The Spaceship That Almost Landed

Most reports of Saturday morning’s flight by SpaceX to the International Space Station note that the primary mission was successful, but that the company “failed” to stick the landing. For example, Andy Pasztor at the Wall Street Journal once again lived up to his reputation of a willing (and often unfair) basher of the company, and initially used the word “botched” to describe it in the lede graf. ("It now says "spoiled," though the headline still says the company "stumbles," but here is a version of the original from Marketwatch.) Though, to be fair, the rest of the piece seems well reported.

But to focus on what went wrong is to ignore the many other things that went astonishingly right, and just how close the company came this time to achieving their long-time goal of recovering the first stage of the Falcon rocket. As with the too-much-bewailed loss of its test vehicle in Texas this past summer (and the destruction of SpaceShipTwo this past Halloween), it is important to understand that this was a test flight and, as with any test flight, they were pushing the envelope to find out how the system worked in practice as opposed to theory. It was an experiment, and the only failed experiment is one in which nothing is learned.

In the '40s, '50s and '60s, the NACA and, later, NASA had many flight-test programs to learn how to do new things, gradually expanding performance envelopes, often breaking airplanes (and killing pilots), but with the advent of the Shuttle and reliance (some would say over-reliance) on computer simulations, the agency has developed a tendency to instead spend months and years in analysis, without flying. The very first (delayed) flight of the Shuttle was not only all the way to orbit, but with a crew. The X-34B program in the '90s was in fact canceled before it ever flew, even though the vehicle was complete, because NASA, in its risk aversion after the loss of a planetary probe, decided in retrospect that it shouldn’t be attempted without sufficient redundancy in key systems.

But SpaceX has picked up where the government left off, building, flying, testing and learning, then rebuilding and reflying. Its first three Falcon 1 flights all failed to deliver their payloads, but with each failure, they learned a lesson that was applied to the next flight. On the fourth flight, they succeeded, as on the fifth. The larger Falcon 9, based on the lessons learned from the Falcon 1 program, has never had a primary-mission failure. Now, they’re flight testing again, with the first stage, but only after it has completed its crucial job of getting the second stage and payload on their way to orbit.

The company understands that the key to reducing launch costs is to minimize the degree to which hardware is thrown way. The first stage, with its nine engines, is reportedly about 70% of the total vehicle cost, and the easiest part to recover, in theory. They’ve been attempting it from the beginning, but it had always broken up on entry. With the realization they’d have to sacrifice some performance in order to save enough propellant to safely enter and land, they increased the size of the vehicle to allow it to carry more. Since then, they have been slowly approaching a successful recovery, first with soft landings in the water, which still destroyed the vehicles, and then, this past weekend, with the first attempt at a precision landing on a hard surface.