Large-mirror testing facility at the National Optical Astronomy Observatories.

A method for testing the surfaces of large mirrors has been developed at the National Astronomy Observatories (NOAO) to be used even when conditions ofvibration and thermal turbulence in the light path cannot be eliminated. The full aperture ofthe mirror under test is examined by means of a scatterplate interferometer that has the property of being a quasi-common-path method, although any means for obtaining interference fringes will do. The improvements in the NOAO method lie in the means for capturing the fringe pattern and the analysis that follows. The method uses a remotely operated CCD camera system to record the fringe pattern from the workpiece. Recording the spot pattern resulting from a Hartmann test screen placed over the workpiece is also possible, but the analysis methods will differ from those discussed. By operating the test equipment remotely, the optician does not cause unnecessary vibrations or heat in the testing area. The typical test is done with a camera exposure of about a millisecond to "freeze" the fringe pattern on the detector. Averaging up to 10 separate exposures effectively eliminates the turbulence effects. The analysis program is based on a method originally developed for examining telescope images affected by atmospheric seeing. From the intensity information, a phase map of the wavefront reflected from the surface is obtained using a new phase-unwrapping technique. The resolution is limited only by the number of fringes in the test aperture. The method described provides the optician with complete numerical information and visual plots for the surface under test and the diffracted image the method will produce, all within a few minutes, to an accuracy of 0.01 μm measured peak-to-valley. The method has been extensively used for a variety of tests of a 1.8-m-diam borosilicate glass honeycomb mirror, where the method was shown to have a sensitivity equal to a Foucault test. The test setup is described, including the analysis technique, and examples of the results obtainable are provided. The method continues to be used at NOAO for testing large optics (currently a 3.5-m honeycomb mirror).