Logo of the Physikalisch-Technische Bundesanstalt
Production sequence of Si-spheres and interferometrical determination of the sphere volume

Interferometric Calibration of Gauge Blocks

Working Group 5.43

 

Statement and comments about the confirmation of tolerance classes for gauge blocks

The most important information on a gauge block is the absolute length at 20 °C and its associated measurement uncertainty. To assess the conformity of the tolerance class (grade) of a measured gauge block (see ISO 3650), it is necessary to take the measurement uncertainty into account (see EN ISO 14253-1). If the measurement uncertainty and the stated tolerances (central length lc and deviations from the central length (variation in length) are of the same order, the tolerance class of a gauge block often cannot be clearly determined (e.g. measured deviation of the central length from the nominal length: 30 nm, measurement uncertainty: 50 nm, tolerance: 50 nm). A reliable confirmation of the tolerance class of a gauge block according to EN ISO 3650 is only possible if the measurement uncertainty is clearly lower than the tolerance and if, in addition, all other requirements (dimensions, material properties, length stability, perpendicularity, etc.) are fulfilled.

Another problem is the comparison of measurement results obtained by different measurement methods. In the case of interferential length measurements, the surface of the gauge block and that of the plane platen are influenced by wringing forces, whereas the gauge block is measured at the 4 corners only in the case of a mechanical comparison measurement. Hence, the influence of the surface topography can lead to different measurement results for the central length and for the variation in length. In the case of a mechanical comparison measurement, the variation in length is measured at the four corners of the gauge block and therefore only provides results with a statistical statement, since the maximum variation in length does not necessarily occur at the four corners. The latter is used especially to check the quality of gauge blocks but it is not suitable to determine a tolerance class with metrological accuracy and with very low measurement uncertainties. It is, however, used in most industrial calibration laboratories all over the world in order to cut costs.

The metrological function of high-accuracy gauge blocks (calibration grade K, grade 0) is to ensure the transfer of the length in the centre of the gauge block's measuring face.

The most important preconditions for a gauge block reference standard are good wringing and reflection properties, a good parallelism in the centre and high length stability of the gauge block. If a gauge block complies with these requirements, then it is possible to calibrate it with the smallest measurement uncertainty in order to ensure its traceability to the SI unit "metre". The best data is then the calibrated central length, together with the indication of the associated measurement uncertainty.

Neither does it make sense, nor is it cost-effective to use gauge blocks only with the indication stating that they comply with a certain tolerance with regard to the central length (conformity) – i.e. that they comply with, e.g., "Calibration grade K" or "Grade 0" – since this does not allow the full information to be exploited. Formally complying with a grade is not metrologically relevant. Since the measured value for the central length is indicated in the calibration certificate, it is not metrologically relevant if the tolerance is slightly exceeded. What is more important than the confirmation of conformity with a grade with regard to the variation in length is that the gauge block is capable of being calibrated without restriction – especially in the centre of the measuring face. The question as to which grade a certain gauge block can be assigned to is often an endless story without any metrological relevance.

For all these reasons, the Physikalisch-Technische Bundesanstalt (PTB) does not normally indicate a value for the variation in length. An exception to this rule is made in the case of gauge block calibrations in which PTB determines the length at the four corners using interferometry and also checks these values by means of a mechanical comparison measurement.

As suggested by PTB, the "Grade 00" classification was deleted from the latest issue of ISO 3650. We are persuaded that the "Grade 0" classification should also be deleted to avoid the problem that occurs when the ratio between the tolerance and the measurement uncertainty is not favourable.