X-33 – the middle path?

I was walking under a beached whale, and inside it, and around, the dangling entrails smacking me in the face, an amateur mistake. I should have known how to move around flight hardware carefully. It was early 1999, and the X-33 was taking shape. With its internal rib-like frame, platforms, curved supports, cables holding all the innards together, and the wiring everywhere like spaghetti, a beached whale is pretty much what it felt like. I had only been on these kinds of advanced projects for a short while, part-time. My day job was still mainly in the Space Shuttle, and the two worlds could not be further apart. Our team’s task was to join them – the present and the future.

Much has been written about the X-33, its cause of death laid bare in many an autopsy, us experts standing around the carcass. Even ending up with parts of the carcass in a drawer. The project bit off more than it could chew. It had too many technologies not ready for prime time, not to mention the political ailments, all explaining its demise. All quite valid. But not enough has been said about why its design and its contractual and teaming arrangements were all so full of promise at the time.

If you are designing a whole reusable launch vehicle, or just part, a booster, an orbiter, a first fork in the road is the form you believe (at the time) follows from the function. Here, the function is to reuse. If you go with a high ratio of length to width, you have a Falcon 9 booster. If you add wings and things and the ratio is lower, you head toward shapes recalling the Space Shuttles. Early competitors to the X-33 did both, a vertical take-off and vertical landing on the one hand and a winged Shuttle-like shape on the other (revived by DARPA in the XSP program decades later). X-33 split the difference and went down the middle, initially at least.

Yet not the blend expected.

Why add separate wings when you can still return with a graceful landing like an airplane by placing all your tanks and innards into a triangular shape that could do double duty for lift. The “lifting body” would be right up Lockheed’s alley, from stealth technology and getting odd shapes to fly that look like they shouldn’t. This was more than a design, as the Skunkworks would be an approach to rethink process and practice, not just tanks, engines, and tiles.

In this sense, the promise of X-33 was non-technical as much as technology. The non-technical points are often lost in the historical record, a contract form to more appropriately align incentives years ahead of the successful format that led to commercial cargo to the ISS. The X-33 was about research, a non-operational demonstrator vehicle. On the technical side were things like robust metallic tile, eliminating an army of tile and blanket technicians as with the Shuttle, and all their necessary support (the higher cost by far). Less advertised, differential throttling would eliminate large actuators for the engine, and where there were actuators, these would be electric and smaller, not a plumber’s nightmare of hydraulics. A big one is that the X-33 would not have toxic fuels – an environmentally friendly step. Simplifying, the X-33’s shape would do double duty, eliminating large active aero-surfaces. The body was the wing. On gathering a team, the flatter organization, a firm fixed price, a bias for action, and a vision of primarily non-government customers (once the full-scale ship was built) rounded out expectations.

In practice, the wings that began as small and stubby got bigger. So much for no wings and no large aerosurfaces with more actuators. The ship got heavier, and the path to a larger, full-scale, orbitally capable vehicle in a single stage eventually seemed a bridge too far. It was too late to matter, but the composite tank materials technology that burst and sounded the end of the X-33 program would eventually work years later.

Too much of a good thing?

Forgotten in all this, individually, most early X-33 decisions did lean toward success. It was only later that it became apparent to almost everyone that too many good choices were the equivalent of wanting it all, fighting a war on too many fronts vs. picking our battles. Many good decisions, it would seem, can end up not adding up any more than many bad ones. Horizontal landing? Vertical landing? Legs and fins or wings and things? Or a spot in the middle? Looking back, the best decision-making analyses and models at the time were isolated when sophisticated or barely connected when simple.

Decades later, this does not have to be the case. While there is no replacement for learning by doing, trial and error, and experimental vehicles (and RUDs), the capability to understand decisions has come a long way thanks to know-how and capabilities we could only dream of in the late 1990s’. The temptation is to believe we could have made better decisions back then anyway, especially in retrospect, once all the information was gathered in one place. Yet we now have AI/artificial intelligence to help in information search, finance, pattern recognition, and knowing and predicting the shape of our private lives (permission granted or not).

Perhaps one day, the complex decisions we face with too many variables to count will all be mixed and combined, and the AI will let us know this set of technology, but not too much, for this result. If you are game. Do you want a reusable launch vehicle making a few daily flights like an airplane? Decisions, decisions. And that’s a topic for more ahead – when we soon have to let the AI’s do what they do well, help us with the design, organize people, and decisions. And we’ll all enjoy the results of our good choices come Monday morning.

Also see:

• November 14, 1997: X-33 Launch Pad Construction Began
• X-33/VentureStar – What really happened by Chris Bergen, 2006