NASA's Roman telescope holds its pointing to within a hundred-thousandth of a degree

"Roman doesn't have a separate guider instrument, like other space telescopes do," said Begoña Vila, Roman's guiding instrument systems lead at NASA's Goddard Space Flight Center ^[9]. Instead, engineers set aside a small section of each of the Wide Field Instrument's 18 detectors to watch a guide star whose position is already known with high precision ^[3]. The attitude control system points the spacecraft at the right patch of sky. The fine-guidance system then tracks those reference stars and supplies the corrections that hold the telescope still during an exposure ^[4]. "Every Roman observation relies on its ability to stay precisely pointed at the correct region of space long enough to collect an image, which can take from minutes to hours for a deep exposure," Vila said ^[5].

The September tests put a number on that. "Our tests confirmed that we are able to keep the observatory very stable for science operations: better than 1/100,000 of a degree for half an hour at a time for Wide Field Instrument observations or for eight hours at a time for Coronagraph Instrument observations, which take much longer," Vila said ^[7]. A hundred-thousandth of a degree is 0.036 arcseconds ^[1]. The coronagraph hold is 16 times as long as the wide-field one ^[2]. The length matters because even a tiny vibration or pointing error can let starlight leak into a coronagraph image and overwhelm the dim planets it is meant to find ^[14]. The coronagraph also stabilises itself on top of what the spacecraft does. "Roman's coronagraph also has its own internal stability process, making it much more stable even than the Wide Field Instrument," Vila said ^[13].

NASA compares the current stability to keeping a laser on a dime from about 150 miles away. Engineers expect to improve it with further tuning, and the goal would stretch that distance to about 230 miles ^[8]. The dime stays the same size, so the goal implies an angular error about 65 percent of today's, roughly a third tighter ^[3].

The coronagraph woke on Sept. 1 ^[15] and took in cosmic light for the first time on Sept. 22 ^[2]. In joint runs with the guidance system on Sept. 22 and 27 ^[11], it captured stars in the Large Magellanic Cloud. NASA says those images confirm the instrument can produce focused images ^[16].

The thing this doesn't tell you is whether the coronagraph can do its job. NASA designed it to demonstrate technologies that suppress a star's light so that much fainter planets and dusty disks around nearby stars can be studied ^[12]. A sharp image of a star shows that the optics focus and the pointing holds. How far the starlight can be suppressed is a separate measurement, and the release does not report one. I think the pointing result is the firmer of the two: it is a measured bound, met in flight, for the eight-hour stretches coronagraph work takes ^[7]. The coronagraph result is a first-light check that the instrument forms images ^[16].

One part of the guidance design is still waiting for its test. Roman will try a form of guidance that NASA says has never been used this way before ^[10]. "Instead of tracking only a star's point-like appearance, it will guide on detailed wavelength patterns called spectra," Vila said ^[17]. Roman already collects spectra for its science, so engineers can use the same data to find and hold the telescope's position ^[19]. "We are looking forward to validating this spectral guiding mode in the coming weeks," she said ^[18].