Despite advances in various metrology tools, interferometry remains the method of choice for measurements of optical surfaces. Fizeau interferometers can achieve precisions of X/100 PV (and better) with proper environmental control. The quality of the reference surface, however, usually limits the uncalibrated accuracy to merely X/10 PV or so. Various methods have been developed for "absolute" (unbiased) surface testing, including the N-position, 3- flat, 2-sphere, and random average tests. The basic principle of these tests is that the reference wave error remains invariant when the part is moved. These tests as a rule require multiple parts and/or measurements at different positions. Sub-aperture stitching requires measurements at multiple positions, and thus in principle can measure reference wave error. QED's stitching algorithm exploits this possibility to produce a measurement of the reference surface along with the stitched full-aperture phase. The precision mechanics of QED's stitching workstation make it an excellent platform for performing conventional reference wave calibrations as well. Results obtained from the QED stitching algorithm are compared with other calibration methods performed on the same workstation. The mean results and uncertainties of the various methods are evaluated, and limitations discussed.
|