Commercial Rockets

XS-1 / PHANTOM EXPRESS — USA

This was envisioned and funded as something just for military use, but if it had worked, there’s no way it wouldn’t be sold commercially and be a hit. And if that had happened, a lot of exciting little New Space startups might have found themselves crushed by big aerospace, just when they thought they were about to have their moment.

The XS-1 program started with a mandate from DARPA: take all the lessons learned from the Space Shuttle and later projects, and make a spaceplane or something that can be what the Shuttle failed to be: rapidly and cheaply reusable. Build something that can take off without needing a fancy launchpad complex, land on pavement, and launch again the next day.

Various small companies submitted competing designs for the XS-1, including Scaled Composites / Virgin in partnership with Northrop Grumman, but DARPA gave the contract to Boeing. Their design was called the Phantom Express. It consists of a rocket plane with a single hydrogen-burning engine, which will take off vertically from a transporter-erector-launcher vehicle, and then land as a glider, as the Shuttle did. But it’s an inversion of the Shuttle scheme: instead of the plane going to space after piggybacking on big boosters, the plane is the booster, with a small second-stage rocket riding on its back. The plane is smaller than the Shuttle orbiter, and lighter because it has two fewer engines and most of its interior is just a hydrogen tank. It would reach a speed of 3 km/s, before releasing the second stage at the edge of space. This is faster than most first-stage boosters can go before separation, and if achievable would keep the cost down by minimizing the size of the expendable second stage. It’s fast enough so that reentry heat protection is an issue, but not so fast that it requires anything cumbersome or heavy to solve that problem. (A secondary use for the plane would be to explore hypersonic flight in general, and in this mode, with no payload on its back, one of the requirements was that it could reach mach ten — nearly 3.5 km/s. That’s around half of orbital reentry speed.)

The second stages might be made by different companies with different sizes and capacities. They aren’t constrained by any particularly strict compatibility requirements, as long as they’re small. Vector Launch was apparently one company that got some funds to make a second stage.

The engine was being built by Aerojet Rocketdyne. It’s called the AR-22 and is derived from the RS-25, the Space Shuttle Main Engine. They’ve made lots of changes to enable it to do more flights without refurbishment, and to make it easier to work on once it needs some. It will still use a lot of the old parts. And the plane would make the engine easy to access, apparently by mounting it on a big hinge. This motor is plenty big, and should have thrust to spare, even if downtuned to reduce wear.

The hope was that the cost of reflying would be only $5 million, including the second stage. And though this low price tag would certainly be a hit if commercialized, for the military its biggest advantage would be not the low cost, but the fast turnaround. Because the vehicle which rolls it out of the assembly building also acts as its launchpad, it doesn’t have to wait for any launch complex to have an opening in its busy schedule. DARPA hoped that they would be able to demonstrate this by performing ten flights in ten days — a schedule more ambitious than even Elon Musk would dare say out loud.

And who knows, maybe someday they could make a reusable second stage. It might end up looking like a micro-shuttle on top of a mini-shuttle.

But woops — in early 2020 Boeing announced that they are ditching the Phantom Express. Apparently the challenge was just too difficult. And in response, the Pentagon decided to end the Experimental Spaceplane Program... for now. The odd thing is, DARPA insists that there was no technical showstopper — that the project remains feasible. This leaves people puzzled over why Boeing quit. They have not enlightened us.

This abandoning of the project is a shame, because even if the extreme goal of ten launches in ten days was not achievable, the basic concept of a spaceplane as a reusable booster stage is a good one, and I think someday it could be a success in some form. Several small companies have at times announced plans to attempt building such a plane — Dawn Aerospace of New Zealand and The Netherlands, for instance, though theirs is much less ambitious in terms of speed, only hoping to reach mach four. I hope one of them succeeds.

And I hope that someone will also find a use for the revised AR-22 hydrogen engine, which was the one part of the Phantom Express project which successfully built usable hardware. Aerojet Rocketdyne proved that on the ground at least, that engine could indeed perform ten full-length burns over ten days.

One company which had submitted a competing bid was Masten, a small outfit which has yet to build anything that qualifies for its own entry on this page, though they hope to send out a lunar lander on a Falcon one of these days. Because Masten’s expertise is in vertical propulsive landings, they went with that approach, despite their booster having wings in order to glide back toward the landing area. Theirs would have had five small methane burning engines, and the design came out tiny compared to the Phantom Express despite having the same capacity. The Northrup/Virgin proposal was a twin motor spaceplane which in some renderings looks bigger than the Phantom Express. They didn’t say anything about its fuel or engines.

As mentioned in the Vector article, one startup called Aevum is also now pursuing a plane as booster, but it’s not a spaceplane, just a fast jet. Because it’s autonomous and uses ordinary runways, they hope it will be able to do launches not just every day, but every three hours. We’ll check back on that ambition once they show an actual rocket to put under the plane.

Phantom Express: mass unknown, diam unknown, thrust ~1800 kN, imp ~4.4 km/s, type ZFh, payload 1.3 t to 1.8 t, cost hopefully $3M/t.