Basic calibrations of sensors combine a known environment tied to an output of that sensor. If a pressure sensor outputs a value of 2 psi when in a vacuum and 102 psi when placed inside a space with known pressure of 100 psi, the operator can calibrate the sensor to output 2 psi less than it currently does. Usually a sensors output is a linear affair between the stimuli and the reading it provides, and in these cases calibration is simple. By calibrating in this way the operator is introducing errors if they are off in any way.
If the testing situations are not accurately known any readings after calibration will have the same discrepancies. The use of a “golden unit” or a standardized unit of measurement to calibrate all sensors against can vastly reduce this type of error.
The calibration procedure should produce consistent results. If the operator can use the same procedure and calibrate the same sensor multiple times to the same end value then a sensor is much more likely to be accurate. If the sensor has varying levels of offset at the end each time, something must be fluctuating between calibrations.
Click here for the full article by Jim McCarty.