Systematic Errors

Systematic ErrorsMeasurement errors can be accidental and systematic. We shall explain the essence of a systematic error by the following example. Let's change resistance of R2 resistor in the circuit shown in the figure. In this case the whole measuring scale shifts over a certain value. So in order to get true measurement values we shall have to add or subtract a certain constant to the acquired value. The other case is when we change R5 resistance without changing R2 resistance. Here we change conversion conductance or characteristic slope. Thus we shall have to multiple instrument readings by a constant factor k. Both types of errors are closely connected with the concepts of additive and multiplicative errors. The first case is an example of an additive error while the second is an example of a multiplicative error. Unfortunately we have to put up with the fact that there are limitations to mathematical methods of error handling. Mathematical statistics handles only accidental errors. Calibration and graduation are the only ways of avoiding systematic errors. Well, of course, we can also select components. Calibration is different from graduation. When calibrating we change circuit constants so that the instrument readings correspond to the readings of a reference or known-good device. When graduating the circuit remains unchanged. We just register the reading bias and find the dependence of a true value on the device's output signal. The dependence can be presented in the form of calibration table on the device's case. However, there is a modern method of presenting dependence in the form of an equation when the value is then calculated on the computer or with the help of a built-in microcontroller. However, such method eventually transforms graduation into calibration though without changing circuit parameters. Thus the corresponding factors are corrected in a computer program.