Change page style: 

Summary of the April/May 2003 commissioning runs

01 July 2003


After a 3 months rework at HIA and at Gemini (December-February 2002), Altair was back on the sky for 2 successful commissioning runs in April and May. The integration with the telescope, TCS and sequexec is basically completed. The system is now robust and has a high level of automation. Altair now achieves 55 to 85 milliarcsec images for seeing below 0.9 arcsec.


The third and fourth Altair commissioning runs took place from April 15 to 22 and May 8 to 16. After three months of rework at HIA and Gemini, these runs has been very successful. One by one, the major functionalities of the system, including the interactions with the telescope control system and the data acquisition sequencer, were successfully debugged, tested and characterized. Altair is now able to: (i) Receive signals from the instrument OIWFS (On-Instrument WaveFront Sensor) and use them to compensate for slow drift of image centering and defocus due to flexures; (ii) Offload at a slow rate the quasi-static shape of the deformable mirror to the telescope primary mirror; (iii) Offload the centering tip/tilt and defocus errors to M2; and (iv) Optimize several of the system operating parameters in real time such as modal gains and WFS centroid gain.

Altair is well integrated both with the TCS (Telescope Control System) and the TCC (Telescope Control Console). The telescope operator may now start and stop adaptive optics corrections from its TCC control screens.

After dealing with system functionalities, the main thrust of the team's effort during the last run shifted toward performance evaluation and improvements. Things are taking shape on this front. Images of full width at half maximum were obtained from 59 to 85 milliarcseconds (mas) on the vast majority of objects observed. This is slightly short of the ultimate diffraction limit (42 mas in H and 57 mas in K), but the main culprits have been identified. We have specific plans for improving these areas to reach our diffraction limited goals.

The team was generally very pleased with the behavior of the system. The rework at HIA and Gemini in the three months preceding the March run- including the refitting of a new lenslet array, work on the NIRI OIWFS gimbal mechanisms and several software improvement and fixes - has brought considerable improvements in the stability and overall performance of Altair. Following these successes, we are more confident than ever that the Gemini facility AO system will be fully competitive with its most serious competitors (e.g. NAOS).

Investigation of how faint a guide star the system can use were done during the May run. We have been able to close the loop successfully on a M6 V=15.4 magnitude guide star at 100Hz. Of course, this is a very red star. For K0 stars, the limiting magnitude is expected to be around 14.5, obviously providing partial compensation only (see the pages on Altair performances for more details).

During the next commissioning run in July, we expect to concentrate on the remaining problems and gray areas, including:

  1. A relatively strong vibration induce image broadening of up to 40 mas per axis. We have been able to identify the NIRI cryo-coolers as the source of the vibration, shaking the Altair bench and mirror mounts. We also found that there is a small vibration of the tip-tilt mirror itself, probably due to the mirror being loose in its mount, but although we plan to fix the latter problem eventually, it does not amount for a significant part of the image broadening now. We have been routinely achieving 60-65 mas at H and K wavelengths.
  2. Some high spatial frequency static aberrations on the telescope secondary mirror are aliasing into the Altair WFS and reduce the performance of the system. This depends on the relative position of M2 with respect to Altair and are therefore difficult to calibrate. We are fully aware of the problem, and are making plans to address and fix it in the next 6 months. In the meantime, we are working on a patch implemented inside Altair that could reduce the amplitude of this effect.
  3. Regarding the isoplanatic angle, the offloading of the Altair DM static shape to M1 has done a lot to improve the performance off-axis. We are fully aware of the importance of this issue. To address this, we regularly observed stellar fields during the runs. The data are been analyzed and a database has been established which will help us characterize the anisoplanatism (see the pages on altair performance for more details).

M13 H band

Left: 22x22 arcsec H band image of the central area of the M13 Globular Cluster (click to enlarge). The stellar core are close to the diffraction limit (60 milliarcsecs). Some traces of anisoplanatism can be seen in the outskirts of the field, where the FWHM is around 82mas. This is a 9 minute exposure, with eighteen 30seconds exposures combined.

V1318 Cygni K prime

Left: 22x22 arcsec K prime band image of V1318 Cygni (click to enlarge).

Last update March 3, 2003; Francois Rigaut.