| Laser interferometrie represents a major tool in machine-tool axis characterization tasks. This work aims at evaluating the robustness for decoupling scale, pitch and yaw errors from linear measurements. Multi-sequential measurement lines make the basis of this method.; Scale error measurements are carried out with linear optics and modeled by a mathematical relation that incorporates the coupling behavior with angular errors (pitch and yaw). A least-square system follows from the multiplicity of measurement action lines.; A study of the measurement uncertainties is then explored using jacobian matrices. These are derived based on the assumption of the superposition principle. Thus it is possible to evaluate the error on the solution for each of the measurement uncertainty added to the equation.; Straightness measurements are approached in a similar fashion. In fact, straightness readings are coupled with another angular error, which in this case is the axis roll error. Laser alignment with the axis nominal direction of travel constitutes a obstacle for accurate decoupling of the horizontal and vertical straightness from the roll error. Upon these statements, few solutions involving many action lines combinations are proposed.; Experimental data of scale errors are obtained from six action lines of a prismatic axis. These are used to populate the identification system which is solved by the Moore-Penrose generalized inverse. Scale, pitch and yaw errors are identified and compared to reference measurements made with angular optics. Robustness proves to be fairly good as shapes of identified errors are close to the reference. Some first order residues are obtained between angular errors.; The jacobian matrix system is used to analyze what uncertainty may be causing such first order residues. The discussion finally points out that thermally induced errors are showing the greatest effect. |
