The difference between actually moved distance and desired distance is measured. Known deviation can be compensated in this way.
If only the linear share (line of best fit) is deducted from the measuring value, then the increase of the fault can be saved in the controller as a simple factor. This so-called linear compensation is specifically suited to systems which work without scale. The accuracy can usually be improved with a factor of up to three.
If accuracies are not linear, then the position-dependent measured fault can be displayed as a polynomial equation and saved in the controller. This assumes that the controller has the ability to generate polynomial equations and to offset these with the setpoint while positioning.
If the controller has this function, then it is also possible to save a table with position-dependent deviations. The controller can then interpolate between these supporting points. The distance between the supporting points depends on the desired accuracy and the spectral distribution of faults in reference to travel. This process is called correction table or error mapping.