Discrete L-spline filtering in roundness measurements

Abstract

One of the most basic geometric primitives of a mechanical object is its circular feature. The departure of a feature from roundness is defined to be the difference between its maximum and minimum radial ordinates. Due to the imperfection of the measuring process filtering is mandatory before measurement results are evaluated to calculate roundness values. But the selection of a special filter type strongly influences the resulting roundness value. The filter prevailing today, recommended for use in roundness measurement evaluations, is the phase-correct 2RC filter. But since the Gaussian filter was introduced in roughness measurement evaluations by the ISO, experts have considered the application of this type of filter in roundness measurement evaluation as well. Recent investigations, however, show that this is not feasible. On the other hand, further use of the phase-correct 2RC filter is not advisable either, because of the disadvantages of this filter type. Therefore a discrete L-spline filter has been developed, which yields a good approximation to the transfer function of the phase-correct 2RC filter, but avoids the disadvantages of this filter type. The filter algorithm has been implemented as a short C-function to be used with any evaluation program, yielding a very fast filtering process. The application of this function to measured data is given for demonstration.