After a scrubbed launch attempt this past Tuesday, April 17, SpaceX today launched the largest, most powerful rocket in the world, Starship. The launch stack consisted of a Super Heavy booster first stage, known as Booster 7, and atop that the second stage/orbital vehicle, Starship SN24 prototype. The launch was a smashing success for about four minutes and to an altitude of about 127,000 feet before it became much less of a success as Starship failed to separate from the booster and then exploded. Oops.
Here’s the full video of the launch, and it’s notable to see just how strangely upbeat everyone involved seems to be even after the spacecraft “experienced a rapid unscheduled disassembly,” as SpaceX put it, soothing the pain with the fun of an overcomplicated euphamism:
They made a note that “everything after clearing the tower was icing on the cake,” and when the rocket started to spin uncontrollably and then exploded, everyone erupted in cheers, which is great that they’re so upbeat about everything and presumably accomplished their key goals, but it still feels a little odd to be that celebratory, but, you know, whatever makes them happy.
Here’s a clip of the explosion itself:
SpaceX Starship explodes after launch#SpaceX
— Crime With Bobby (@crimewithbobby) April 20, 2023
The Super Heavy booster is interesting in that instead of using a few large rocket engines, like a Saturn V moon rocket, it uses 33 Raptor engines firing all at once to generate 16 million pounds of thrust. It appears that six of these engines were not running, according to both video screengrabs:
#SpaceX looks like 6 engines were out pic.twitter.com/ImORSUXA8Z
— Miles O'Brien (@milesobrien) April 20, 2023
…and telemetry from the rocket, which seemed to show 5 engines out:
It seems that even with five or six engines out, everything was still within parameters, and I don’t expect that this was related to the eventual explosion. Synchronizing and getting many engines to work together is not a trivial achievement on its own; the failure to solve this is largely what doomed the Soviet moon rocket, the N1 in the late 1960s and early 1970s.
The rocket engines seemed to be working together, but clearly something else was up. The separation of Starship from the Super Heavy booster didn’t happen properly just before everything exploded, so perhaps that’s a clue as to what went wrong. I’m pretty certain there are many many smart people at SpaceX figuring all of this out as we speak.
While it’s easy, perhaps even fun, to roll our eyes and indulge in a bit of schadenfreude when reading tweets from Elon superfans like this one:
Good luck to those betting against Tesla and Elon Musk pic.twitter.com/6OZyTksFbo
— Whole Mars Catalog (@WholeMarsBlog) April 20, 2023
… I’m not sure that’s really all that useful. I’m not exactly the biggest fan of Elon Musk, but I’m a very big proponent of space exploration, so I’d have liked to have seen this thing make it to orbit. I’m sure that’s coming, and none of this is easy, so this kind of thing is to be expected. That’s pretty much exactly what NASA chief Bill Nelson had to say about it, too:
Congrats to @SpaceX on Starship’s first integrated flight test! Every great achievement throughout history has demanded some level of calculated risk, because with great risk comes great reward. Looking forward to all that SpaceX learns, to the next flight test—and beyond. https://t.co/ZYsh5VkxsA
— Bill Nelson (@SenBillNelson) April 20, 2023
Rockets explode sometimes. Happily, this one had no humans (that I’m aware of) on board, and hopefully the debris from the explosion has only damaged various cameras around the launchpad, though I have heard scattered reports about innocent car windows being shattered. In fact, thanks to a commenter, here’s a link to a poor Mopar minvan taking damage from debris:
Car getting pelted by debris from Starship pic.twitter.com/yI61CQ3NFz
— Latest in space (@latestinspace) April 20, 2023
There’s the real victim, here.
I’m sure all sorts of data is being gathered and studied so that the next launch may go a bit better. One important bit of data I think has been gathered, at least, and that’s confirmation of Elon’s rather blank dammit-everything-just-went-to-shit-but-I’m-in-public facial expression, as we can see here in two separate provocation incidents, the one on the right from today, and the left one from an earlier event:
So, that’s useful information to have.
Elon did seem upbeat about everything overall, though, tweeting:
Congrats @SpaceX team on an exciting test launch of Starship!
Learned a lot for next test launch in a few months. pic.twitter.com/gswdFut1dK
— Elon Musk (@elonmusk) April 20, 2023
Good luck next launch, SpaceX.
Godspeed, Jebediah Kerman. See you again when you inexplicably pop back up on the roster.
NASA Spaceflight is doing a livestream, discussing the van that you mentioned and the damage that it took after it got blasted by Starship. (Starts around 5:30 Central Time). PLEASE talk to them about it and Starship before you release an article. The rear end is smashed, but the van was drivable!
A bit of a shame to have it fail from something simple(ish) but no big deal.They’ll get it right in the end.
Something funny i have to admit: When they showed the closeup of the engines i had zero clue what i was looking at.I was wondering why there were a bunch of cheap LEDs onscreen XD
IMHO multiple engines is just begging for trouble. Anyone worth their salt with a multi-engine rating will tell you that more than one engine is less than optimal in terms of handling, simplicity, durability, reliability, etc.
Unless the engines are all replaced at the same time they will wear differently resulting in one producing more power than the other which complicates controllability a good deal. Contrary to popular belief extra engines generally do not provide redundancy, instead they increase your likelihood of engine failure relative to the number of engines. For example with 2 engines you have twice the likelihood of engine failure, with 3 engines you have 3 times the likelihood of engine failure, so on and so forth. With Rockets since you cannot have an inline push-pull setup (unlike with prop planes) you have to worry about asymmetric thrust from engine failure much like you do with engine failure in a traditional multi-engine plane.
The Falcon 9 has nine engines, so they’ve managed to deal with that many just fine. However, I guess scaling up to 33 is more than three(.6) times more complicated.
They do have one thing on their side that the Soviet engineers building the N1 didn’t have though, which is a place where they can test all the engines at once, while fitted to the vehicle. The Soviet engineers at the OKB-1 design bureau couldn’t get the funding to build an engine test stand big enough (Space X have just used their launch platform instead), and that’s been put forward as a key problem with the N1. The first time they fired up all the engines together, at full thrust, was on the first launch.
The Soviets launched four N1 rockets, and all of them failed, very spectacularly*, so my prediction is that SpaceX will need at least three before they get a mostly successful flight.
*-The second N1 exploded before it had even cleared the tower, and was probably the biggest rocket explosion in history. Starship might be bigger, but it was almost out of fuel when it exploded, so it was a smaller boom than the N1. Sorry Musky, but the N1 is still number one at exploding!
Honestly, the N1 was so cool and I could blather on about it all day.
I’d challenge a few points there. The first is that SpaceX already operates a 27-engine rocket in the Falcon Heavy, which is 5 for 5 in successful launches. Granted, it’s 3 F9s strapped together, so you don’t have all of the heat dissipation issues that you do with Superheavy’s tighter cluster.
Second, they do extensively test their engines on the ground at their facility in McGregor, TX, not just on flight day. 4-6 tests a day typically go on there.
Finally, the difference between F9 and Starship isn’t really the *number* of engines, but their complexity. The Merlin on F9 is a dead-simple kerolox gas generator cycle. It’s about the easiest option for a practical engine. Meanwhile, the Raptor on the Starship uses a Full-Flow Staged Combustion cycle, which is a new and very complicated (but much more efficient) option.
I’m definitely in the ‘They NEED a water deluge’ camp. Firing 33 engines straight into the concrete was just asking for trouble (which Elon admitted a while ago). They need a proper flame trench and water system in order to prevent debris from damaging the engines before they leave the ground.
Doh! I’d forgotten about the Falcon Heavy, good catch. I should have made my point better about testing, SpaceX test their engines individually, and also all at once on the vehicle on the ground. OKB-1 could only test the N1’s engines individually on the ground.
It’s not just the heat from clustered engines, it’s the vibrations as well. At least one of the N1 failures was due to a “water-hammer” effect smashing the plumbing when several engines were powered off at once.
Switching to new engines will definitely be adding to their complications, but at least those are a lot easier to test on the ground.
So, there are a few points here where airplane engines differ from rockets and a few completely separate reasons why multiple engines can be a good thing.
On the differences:
1. You’re not talking the difference between one engine and two, or even two and 4. You’re also not talking about an hour-long flight with multiple abort options. A rocket either reaches its destination or the mission is a failure and the entire cost of the vehicle and mission are lost. Going from the 2-5 engines of many older rockets to the 9+ that is common on newer ones means that you have engine-out capability almost from liftoff. You can still complete the mission, even if an engine fails.
2. Asymetric thrust is definitely a thing, but you don’t have nearly as large of a moment arm between the opposite ends of the diameter of a single stick rocket as you do on even a medium plane. Plus, those rockets weigh hundreds or even thousands of tons at launch and, as they ascend, throttling down is less and less of a problem because that weight drops rapidly.
3. Handling actually improves with multiple smaller engines because most modern rockets use thrust vector control, rather than fins for directional control. Especially if you’re building a reusable rocket with propulsive landing, having many smaller engines means quicker, more precise control, as opposed to fewer, larger ones.
In terms of factors unique to rockets and, especially reusable, propulsively-landed ones,
1. Rocket engines can only throttle so deeply, typically 30-50% of max power. That means that, once a rocket has burned 90%+ of its mass on ascent, it requires smaller engines to land propulsively. Falcon 9 usually lands on one of its 9 engines and, even then, it can’t hover. Hence the ‘hover slam’ where it decelerates appropriately to reach zero velocity at zero altitude. For Starship, this is even more extreme because…
2. ‘Newspace’ rockets frequently share common engines across all stages in order to simplify manufacturing. For Starship, this means that the engines don’t have to be small enough just to land the 33 engine booster. They also need to be small enough to land the 6-engine upper stage, based on the constraints of 1 engine.
3. Space is a limitation on rockets. Your mass is limited by how much thrust you can generate per square foot of area covered in rocket engines. Smaller engines to a better job here than larger ones. Starship currently has 33 and could, theoretically, go to about 37 engines. The only single-engine super-heavy vehicle seriously proposed (
4. Smaller engines are less prone to combustion instabilities, which make them difficult to start. A descending booster engine must be able to restart and burn 2-3 times, which favors smaller, more numerous engines.
Overall, it’s just a difference of scale and numbers. Would you propose a heavy-lift transport with one engine? Rockets take that an order of magnitude further. The only super-heavy rocket ever seriously proposed with a single engine was Sea Dragon, and they handwaved away the combustion instability issues that plagued the F1 engine (largest single-bell engines ever built) on the Saturn V. Once you go above a couple engines, you get into the realm where loss of one or two doesn’t scrub the mission. That is the area in which rockets will always operate.
I should also point out that, after a couple days of analysis, the fault was pretty clearly SpaceX’s decision to try to launch the Superheavy with nothing but high-temp concrete below the pad. Film analysis shows 10ft chunks of it blowing 300ft in the air, directly alongside the rocket and a 20ft deep crater dug under the launch pad. It’s pretty clear that the vehicle was impacted by this debris, causing the three engine failures before they even left the pad and, due to a burning, collapsing hydraulic system, the cascading failures over the next few minutes.
EDIT: I wrote most of this as a way to spend a mildly inhebrehated on a pleasant Friday evening on my back porch. Please let me know if it makes sense.
Big bada boom. Bada BIG boom.
Like the reference! Hate it blew.
When I was in college (aero and mech engr) in the late 70’s we had a guest lecturer from NASA. The one significant factoid I remember was that we were told was that NASA fully expected that 50% of the Apollo rockets would fail in some manner or another. This was why the later Apollo missions really didn’t have much new to do, as the earlier missions accomplished mission objectives and redundant actions were on each of the missions.
I’ve been in engineering product design for over 30 years on much less complex stuff. If you are not failing in development you are not ambitious enough. I’ve seen my designs fail 100’s of times in testing as they are refined before they hit production. If you innovate, you will see failures in testing. If you just put new badges and paint on existing products then call them new, then you won’t see any development failures. You have to take risks to really see success.
Ok, what was the cost of your failures? Was it part of the development phase?
Almost all the test failures happen during the development process so it hard to break out costs. Sometimes you spend a year designing and testing a product only to have the project cancelled. The systems I design parts for cost the end user $20,000 to $30,000 which is literally orders or magnitude less than a rocket. The closer you get to production the more expensive it is to make the parts you are testing. Prototype parts take a lot of person-hours to design, then fabricate, and finally test. By the time you get to your final pre-production build of units made from the final manufacturing processes but that still cannot be sold to customers you can literally have spent a million dollars on parts that broke during testing if you count the labor to design, fab and test. The current philosophy is to iterate quick and fail often and early to get to a final design. The other thing to keep in mind is for the stuff I work on the safety factors are 4+ meaning our parts and assemblies will survive and are tested to 4 times the expected loads. Rockets are more like 1.25 or less for safety factors because a rocket with a 4X safety factor would be too heavy to leave the surface.
what was that car doing that close to the launch pad (and that was shit being blown across the ground by rocket blast at launch, not debris from the explosion, which happened well out over the gulf)
That was the NASA Spaceflight (news organization that covers spaceflight) press van. You’ll notice that the whole field is littered with wrecked equipment. It was a designated space inside the exclusion zone for automated cameras. NSF went whole-hog and rigged up that old van with several remote-controlled cameras, along with solar and batteries to power them and high-bandwidth communications. It was expected that it could be damaged.