In most cases, the selection of temperature sensors should consider the following aspects:

(1) Whether the temperature of the tested object needs to be recorded, alerted and automatically controlled, and whether it needs to be measured and transmitted remotely.

(2) The size and accuracy requirements of temperature measurement range.

(3) Whether the temperature measuring element size is appropriate.

(4) When the temperature of the measured object changes with time, whether the hysteresis of the temperature measuring element can meet the requirements of temperature measurement.

(5) Whether the environmental conditions of the object under test are harmful to the temperature measuring element.

(6) How about the price and convenience?

The choice of temperature sensor is mainly based on the measurement range. Platinum resistance sensors can be used when the measurement range is expected to be within the total range. Narrow ranges generally require that the sensor must have a fairly high base resistance in order to obtain a large enough resistance change. Thermistors provide enough resistance variation to make these sensing elements very suitable for narrow measurement ranges. Thermocouple is more suitable if the measurement range is quite large. It is best to include freezing points in this range, as thermocouple meters are based on this temperature. The linearity of the sensor within the known range may also be an additional condition for selecting the sensor.

The response time is usually expressed as a time constant, which is another basic basis for selecting a sensor. The time constant is less important when monitoring the temperature in the tank. However, when the temperature in the vibrator must be measured during use, the time constant becomes the determining factor for selecting the sensor. The time constants of beaded thermistors and armored outcrop thermocouples are relatively small, while those of immersion probes, especially those with protective sleeves, are relatively large.

The measurement of dynamic temperature is quite complicated. Only by repeated testing and simulating the conditions that often occur in the use of the sensor as close as possible can a reasonable approximation of the dynamic performance of the sensor be obtained.