Week 24 of the Early Operations system optimization period
At least 100 pre-LSST visits were acquired on each night of the past week, and more than 3000 total pre-LSST visits have been acquired since resuming intermittent pre-LSST observations on 3 April following a period of intensive on-sky engineering in March. The team set a new single-night record of 842 science program visits on 5 April. During the past 2 weeks, on the 4 nights that were fully devoted to pre-LSST observations, the ratio of acquired pre-LSST visits relative to the expected number of visits for an ideal full night with the current baseline LSST survey cadence (v5.1) varied between 0.75 and 0.9. For comparison, the 10-year average value of this ratio for the baseline LSST simulation is 0.63, considering expected losses for weather, engineering, and downtime.
A total of more than 27K pre-LSST visits have been acquired since the Early Operations period began in late October 2025. For comparison, ~23K science program visits were acquired during the on-sky commissioning period from April-September 2025.
The pre-LSST observations of the past week included observations of the COSMOS LSST Deep Drilling Field that produced alerts on 6 distinct nights, with >30 total visits acquired across the griz bands. The testing this week showed that median run times for Alert Production are now below the SRD specification of 120 seconds, meaning that alerts are consistently being distributed within 2 minutes of shutter close.
During the pre-LSST observations this week, the per-night median PSF FWHM varied between 0.9-1.1 arcseconds, the per-night median estimated instrumental contribution to the delivered image quality varied between 0.45-0.5 arcseconds (estimated from the optical contribution and diffusion of photo-electrons within the LSST Camera), and per-night median ellipticity varied between 0.06-0.09.
On-sky testing during the week included further studies of the Active Optics System (AOS) control loop running for periods of an hour or more at fixed telescope pointing, rotator angle, and filter to characterize the optical convergence and stability of the system. Specific tests included runs with a reduced number of active degrees of freedom to evaluate the impact of potential degeneracies, as well as runs to evaluate a range of proportional and integral coefficients for the proportional–integral–derivative controller. The team directly compared performance with various configurations for observations at fixed telescope pointing versus rapidly scanning the sky during pre-LSST observations.
Rubin held an intensive 3-day workshop on the AOS that included focused efforts on
- control loop characterization,
- validation and comparison of wavefront estimation algorithms,
- accuracy of predicting PSF sizes and shapes from the corner wavefront sensors and with intra-, extra-, and in-focus triplets for the science array,
- variations of optical aberrations across the field of view, their dependence on telescope elevation and camera rotator angle, and their stability across time, and
- computational performance of the wavefront estimation pipeline at the summit.
Findings and discussions from the workshop are informing priorities and next steps for both offline analysis and software development as well as on-sky engineering.
The team continues daytime engineering at the summit to improve control of the thermal environment. As part of broader efforts to reduce heat sources near to the optical path, the team installed a thermal shroud around the hexapods used for fine-positioning of the LSST Camera. Work continues inside the cell of the primary-tertiary (M1M3) mirror cell to fine-tune the air circulation and thereby reduce thermal gradients in the glass during diurnal heating and cooling cycles. The team is also preparing to test another set of dome louvers.
The next week is expected to include a blend of pre-LSST observations and dedicated on-sky engineering tests.
The currently installed filter set is ugriz. The next filter swap to exchange u for y is planned for 21 April.