**1. Introduction**

Metrological verification using optical equipment is of increasing interest in industry. In this sense, one of the most widespread technologies is laser triangulation sensors, either mounted on coordinate measuring arms or on tridimensional coordinate measuring machines, or even as independent sensors that can be integrated in multiple industrial applications. When a laser triangulation sensor is used for scanning (Figure 1), a laser beam is first projected onto the surface to scan, then the projected intersection line is captured by a digital camera, and finally the coordinates of the points are determined by trigonometric calculations (hence the term "triangulation"), taking into account the angle between the laser beam and the camera orientation, and the light intensity captured by the digital camera, among other parameters [1]. The highly extended deployment of these sensors has also been possible due to improvements in the accuracy and capabilities of these devices. Manufacturers have enhanced these instruments through adjustment and calibration processes, apart from increasing the quality of designs and materials for the internal components. Currently, many users and researchers employ these sensors for measurement (Geometrical Dimensional and Tolerancing - GD&T- verification) as well as for reverse engineering typical tasks.

**Citation:** Meana, V.; Cuesta, E.; Álvarez, B.J. Testing the Sandblasting Process in the Manufacturing of Reference Spheres for Non-Contact Metrology Applications. *Materials* **2021**, *14*, 5187. https://doi.org/ 10.3390/ma14185187

Academic Editor: Gilles Dessein

Received: 22 July 2021 Accepted: 8 September 2021 Published: 9 September 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Figure 1.** Principle of laser triangulation scanning.

The improvement in accuracy, including traceability assessment [2–4], is a crucial factor to consider. It is precisely in this line of research where several factors that have an influence on laser triangulation sensors have been analyzed. Apart from the equipment's own parameters and specific tests [1,5,6], other parameters have also been taken into account, such as scanning speed [7] and scanning strategy [8,9]. Moreover, external parameters, like those derived from the part or object acting as the measurand, such as the material [10,11], color, surface roughness [12,13], ambient light [14], or even the type of the geometry to scan [3,15,16]. The idea is to assess the measurements that can be made with these non-contact technologies, thus extending their application beyond typical reverse engineering applications to metrological (GD & T inspection) ones.

Two of the fundamental parameters in the scanning quality of this type of sensors are the orientation of the sensor with regard to the target surface [8,17] and the surface finish of the part [12,18,19]. Shiny surfaces cause defects in the captured pointclouds, whereas, on the contrary, matte surfaces enhance the coverage of the capture. To achieve this type of surface finish over metallic surfaces, sandblasting is one of the most promising processes. In fact, for certain optical instruments, sandblasted surfaces are among the few surfaces that allow a measurement accuracy comparable to that achieved by contact measurement [18].

The aim of this work is to validate the sandblasting process as a process for modifying the surface condition of precision spheres in order to use these spheres as reference artefacts for adjustment, verification and/or calibration of optical sensors and non-contact reverse engineering equipment. Only if this process is sufficiently conservative with respect to the geometry of the precision spheres can it be valid for creating reference objects. On the contrary, if the damage or modification (dimensional or geometrical) is excessive, the process will not be suitable for this purpose.

In particular, a laser triangulation sensor (3D scanning) mounted on a Coordinate Measuring Machine (CMM) was used in this study. The CMM permits automation of the scanning process, avoiding errors derived from manual operation (as occurs with laser triangulation sensors mounted on coordinate measuring arms). Thus, factors like focal distance, scanning orientation, or scanning density were not considered as variables that can influence the experiment.

Although the finishing process is identical, the essential difference with respect to the work presented in [19] is that in that work the geometrical quality of the workpieces was qualified as very low, with a high surface roughness. The spheres were manufactured by metal laser sintering (SLM), and the post sandblasting was aimed at improving the geometrical quality, which was proved by both contact and laser measurements. In fact, the improvement ascertained by the contact measurement was very high.

However, the situation is the opposite in the case presented here. The original workpieces possess a high dimensional and geometrical accuracy, with very low deviations for the diameter and sphericity, due to the application of superfinishing processes and a final polishing. Conversely, this excellent surface finish makes it very difficult (and can even impede) the measurement of these spheres with laser triangulation sensors. By sandblasting these spheres, the purpose is to enable their use in laser measurement, but assuming a probable and unavoidable loss of accuracy caused by sandblasting.

Specifically, the objective of this article is, on one hand, to quantify the loss of dimensional and/or geometrical accuracy, and, on the other hand, to quantify the improvement in the laser capturing. The sandblasted spheres are suitable as reference elements only when the loss of accuracy is not too high and, on the contrary, when the improvement in laser capturing is substantial.

The final target of this research is to find out whether a low-cost process (manual sandblasting) can be applied to stainless steel precision spheres, of very low cost as well, to materialize calibration spheres for non-contact metrology. It is expected that the loss of precision in both diameter and form error of the post-blasting spheres will be low enough for this purpose. Ideally, the form errors of the sandblasted spheres should be at least one order of magnitude lower than the measurement uncertainty of the optical equipment. In any case, this experimentation will reveal which equipment is suitable for being calibrated with this type of sandblasted spheres.

Nowadays, the manufacturing of precision ceramic spheres (grades G3, G5 or G10, with sphericity < 0.25 μm, *Ra* < 0.020 μm, according to ISO 3290/DIN 5401 [20]) is very costly. This is mainly due to the fact that they are built specifically for this purpose, starting from ceramic powder, which is sintered and subsequently polished. They also require high hardness and wear resistance, using materials such as ruby, alumina, sapphire, or zirconia, among others. However, when they are intended to be used as reference elements for non-contact measurements, neither high hardness nor high wear resistance is required. In the context of the aims presented in this work, it is in fact sufficient that they are made of stainless materials, for example aluminum alloys or steels of qualities such as AISI 304, AISI 306L, or similar.

The spheres used in this research are stainless steel precision balls commonly used in the bearing industry, the cost of which is lower, but which also feature worse manufacturing qualities (G50 or G100 [20], with sphericity < 2.5 μm, *Ra* < 0.1 μm).

For optical applications, G100 accuracy or lower (according to ISO 3290 [20]) is enough unless the spheres show excessive brightness. Specifically, the idea of the experiment is to eliminate the very shiny finish (mirror-like) of the sphere surface, checking the variations in both diameter and form error caused by the sandblasting process. Another very important objective of the experiment is to quantify (if it exists) the improvement in the quality of the point clouds achieved by a laser triangulation equipment with respect to the cloud obtained on the polished, pre-sanded sphere.

The shot peening process, in this case sandblasting, will be carried out in a manual sandblasting machine using sand with fine grain size. Obviously, a certain variability is introduced even though the process variables are controlled (grain size, exposure time, distance, and the incident direction of the sandblasting stream onto the spheres). This variability must be studied so that the detected wear (or surface attack) will be assessed as rigorously and objectively as possible.

Therefore, the work includes sandblasting tests on stainless steel spheres of different diameters, evaluating (by CMM measurement) the variation of the mean diameter value and especially the loss of form error. In addition, it will also be interesting to analyze the variation of the standard deviation of the point cloud, as this is a crucial parameter in the measurement of the quality of the laser point cloud [19].
