351 nm Laser Guide Star Pupil Image
13 Aug 2004, telescope 20 degrees off-zenith,
air mass = 1.20. Image captured just prior
to tests on Alpha Ophiuchus
|Laser Guide Star System Upgrades|
During the UnISIS development, obtaining sufficient laser guide star pupil brightness
became one of the major challenges. As the image at left shows, this problem is no
longer an issue. Many improvements were made along the way. The original UV wavefront
sensor used only one polarization; it now uses both. The original wavefront sensor
had a QE of 49%; a new one was purchased that has QE of 63%. Many very critical optical
alignment and focus issues were solved during the instrument development period. Working
at a UV wavelength of 351 nm makes the alignment issues a great challenge: the UV light
is invisible to the eye and the short wavelength means that the tolerances are very tight.
Note: As in all AO systems, the pupil is illuminated with an array of Hartmann images. If the wavefront CCD is moved laterally (in the pupil plane) 0.5 pixels in both x and y axis, the image at left would immediately take on a bright-dark-bright-dark checker board look because the individual Hartmann images would fall in the center of single pixels. Instead, you see a relatively smooth intensity distribution in the image on the left because the wavefront sensor is positioned so that the Hartmann subimages sit at the centers of 2x2 quad-cells. The irregular illumination captured in the image shows the extent of the aberrations present in the wavefront.
Please note that the images shown below -- except for the very last one -- were taken prior to Oct. 2004 when the UnISIS control computer was switched from the old slow system to the new Real-Time Linux system. A significant improvement in performance is expected from the new high-speed AO system under good atmospheric conditions.
|Altair: Left Panel Shows AO-Off while the Right Panel Shows AO-On|
|The two-panel image to the left demonstrates the Natural Guide Star AO performance of UnISIS. On the far left is the AO-off image showing that the natural seeing on this night was about 2 arcsec. The panel on the right shows AO-on with a nice tight image core. Note that this AO-on image was taken with the star positioned near the center of a 2.11 arcsec diameter focal plane aperture. This aperture was created on a glass substrate by NOT aluminizing the 2.11 arcsec diameter aperture. So the AO corrected star light is seen reflecting off the bare glass surface where the reflectivity is in the range of 2.5% whereas the exterior region around the 2.11 arcsec aperture has ~90% reflectivity characteristic of a standard aluminum coating.|
|H-Band Image||Z-Band Image||Explanation|
|The images on the left provide another example of UnISIS Natural Guide Star AO performance, this time in the dual wavelength mode. On the far-left is an H-band image at 1.65 microns (Strehl 0.38) and on the near-left the Z-band image at 0.85 microns (Strehl 0.20). The natural seeing was variable between 1.2 - 1.4 arcsec during these exposures. The 2.11 arcsec pinhole sets the image scale. Note that the NGS tip-tilt signal comes from the wavefront sensor data.|
|H-Band Image||Z-Band Image||Explanation|
|Immediately after the two above images were taken with the NGS side of UnISIS, control of the system was switched to the LGS side at 167 Hz. The H-band image (again far-left) shows a Strehl ratio of 0.13 which seems to be characteristic of nights when the raw image quality is low. Only when the natural seeing is 1 arcsec or better does the UnISIS LGS performance reach its "theoretical" limit of Strehl = 0.28 at H-band. Why? When the seeing is poor, the LGS itself is blurred and hence the AO system cannot correct as well. There is no doubt that further performance loss came from the slow (167 Hz) correction rate.|
|Log-Stretch of H-Band Image||Line Cut through Image||Explanation|
|After the upgrade to the fast Real-Time Linux AO control computer, UnISIS was run in the NGS mode for a single night. The seeing was terrible with uncorrected image FWHM = 2 arcsec or perhaps even worse. With the AO system running at 622 Hz, the image at the left was obtained on this second magnitude star. Because of the very poor seeing, it was difficult to precisely adjust the AO system and hence the tight image core is surrounded with a set of 4 faint "satellite" images from what is called a "waffle mode". Strehl = 0.2.|