Error reductions for stitching test of large optical flats

Stitching test of large optical flats is still quite challenging. In this paper, aiming to reduce the 2nd order error accumulation effect, separation of the reference surface error from the subaperture measurements is proposed in stitching of large flats. The reference error is decomposed into power and astigmatism, residual Zernike polynomial surface and high frequency component, which are estimated or calibrated, respectively. The error separation procedure is then incorporated into the subaperture stitching algorithm. Finally experiments are presented to show the effect of error reductions in stitching test of a 600 mm-aperture flat, with reference surface separated and verified by the factory report of the transmission flat.

► Reference error is decomposed into three different parts.
► Power and astigmatism are calibrated by modified three-flat testing.
► Approximately estimate the high frequency component by averaging.
► Residual Zernike polynomials are self-calibrated during stitching.
► Experimentally verified by stitching test of a 600 mm-aperture flat.