During the second week of on-sky commissioning with LSSTCam, the Rubin Observatory team concentrated on testing the Active Optics System (AOS) open loop and closed loop components. The objective was to improve stability of delivered image quality while tracking targets over a range of azimuth and elevation angles, and while executing translational dithers, rotational dithers, and filter changes. A typical observing pattern consisted of slewing to a target field, using the AOS closed loop to bring the optics into focus and alignment, and then using the Feature Based Scheduler (FBS) to queue up many repeated visits of the target while the “survey mode” AOS closed loop refines the delivered image quality using measurements from the corner wavefront sensors. These sequences of maneuvers are needed to deliver the early datasets for testing the coaddition and difference imagining processing steps with the LSST Science Pipelines. This mode, with frequent slews and dithers, is naturally more challenging for the AOS, so the typical image quality this week has been somewhat worse than last week, with many improvements still to come.
Acquiring these data during commissioning has some special considerations. For example, during the 10-year LSST survey, each patch of the survey footprint will be observed from multiple Camera orientation angles relative to the sky with observations on many different nights. In order to mimic this pattern during the brief on-sky commissioning period, the team is using rotational dithers with the Camera rotator to efficiently sample sky rotation angles. Early in the AOS commissioning process, the team observed that an accurate rotation angle look-up table (LUT) is required to maintain stable delivered image quality as the physical rotator angle changes, as expected. The team took both in-dome laser tracker metrology and on-sky measurements to refine and validate the rotator angle component of AOS open loop system. AOS testing also included building a sensitivity matrix for the Camera and secondary mirror (M2) hexapods as well as the primary-tertiary mirror (M1M3) and M2 bending mode degrees of freedom by perturbing one degree of freedom at a time and measuring the system response. Collectively, these tests are helping to validate the many coordinate systems used in the telescope control software, and prepare the way for further AOS optimization. Additional AOS work included optimizing the closed loop, e.g., tuning the set of enabled degrees of freedom and adjusting the gains, to reduce ellipticity across the field of view. This work is ongoing.
As the AOS improved capability to converge towards stable image quality across the full field of view, the team increased the duty cycle of FBS-driven observations, gathering up to 372 science program visits on a single night across the ugri bands by the end of the week. To date, there have been more than 5,000 sky exposures, amounting to 15 trillion pixels.
While the top priority of the week has been supporting sustained on-sky observations, the team also collected first calibration images with the collimated beam projector.
LSSTCam continued to run smoothly. Work continues in parallel on the remaining 3 science sensors (out of 189) that are not yet fully operational.
The team continues taking initial datasets for Science Pipelines commissioning. Looking forward, the team plans to use the upcoming bright time around the full moon in mid-May to continue AOS commissioning with a focus on refining the pointing model and LUTs of the open loop system, and to continue commissioning of the in-dome calibration systems. The currently loaded filter set during dark time is ugri together with the pinhole filter to help characterize stray and scattered light from both celestial sources and sources within the dome. As we enter bright time, the team plans to swap in the z filter and y if time allows.