Intercomparison of Measurements on Surface Roughness Standards

In this project we performed from 2001 to 2003 an intercomparison in the field of surface texture. 14 NMi’s carried out measurements on 1 ISO 5436-1:Type A2 depth-setting standard, 3 Type C calibration standards, 3 Type D1, and 1 Type 2. The measurements performed covered a number among the most interesting parameters, i.e. Pt, D, Ra, Rz, RSm, and Rk. Additionally, a first comparison using ISO 5426-2:Type F1 software gauges was included, taking into consideration that the contribution of the software to the results becomes more and more important. The participating laboratories provided a report with a short description of the measurement set-up and conditions, the results of the measurement, and the combined standard uncertainty. The uncertainty of the measurement should be calculated following the rules in the GUM. Contributions to the uncertainty budgets depend on the method and the instrument used.For each parameter a reference value was calculated as the weighted mean of all measurements yi. The weights were u-2(yi). To set up the |En | £ 1 criterion, the expanded uncertainty U with a coverage factor of k = 2 was used. A considerable part of the over 600 measurement results did not fulfil the En<1 criterion. Problems could be related to the determination of the depth on the depth setting standard. This was really not expected at the beginning of the comparison. In one case the evaluation revealed that the measured values of an institute failed completely. Further problems could be related to the definitions of some 2D roughness parameters, e.g. RSm, in ISO 4287 and ISO 13565 which are not precise enough. The results obtained for three software gauges (ISO5436-2 Type F1) were compared with the results calculated using the certified software RPTB. This shows that some problems obviously occurred for the calculation of the Rp, Rv and Rt parameter.Overall the intercomparison showed that it is really important, for accurate calibrations, to have accurate instrumentation as well as accurate software. Therefore both have to be checked by appropriate means. The last intercomparison in the field of surface roughness was finished in 1989. The institute of Prof. De Chiffre, Technical University of Denmark, and the laboratory Mikro- und Nanotopographie, Physikalisch-Technische Bundesanstalt, Germany, acting as pilot laboratories are preparing an EUROMET round-robin comparison for surface texture measurements. We plan to measure some roughness standards of type B, C and D as described in the ISO 5436, together with a type A standard. Since the software for data analysis is becoming more important we will include some data sets which should be analysed by the participants' software.