is very application-dependent. So when you get right down to it, accuracy without repeatability is worthless. If you start with a sensor that is ±0.03°C at 0°C but is found to have repeatability only around ± 0. 5°C, what you have is a sensor whose readings are far less reliable than a standard-accuracy probe with good repeatability. A high-accuracy RTD installed in a field application also does not ensure that you will be getting a highly accurate signal back at the control room.
Most 4-20 mA transmitters and many display units and controllers have adjustable zero and span controls that if improperly adjusted will destroy the high accuracy of the RTD signal.
The best solution for applications of this type is to have both the RTD and the transmitter, or display, or whatever, calibrated as a unit by a certified calibration laboratory.
Fortunately, the requirements for this degree of accuracy best solution for applications of this type are few and far between. For more on this subject see, Accuracy Standards.
Our final two parameters are application dependent and vary from the specification of a bare resistance element to a large industrial assembly with thermowells, connection heads, and possibly field -mounted transmitters. We will discuss only the most basic areas: physical dimensions and size restrictions, and material compatibility.