A variety of accuracy-relevant information describes the accuracy, which itself does not exist in the way we use it in everyday language. About 68.27% of the devices used by the user comply with the typical value of "accuracy".
Paradoxically, a pressure sensor with 0.5% error at maximum non-linearity after limit settings is as accurate as a pressure sensor with 0.1% after non-linearity after minimum settings.
The following describes properties that can lead to deviations in the measurement data. It should be noted that their use is not standardized. This means, for example, that three identical specifications for non-linearity can describe three completely different accuracies.
The most frequently mentioned specification of accuracy is non-linearity. This defines the most extreme deviation between the characteristic curve and the reference line. After setting the limit point, the reference straight line goes through the beginning and end of the characteristic curve in the case of non-linearity. In comparison, there is the BFSL (Best Fit Straight Line) method, in which the reference line is aligned in such a way that the maximum positive and maximum negative deviation is the same.
The non-linearity after limit setting offers the largest possible error in absolute terms, but it is the easiest for users to understand, even if the minimum value setting provides a more precise value in most cases.
The measurement deviation that comes closest to the true value is the measurement deviation. The reason for this is the ability to read the following immediately:
- Measurement deviation at the beginning and end of the measuring range
The largest difference between the actual and the ideal characteristic is the measurement deviation. The hysteresis defines the maximum variance in up and down gears.
Unfortunately, hysteresis and non-repeatability can neither be completely eliminated nor minimized.
Remove zero error
The zero point error can be read in the depressurized state. This is entered as an offset in its evaluation unit.
Remove span errors
In order to eliminate the span error, it is assumed that the pressure at the end of the measuring range has been approached exactly. This requires a pressure reference that is at least three times more accurate than the desired accuracy.
Non-linearity can be minimized by, for example, using support points in the downstream electronics to calculate errors. However, this technique requires an extremely precise standard.
The accuracy of sensors can be affected by a variety of factors. For this reason, a comprehensive analysis is recommended in advance. In this way, possible sources of error can be identified and eliminated in advance. This applies to the installation of the measuring device itself, but also in advance to the circumstances under which the accuracy in the data sheet was generated.
In conclusion, the effort you put into the accuracy of your gauges is well worth it. Not only does it increase security, but it also delivers precision and flawless continuity.
All accuracy specifications as described here have been defined at room temperature. However, if you want to measure at temperatures that are above or below room temperature, you also have to calculate the temperature error, also known as the temperature coefficient. A measuring device that offers sufficient accuracy at room temperature can show an error that is twice as large from as little as 10 K.
Long term drift
What happens between the production site of a measuring device and its arrival at the user? Even storage and transport can permanently manipulate the accuracy. The meter can also change over time due to possible magnetic interference or continuous vibrations at the measurement site. For this reason, manufacturers recommend annual calibration to record the so-called drift, the change in the device.