Uncertainty estimation of shape and roughness measurement

Abstract

One of the most common techniques to measure a surface or form is mechanical probing. Although used since the early 30s of the 20th century, a method to calculate a task specific uncertainty budget was not yet devised. Guidelines and statistical estimates are common in certain cases but an unambiguous method for all kinds of measurements and measurement tasks is absent.Anew method, the virtual measurement machine, already successfully implemented in CMMs, is now applied on a specific group of stylus measurement instruments namely for: • roughness; • roundness; • contracers (form measurement). Each of these types of machines use the same measurement principle; a stylus is pressed against the object with a well specified force, moved across the object and the trajectory of the stylus tip is registered. The measurement process and its disturbances can be described theoretically and mathematically. Each disturbance or influencing factor which contributes to the uncertainty of the measurement is modeled and with this model simulated (virtual) measurements are generated. The virtual measurement depends upon the magnitude and range of the influencing factor. Some examples of influencing factors are; tip geometry, measurement force, probe gain factors, squareness of measurement axes, etc... The sensitivity of each factor upon the measurement is calculated with so-called virtual measurements. Recalculation of the describing parameters of the measured object with the virtual measurements gives the amount of uncertainty attributed to the influencing factor or machine parameter. The total uncertainty budget is composed out of each contribution in uncertainty of each machine parameter. The method is successfully implemented on two machines: the SV 624-3D (roughness and shape) and theRA2000 (roundness, form and cylindricity). It is shown that an on-line uncertainty budget can be calculated specifying each contributor. As not only gain factors need to be calibrated, but more input variables, e.g. calibration data of machine parameters, are required by the uncertainty calculation, calibration artefacts are developed to perform such a task. The artefacts can be used to perform a total and fast calibration on the shopfloor directly traceable to the appropriate primary standard. Combining the virtual measurement machine, implemented for roughness, roundness and form in high quality software, with the calibration artefacts, a powerful measurement tool is realised which allows to calculate a task specific uncertainty budget for these types of machines and creates a traceable measurement result which can be accredited by accreditation organizations.