During the past 60 days, the solar adaptive optics program at NSO/Sac Peak has achieved two major milestones: in mid-September, the control loop was closed for the first time on the bench, and the loop was again closed with solar granulation and small pores as the wavefront sensing target during the system's first tests at the Dunn Solar Telescope (DST) in early November.
The solar adaptive optics program, a joint effort involving both the NSO and the Air Force detachment at Sac Peak, is currently concentrating on converting the low-bandwidth active optics system developed previously into a higher bandwidth, low-order adaptive optics system capable of compensating about 20 spatial (Zernike) modes through atmospheric turbulence. A correlating Shack-Hartmann wavefront sensor with 24 subapertures, capable of using solar granulation or other time varying, low contrast, spatially extended targets, is used to measure the wavefront aberrations. The hardware design features parallel processing using off-the-shelf DSP components. This approach will allow the system to be expanded to more spatial modes later. A modal wavefront reconstruction algorithm is used to derive drive signals for the 97 actuators of the XINETICS deformable mirror from the wavefront sensor data. Currently, the development effort is concentrating on creating a functional system that will demonstrate the concepts and validate the approach; a user system will be implemented at a later stage.
For both the bench and the telescope tests, the servo loop was closed at an update rate of 800 Hz and a lag of more than 2.0 milliseconds, resulting in a system bandwidth of about 25 Hz. This is inadequate to fully compensate the atmosphere except under the best seeing conditions. The loop bandwidth is currently limited by both the frame rate and read-out speed of the wavefront sensor camera and to a lesser extent by the processing hardware.
These bottlenecks will be alleviated in the near future through the purchase of a faster camera and vendor improvements to the hardware. The small number of subapertures presents another performance limitation that will be more costly to deal with, and the current reconstruction and control algorithms, which remain imperfectly understood, are certainly not optimal. In particular, controlling the outer edge of the adaptive mirror has proven difficult. In the face of these and other difficulties (e.g., vibrations in the preliminary optical setup), the partial atmospheric compensation that the system achieved under good, but not exceptional, seeing conditions at the DST was genuinely exciting: improved and more stable resolution and a reduction of residual jitter were both clearly evident in the corrected image. Now that the critical milestone of closing the atmospheric loop has been achieved, attention can be turned to improving the performance of the current system and planning the next steps for the overall program.
T. Rimmele, R. Radick, R. Dunn, K. Richards