Week 35 of the Early Operations system optimization period
The team used two nights of on-sky observations with the Simonyi Survey Telescope this week for a combination of on-sky engineering and pre-LSST observations. There were four nights of unscheduled engineering downtime for maintenance on compressors for the LSST Camera cryogenic refrigeration system, and one night of closed dome testing due to clouds.
On-sky engineering activities for the Active Optics System (AOS) included full-array mode (FAM) measurements in the u and g bands to examine dependencies of the optimized optical state on Camera filter. The team also deployed and validated several updates for the AOS, described in turn below.
First, the team has been developing and testing a series of AOS software infrastructure upgrades in preparation to deploy an improved optical model of the reference wavefront that defines the target optical state the AOS aims to converge toward. The reference wavefront minimizes wavefront errors across the full field of view, limited by the intrinsic aberrations of the as-built system. The refined reference wavefront to be deployed is based on measurements of the optical wavefront across the full field of view using FAM observations with the entire focal plane pistoned to intra- and extra-focal positions. The FAM data show that there is one component of the measured intrinsic wavefront that is fixed in Camera focal plane coordinates (at least partially attributed to micron-scale variations in the detector heights across the focal plane) and a second component of the measured intrinsic wavefront that is fixed in the observatory coordinate system related to the telescope optics, as expected. Accordingly, the measured reference wavefront includes these two separable components to accurately model the optimized optical state for different Camera rotator angles.
Second, the team deployed an update for the wavefront estimation pipeline that introduces a new algorithm to more robustly model the telescope pupil during the iterative fitting process. The new algorithm projects the telescope pupil model to the focal plane during each step of the fitting, rather than solving the inverse problem of mapping the observed “donut” image in Camera pixel space back to the telescope pupil. The new approach is expected to yield more numerically stable and repeatable wavefront estimates.
Third, the team took advantage of recent refinements of the telescope pointing model to enable selection of “donut images” for wavefront estimation using a catalog of pre-selected bright, isolated stars, and incorporating other quality selection criteria, to further improve the robustness of the wavefront estimation pipeline.
Closed dome calibration work included wavelength scans using the collimated beam projector (CBP), expanding the set of measurements in multiple bands and multiple pointings. The team also took measurements with a separate “traveling collimated beam projector” (TCBP) that will be used to calibrate the photometric response using on-site measurements at multiple observatories, including Rubin Observatory and the Zwicky Transient Facility.
With the return to on-sky testing following a couple weeks of winter weather and closed dome engineering, the team plans to run sustained survey observations for the next several nights. The Rubin Observatory Director has accepted the recommendation from the Survey Cadence Optimization Committee (SCOC) for the survey strategy at the start of LSST, and Summit Operations is now working to deploy the Feature Based Scheduler (FBS) configuration to implement this strategy at the summit. The SCOC continues to welcome input from the science community regarding the LSST survey strategy.
The currently installed filter set is ugriz. Due to other daytime engineering needs, the filter swap initially planned for 19 June did not occur, and the team elected to pass on filter swaps for the current lunation. The next filter swap to remove u and install y is now planned for 21 July.