Linearity is defined as the maximum deviation of the calibration curve (an average of the upscale and downscale readings) from a straight line so positioned as to minimize the maximum deviation.
Platinum resistance elements have a nearly linear output while nickel and nickel-iron (Balco) sensors are quite curved. Copper elements are also nearly linear over their narrow temperature range.
Stability is the relationship of a sensor’s original resistance curve to its curve after being in service. Drift rates published by a manufacturer must be assumed to be applicable to high purity laboratory environment probes. The published drift rates of 0.0°C are to be considered general and not necessarily quantitative.
Several parameters affect stability in a platinum sensor used in industrial processes. Thermal and mechanical treatment cause physical changes in the crystalline structure of the platinum causing different resistances at different temperatures. Chemical reactions involving platinum and impurities as well as migration of internal materials can affect a sensor output. A shunting effect due to insulation resistance deterioration is another influencing occurrence.
The drift caused by these conditions is not normally catastrophic except in rare instances. Attempts to establish a statement of stability in industrial applications would result in an ambiguous approximation at best.
Since an RTD measures temperature by passing a current through a resistor (the RTD), the error known as self-heating occurs.
Primarily the sensor’s mass, its internal construction, the measurement current and to a large degree environmental conditions determine the magnitude of this error. Normally a very small current, usually 1-5 milliamps is used in the excitation circuit to minimize this joule heating of the sensor. Thermo Sensors’ internal construction technique maximizes heat transfer quality to further reduce the effect.
An installation condition requiring large mass hardware such as thermowells or protective tubes coupled with an environment of still or slow moving air is going to experience a great deal more self-heating than the next example. a small diameter (.250″ O.D.) direct immersion probe mounted in an environment of flowing water (min. 3 ft./sec) could totally dissipate the error.
Fortunately if a small measuring current (1-2 ma) is used, selfheating errors will be well within acceptable levels for industrial applications.
To approximate the amount of error; consider that normally the dissipation constant will be of the magnitude of 20-100 mw/0°C, and use the following formula.
|Self-heating error =||Power|
Measurement current – 2 ma
Resistance of sensor – 140 ohms
dissipation constant – 50 mw/0°C
Power = 12R
= (.002)2 (140) = 0.56 MW
|Error =||.56 mw||.011°C|
Time constants are values used to indicate the time it takes a sensor to read 63.2% of a step change in temperature. This test is conducted in water flowing at 3 ft/sec or 20 ft/sec in air. Typically this measurement is made by plunging a sensor at room temperature into a bath at 80°C and noting the time required to reach 63.2% of that step change. Generally speaking, it takes approximately five (5) time constants before 100% of the step change is realized.
Several variables affect the response time of sensors. Diameter of the sheath, material of the sheath and internal construction for different temperature ranges are the most variable. It is possible, however, to approximate the time constant for a particular group of sensors based on diameter and assuming the sheath material is a 300 series stainless steel.
|These approximations are:|
Note: elements capable of a lower range of -250°C (to +600°C) have similar time constants.
These time constants should serve only as a general approximation for direct immersion sensors. Sensors installed in thermowells, protection tubes or that are mounted in conditions allowing appreciable stem losses are not subject to even these general constants.
In the rare instance where the response time absolute needs to be known; response time testing must be conducted to provide a time constant.
To prevent an unacceptable shunting effect between the sensing element and the probe sheath, care must be taken to assure good insulation quality.
In all sensors and particularly those in industrial service, high temperature operation, contamination and moisture absorption are potential problems.
To eliminate the effects of these occurrences, Thermo Sensors adheres to stringent manufacturing procedures. Reliatemp’s Insulation Resistance will always be > 2000 megaHMS at or below 100°C.
By definition repeatability of a sensor is the relationship of the original resistance at 0°C and any different resistance at 0°C after being subjected to the following test.
The sensor shall be brought slowly to the upper limits of its temperature range and then exposed to air at room temperature. It shall then be brought slowly to its lower limit, and exposed to air at room temperature.
This procedure is repeated ten times. The resistance of 0°C is then measured and the difference from the pre-testing resistance is 0°C is noted.
For a typical platinum probe, the resistance should not change more than 0.3°C for a 0.12% sensor or 0.15°C for a 0.06% sensor. The 0.12% and 0.06% are original resistance tolerances at 0°C of the element.