The advent of adaptive optics systems, which correct for the aberration caused by atmospheric turbulence, has dramatically improved the resolution of ground-based telescopes. Before a new design can be put into full-scale production, researchers must test it with simulations of aberration to make sure it works. Now a team at the University of North Carolina has developed a method based on inexpensive deformable mirrors to test the high-grade ones used for astronomy. Robert K. Tyson, who built the testing system with Benjamin West Frazier, explained that some researchers have placed electric hot plates near the beam path to generate dynamic optical turbulence, but that that technique is not repeatable. Others, he said, have mounted spinning, etched-glass plates with a spatial spectrum that matches the atmosphere in the beam path, but it is difficult to change the plates. Deformable mirrors based on the relatively inexpensive microelectromechanical systems (MEMS) are not advanced enough to be used as adaptive optics systems themselves, but it is now possible to test the systems with them. Developing the method was a two-step process, Tyson said. The researchers first developed a set of algorithms to drive the MEMS mirror into shapes that approximate atmospheric turbulence. This involved setting up a conventional adaptive optics system with a controlled deformable mirror and a wavefront sensor to measure the phase profile. "We had to be careful to make the MEMS mirror plane an optical phase conjugate of the wavefront sensor plane so that the information was not corrupted by other diffraction effects," he explained. "When the MEMS mirror produced a phase that was sufficiently close to the expected spatial spectrum of the atmosphere -- the Kolmogorov spectrum -- we saved those drive commands." The researchers then used these repeatable commands to assess the performance of other deformable mirrors, wavefront sensor configurations and closed-loop algorithms. Out of the sandbox Tyson said it is difficult to determine how this method compares with other techniques because they typically address a specific application of adaptive optics. The academic research environment, in contrast, required a more flexible solution. "We're working in a university research lab -- a sandbox, so to speak -- and we are constantly changing our configuration, our beam sizes, our diagnostics," he said. Besides astronomy, the technique may find a place in testing the performance of a free-space laser communications system, as well as other applications. "We can also use our MEMS mirror to simulate non-Kolmogorov disturbances such as those found in laser resonators or machine vision," he said. Students in his lab have been working with Xinetics Inc. of Devens, Mass., to develop a robust control system. Tyson and Frazier described the technique in the May 1 issue of Applied Optics.