Assessing the effect of penetration length variations on dimensional measurements with X-ray computed tomography

From discussions in various standardization committees, testing material thickness influences on dimensional X-ray computed tomography (CT)-through variations of X-ray penetration-path lengths-has become a matter of great interest. Although this topic does not seem to be fully covered by the VDI/VDE 2630-1.3 guideline, other emerging documentary standards (ASME B89.4.23 and ISO/DIS 10360-11) are trying to address it in their guidelines. This paper evaluates variations in CT dimensional data when, to provide different penetration-path lengths to X-rays, enclosures of varying aluminum thickness are added ('plus-shell' condition) to surround three test workpieces during measurement. The test workpieces are two hole plates of square-shaped geometry with 28 drilled holes, one made of aluminum (48 mm x 48 mm x 8 mm) and the other made of steel (6 mm x 6 mm x 1 mm), and a multisphere artifact containing twelve ruby spheres of 4.0 mm nominal diameter. Geometrical features in the 0.5-66 mm range are studied. The main results show that while the deviations between CT and reference/calibrated data for uni-directional (center-to-center) lengths can be confined to a range of +/- 5 mu m, with no significant differences in measurement results between no-shell and plus-shell conditions, bi-directional (edge-to-edge) lengths and form measurements have substantial differences-with deviations of up to 16 mu m for form measurements in the aluminum hole plate in the no-shell condition and 42 mu m when the same hole plate is housed within a highly irregular (gear-like) aluminum enclosure.