Surfaces with nanometric texture, machined by ultra-precision diamond milling (UPDM), have wide applications for their functionalities. Surface texture characterisation is crucial for their quality evaluation and it would vary in the measurement data processing under various scales, which is termed scale effects. However, the scale effects have not been comprehensively studied. In this work, we focus on the scale effects on surface texture characterisation of UPDM. Firstly, a series of experiments were conducted at different material removal rates to produce surfaces with nanometric texture. Secondly, a discrete wavelet transform (DWT)-based method was employed for surface filtering and surface texture parameters (Sa, Sq, and Sz) were evaluated according to ISO standards. Thirdly, power spectral density (PSD) analysis and surface filtering comparison were performed to realize a scale observation. Furthermore, the effects of the filtering scale, the sampling scale, and the subarea scale on surface texture characterisation were investigated. Finally, the results showed that surface topographies were governed by various machining factors which resulted in multi-scale topographical components. The PSD analysis enabled discrimination of the multi-scale topographical components to the underlying machining factors. The DWT-based method achieved a relatively small deviation (<4%) for surface filtering as compared to the zero-order Gaussian regression filter. The filtering scale affected the filtered topographical components and it visualised and differentiated those components with the machining mechanisms. The sampling scale determined the filtered components to yield an impact on the characterisation results and Sz was more sensitive to this effect than Sa and Sq. The subarea scale influenced the probability distributions and statistical values of surface texture parameters within its contained components and it indicated that the increased material removal rate resulted in a deterioration of surface uniformity.
Read full abstract