Deep Space Tuesdays · your week in deep space

From The Decisional Pass

Three weeks ago I closed this newsletter on two observatories. Swift had just caught a supermassive black hole shredding a star out in a galaxy's suburbs, and its principal investigator had to talk about what came next in the conditional, because the spacecraft wasn't doing science. A few paragraphs down, Roman had hydrazine in it and a launch date.

Both moved this week, in opposite directions.

Sunday morning, 7:26

On August 21, NASA and SpaceX finished the Flight Readiness Review for the Nancy Grace Roman Space Telescope. The same day, the observatory went inside the Falcon Heavy fairing.

Launch is no earlier than 7:26 a.m. EDT this Sunday, August 30, from LC-39A at Kennedy, bound for Sun-Earth L2.

Here's the number I had to read twice. Roman's formal commitment was to be ready to launch no later than May 2027. Sunday puts it nine months early. I've watched enough flagship science missions eat their own margin and then start on next year's that I didn't expect to type that sentence.

The optics are familiar. The mirror is 2.4 metres, same as Hubble's. What's behind it isn't: the Wide Field Instrument is a 300-megapixel near-infrared camera covering 0.281 square degrees in one shot, a field at least 100 times wider than Hubble's. Over a five-year primary mission it should produce roughly 20,000 terabytes.

Nobody gets a head start

That's the spec sheet. This is the part that got me.

Roman has no proprietary period. None. Every other big observatory hands the team that won the observing time a year of exclusive access to their own data before it goes public. That's the currency of the whole system: you write the proposal, you win the time, you get first crack at the paper.

Roman releases everything as it comes down, same day, to anybody.

The time itself was handed out in the open too. More than 100 science pitches in early 2023, then 70-plus technical white papers, then definition committees, then a time allocation committee that reported in April 2025. Three Core Community Surveys eat most of the first five years: 17 months imaging over 5,000 square degrees, 15 months on the galactic bulge that should turn up more than 1,000 wide-orbit planets by microlensing, the population Kepler and TESS structurally can't reach, and 6 months chasing Type Ia supernovae.

So Roman isn't really a telescope you book. It's a survey machine you query.

That's a different engineering problem, and the hard part isn't the mirror. It's the ground system. With no embargo, the archive is the mission. There's no quiet year to fix a soft calibration in private. Day one, everybody sees the same pixels, including the people who will tell you your pipeline is wrong.

I like that a lot, and it's the riskier choice, which is most of why I like it.

The one that can't be caught

The other half of the week is worse.

On August 19, NASA and Katalyst Space called off the plan to boost the Neil Gehrels Swift Observatory. Swift launched in 2004 on a two-year mission, has run 21 years, and its orbit is decaying faster than anyone modelled, because high solar activity puffed up the upper atmosphere and raised the drag. Without a boost, NASA expects it to reenter later this year.

The rescue was aggressive and I still think trying was right. NASA gave Katalyst $30 million in September 2025 to design, build and fly a servicing spacecraft in under a year. LINK launched July 3 on a Pegasus XL out of Kwajalein.

Then commissioning went sideways. Over the weekend of July 25, LINK lost two of its three reaction wheels, took a partial failure in its cold gas thruster system, and ended up tumbling at about 9 degrees per second, with comms going sporadic. Every other subsystem was fine.

What the team did next belongs in a training deck. With most of the attitude hardware gone, they used the electric propulsion thrusters meant for changing orbit to kill the spin, and had it down to 1.47 degrees per second by August 5. On August 11 they uploaded new attitude controllers written for the surviving configuration. That is genuinely good work under real pressure.

It's also the trap. Every second of despin spent propellant budgeted for the rendezvous. You can save the vehicle and still lose the mission, because what you burned saving it is what the mission needed later. I spent a lot of hours in JPL's Mars Yard learning that recovery is never free, and that margin you dip into during commissioning is margin somebody already planned to spend.

LINK will still attempt rendezvous and proximity operations near Swift. No capture, no boost, just approach. Sounds like a consolation prize, and it isn't, quite. There's very little flight data anywhere on closing in on a client that never agreed to be approached, carries no docking aids, and is sinking. That data is the product now.

Swift still burns up. Twenty-one years of gamma-ray burst alerts, and the last one is already logged somewhere.

References

• Teams Complete Flight Readiness Review for NASA's Roman Telescope — NASA Science, August 21, 2026
• NASA's Roman Telescope Enclosed in SpaceX Falcon Heavy Rocket Fairing — NASA Science, August 24, 2026
• Cutting-edge infrared space telescope Roman set to launch — ESA, August 25, 2026
• NASA Completes Nancy Grace Roman Space Telescope Construction — NASA/JPL, December 4, 2025
• Instruments and Capabilities — NASA Goddard / Roman Space Telescope
• Core Community Survey Definition — NASA Goddard / Roman Space Telescope
• NASA Updates Next Steps for Commercial Swift Boost Mission — NASA, August 19, 2026
• No-Go for Katalyst's Reboost of NASA's Swift Observatory — SpacePolicyOnline, August 19, 2026
• Katalyst Space and NASA abandon plans for Swift reboost — SpaceNews, August 19, 2026
• Commissioning Update for Spacecraft to Boost NASA's Swift — NASA Science, July 28, 2026
• LINK Spacecraft Recovery Progressing for NASA's Swift Boost — NASA Science, August 6, 2026

Forwarded this? Get The Decisional Pass in your inbox: