Measurements of surface form involve a comparison between the form of the test surface and a reference. In coordinate measuring machine (CMM) measurements, this reference is the machine geometry. The measurement results typically show a height map representation of the test surface but these contain machine error contributions. This research demonstrates and applies an in-situ self-calibration technique, Shift-Rotation, to CMM-based freeform optics metrology. This technique minimizes task-specific uncertainty by separating machine error contributions from test surface measurements. Simulations were used to demonstrate the error separation of freeform measurements into components of the test surface and machine error contributions. Simulations also estimated uncertainties associated with applying the Shift-Rotation technique; these were Monte Carlo simulations that evaluated standard uncertainty contributions from potential sources inherent in the shift- and rotation-measurement process.The Shift-Rotation technique was experimentally validated by measuring a freeform surface on a Zeiss F-25 micro-CMM and a Mahr MarSurf LD 260 surface profiler. Error separations of these measurement results led to best estimates of the freeform surface, without machine error contributions. Zernike coefficients of these best estimates from the tactile machines were compared to those from optical measurements of the same freeform surface. The optical measurements were from a Zygo Verifire Fizeau interferometer and a Zygo NexView Scanning White Light interferometer.
