"I’ve spoken with plenty of the earliest engineers who worked at SpaceX, and almost all of them have noted that Musk tackles the hardest engineering problems first. For Mars, there will be so many logistical things to make it all work, from power on the surface to scratching out a living to adapting to its extreme climate. But Musk believes that the initial, hardest step is building a reusable, orbital Starship to get people and tons of stuff to Mars. So he is focused on that.
He knows he won’t get Starship right at first. He employs some of the smartest engineers on this planet, and they’re still, in many ways, fumbling toward solutions for the extremely hard problem of getting a super-large vehicle out of Earth’s gravity well into orbit—then to land it and fly it again. Musk has come to believe the only way to realistically achieve this is through trial and error, by iterating closer and closer to the right design."
I always found the dichotomy of Soviet culture interesting - that engineers were the most pragmatic I've ever seen, and the politicians were just...fucking idiots.
The best example being Korolev, possibly the greatest rocket designer in history (certainly up there with Von Braun and Xuesen, and in my opinion more impressive), who died because he was sent to the Gulags by a bitter politician (this is an incredibly short edit of the full story, but that is a factual statement).
Korolev built working rockets the US said were theoretically impossible, and he did it because he couldn't afford to build the rockets the way the US were doing it. So he got resourceful.
Had he not died in 1966, I suspect the Soviets would have beat the US to putting a man on the moon.
>> I always found the dichotomy of Soviet culture interesting - that engineers were the most pragmatic I've ever seen, and the politicians were just...fucking idiots.
I'm not familiar with that history, but I'm quite familiar with Corporate America. One thing I've seen is that when Engineering is great, it gives the company great momentum and allows the Management/"Business"/Strategy/etc functions to lag and get lazy. Good Engineering momentum essentially produces enough residual value for others to leech off from.
Which may be an insight on how SpaceX/Tesla avoid some of those traditional business issues- the business is this case is a means to an engineering end, not a means to perpetuate the power/wealth/comfort of management.
> that engineers were the most pragmatic I've ever seen, and the politicians were just...fucking idiots.
The engineers would have big problems if they spaceship doesn't work. The politicians are not idiots, its simply that they don't have the same goals as the engineers so the politicians seem like idiots to engineers.
> Had he not died in 1966, I suspect the Soviets would have beat the US to putting a man on the moon.
This is an false take. Many people have looked at this and it simply doesn't hold up. The resources they were investing was not big enough and having korolev couldn't have solved all the many problems they would have still faced.
They had some great engine technology but its a long way from there to the moon.
Stalin spent the latter half of his career killing all the talented politicians, whereas the engineers got left alone. Then the not-so-talented politicians that got left over (some of whom were not so bad) had to fight against a culture of fear, compulsive ass-covering, and magical thinking.
The only fucking idiot among top Soviet politicians was Gorbachev who sold one of the world's biggest economic, scientific and military powers just for 30 pieces of silver. His predecessors were no idiots at all. They are depicted as complete morons by the western propaganda - it's true. But propaganda and reality are two completely different things. I would advise you to learn history instead of blindly believing the yellow press and political articles on Wikipedia.
> On 23 March 1938, Valentin Glushko was arrested. To reduce his charges he denounced Korolev, which resulted in his arrest on 7 June and a sentence of ten years forced labour. From that day on, the two men were bitterly opposed to each other.
So as you can see he was not sent to gulag out of spite by Stalin but out of envy and professional competition by his fellow rocket engineer.
While in gulag he was not tortured or sent to gold mines. He and other scientists continued their work in relatively comfortable conditions in so called "sharashka": https://en.m.wikipedia.org/wiki/Sharashka
I'm not saying that Stalin was a saint or a winged seraph. But definitely he was anyone but an idiot.
Also Korolev did not died because he was sent to the Gulags. He died from a chronic disease from which he started suffering long before he was imprisoned.
An engineer was imprisoned because another engineer was spiteful. You don't consider this a dumb political decision? I consider the default judgment of "everyone goes to the gulags" to be shitty governance, and also quite stupid.
My understanding is Korolev's health greatly declined because of his time while imprisoned, and had he never had to endure it he likely would have lived much longer.
I must point out that username ilyich is a transliterated name likely referring to Lenin – real name was Vladimir Ilyich Ulyanov. So ilyich seems to be a fan of USSR.
I read Stephen Kotkin's excellent three part series on Stalin this summer and Kotkin (who is by no means a Soviet apologist; he's a conservative) thinks Stalin was a political genius, a modern Machiavelli. For example, in this interview [1] he explains how Stalin's administration was what he considers a "transcendental work of art" while also explaining how it was psychotic and evil.
I highly recommend Kotkin's books. You can jump into the second, Stalin: Waiting for Hitler, 1929-1941 if you want to learn more about him as a leader.
Gorbachev saw an totally undemocratic nation that had long flattened out in terms growth and innovation and wanted to make it democratic, he didn't realize that all these rich people in the society had more interest in keeping the system in their control and they had no interest in his vision.
> They are depicted as complete morons by the western propaganda - it's true. But propaganda and reality are two completely different things.
They were not idiots, but their intensives didn't align with the goals of the engineers.
It's well known that the Soviets advanced in the Space race because they neglected people's lives. Cosmonauts were launched in half-baked, hastily assembled spacecraft. It's a gruesome example of iteration.
The quote that explains why it's important to build one Starship a week:
> “A high production rate solves many ills,” he said. “If you have a high production rate, you have a high iteration rate. For pretty much any technology whatsoever, the progress is a function of how many iterations do you have, and how much progress do you make between each iteration. If you have a high production rate then you have many iterations. You can make progress from one to the next.”
It turns out orbital launchers scale rather well with size.
The rocket is already 90+% propellant (fuel & oxydizer), with the few percent left being the engines, structure and payload. So if you want the rocket to have a meaningful payload, it needs to be big.
For example the Falcon 9 weights 350 tons and can place 22 tons to low earth orbit (LEO). So about 6% of initial launcher weight ends up in orbit.
If you want to place the 100+ tons estimated for a reasonable Mars trip, your launcher will be either huge or you will need to do a lot of launches and assemble your ship in orbit with all the overhed that requires.
Starship also aims to be fully reusable, which will eat to your payload yet again but provides a huge benefit of not throwing away any part of your rocket, just burning a lot of propellant, which is dirt cheap compared to rocket hardware (estimates say about 5% of conventional rocket launch is fuel, the less is hardware that you normally simply crash into the ocean after use).
So in short - big payload -> big rocket. Reusable rocket with big payload -> humungous rocket that is really really cheap to operate.
Efficiencies of scale. Too small and you can't pack enough food for the trip or proper redundancies, still too small and you can't build the base fast enough, don't bring enough stuff on the trip and you need to make more trips..which are time limited due to the rendezvous. You end up inclined to 'as big as possible'
Basically to drive down the cost of payload mass as much as possible. First you have to build a lot of rocket to get anything up into space. If it only launches 1kg of payload, that's going to be a very expensive kilogram. As you continue to scale the rocket up, assuming you can design for a decent payload ratio, the amount of cargo capacity increases and the cost per kilogram drops.
People need to live in it. A 1 Kg probe can take some pictures, the tiny Apollo Lunar Module was 16,400Kg and only designed to support 2 people for 75 hours. Meanwhile Mars (3.72 m/s²) has significantly more gravity than the Moon (1.62 m/s²), making a trip back from the surface even more difficult.
Launch vehicles tend to scale payload capacity at a greater rate than dry mass. So a small launch vehicle like RocketLab's Electron might have a wet mass of 12,000kg at launch and an LEO payload capacity of 225kg (225/12000 ~ 1.8% payload), the SpaceX Falcon9 has a wet mass of 550,000kg and an LEO payload capacity of 22,800kg (22800/550000 ~ 4.1% payload).
This just tends to mean that, very generally and with exceptions, larger vehicles can get mass into space with less fuel.
A Falcon 9 has a LEO payload capacity of 22,800kg in expendable mode, something SpaceX does not like to do. The largest payload mass they've carried to LEO was Starlink-2 at approximately 15,600kg (60 satellites at 227 kg each).
The rocket equation [0] disagrees: "In what has been called 'the tyranny of the rocket equation', there is a limit to the amount of payload that the rocket can carry, as higher amounts of propellant increment the overall weight, and thus also increase the fuel consumption"
The pure, theoretical rocket equation still scales linearly, the tyranny lies in the hard limit of how much mass you can theoretically accelerate out of a gravity well per unit off fuel.
But bigger rockets can get closer to that theoretical optimum because some dead weight components remain fixed-size and because some masses scale with surface instead of volume. Rockets can scale better than linearly with size, but they will never exceed the rocket equation.
I'm having a mental blockage on this, as you can always launch 10 rockets at the same time and get 10 times the payload. Having those 10 rockets strapped together should not change the payload of them flying independently.
What am I missing here -- I'm sure I'm overlooking something.
If you strap 10 rockets together, you don't need to keep all 10 copies of some components. You don't need 10 computers, or ten emergency abort systems, or ten communications systems, etc. One large tank will have less mass/volume than 10 small ones, etc.
Interestingly there are couple rockets that are close to "strapping couple rockets together" for various reasons:
The R7 rocket
- mostly to avoid engine ignition in flight & due to having a lot of nozzles you need to fit on a rocket
- this was then kept for Soyuz even after an additional stage has been added that is started in flight
- Soyuz 2V uses a more efficient engine & uses only the central stage
The proton Rocket
- the central tank has the maximum diameter you can ship via rail from the factory to Baykonur
- by mounting smaller tanks around it that are shipped separately you can avoid building an overly long rocket, that could be fragile and unstable
Saturn 1/1B
- basically a stop gap using existing Jupiter and Redstone rocket tankage tooling
- strap 8 Redstone tanks around 1 Jupiter tank and you get the S-I first stage
Delta IV Heavy/Falcon Heavy
- you have a rocket that can launch by itself with smaller payload or by strapping 3 first stages together can launch a bigger payload
- better economies of scale as you can doe more with a single rocket design instead of maintaining 2 separate one (big and small)
- in Falcon Heavy case you can also save all the first stages from more demanding trajectories where you would otherwise have to expend the regular F9 first stage
OTRAG
- make dirt cheap and as simple as possible standardized "rocket tubes"
- strap a lot of them together
- fire and jettison in the right order to achieve orbit (check your staging! ;-) )
While this does work, the Falcon Heavy has convinced Elon not to go into that direction again. Its actually not that simple to strap these together and just making a bigger tank is actually easier.
Poorly for the current market using a FH makes sense, but to break to another level, building one big rocket is better. Both the US Saturn 5 and the Soviet N1 were big single tube vehicles.
Correct. You can (super¹)linearly increase launch mass by adding fuel. The rocket equation hits when you're trying to go farther, rather than more massive.
¹ Superlinear because tank mass scales with a lower exponent than tank volume.
You aren't missing anything, the comment you are responding to is confusing the per unit of fuel payload limit to be the total limit of the rocket. You can add more engines/ use bigger ones which would allow for more fuel and thus more payload( in theory at least).
That's correct so far. Now think further: a single, larger rocket has the advantage of leaving away the tank walls that would be in the inside now. Less mass needed for the tank walls (the outer walls need to become stronger though), more available for payload, extra fuel and oxidizer or extra stuff like things needed for reusability.
This makes total sense to me. If you can solve the hardest problems, it is reasonable to expect that you can solve the rest as well. If you fail at solving the hardest problems, the system will not work anyway.
The science, then, lies in identifying the hardest problems. The art lies in identifying people who can solve these problems.
No, not in aerospace. The common approach is to spend years designing, simulating, and then actual tests come at the very end, and sometimes you don't even test because you can't or it would be too expensive. That's what Boeing are doing with their own rockets; Mars missions until now didn't really go through an iterative process; satellites are usually manufactured only by one and they better work once in orbit.
SpaceX built an incredible custom simulator in order to design the methane engines used in Starship[1]. Right now they are experimenting with manufacturing techniques, which is much less amenable to simulation
Elon has stressed this. Design is easy, building is hard, building at scale is harder. Thus the reason he iterates on the design is to make each design iteration also an iteration in the manufacturability, not just the functional design.
You're telling me we went to the moon on the first try? We went into space on the first try? The Wright brothers were successful on their first flight?
Simulation is not the same as trial and error. There's no such thing as a perfect simulation, and imperfect simulations without real world testing and iteration result in disasters. There are many examples, but the Mars probe that crashed because of a unit conversion error comes to mind.
You're not differentiating between an iterative design process, during which failures are expected, and a monolithic one where the end product is final. When Starships blow up, Musk can build new ones and try again. When the probe crashed, that was the end of the mission and there was no more Mars Climate Orbiter. It was a one-off. This is a common problem in spaceflight.
The Webb Space Telescope is another example. If it eventually does launch and something goes wrong, we'll have paid >$10 billion for absolutely nothing. These aren't trial and error because the designs are never improved based on the errors. They're just scrapped. SpaceX works in a fundamentally different way.
They are not that different from other space startups.
> "We succeeded in launching the rocket," Zhang told the media. "The experience we gained from evaluating the rocket's flight conditions will help us remodel the rocket as well as advance new rocket research and development."
When's the last time that someone tested a series of satellites (or launch vehicles), with a number of up-front failures? Prior to SpaceX, of course. To me it looks like that sort of iterative trial-and-error design process died with the end of the space race.
For instance, looking at LandSpace - that quote relates to the launch of Zhuque-1, of which there has never been a repeat despite a maiden flight in 2018, and Wikipedia says that a second flight ever happening is doubtful. The company's second launch (and probably any after that, too) will likely be with a completely different, liquid-fuelled design.
THAT's what failure normally looks like in aerospace. Whoops, we fucked up, and so now that design never flies again, or the company goes bankrupt, or everyone panics and goes through a massive design review because everything was supposed to be perfect and now obviously it wasn't.
You make it sound that everything that may involve an "error", however small (that is, absolutely everything) is "trial and error". This is not what most people mean - I argue that interpretation would make the whole concept meaningless.
In this context, "trial and error" means SpaceX builds things out and tests them in real-life not virtually (as opposed to others that tend to validate things virtually for much longer before actually building stuff).
You jest, but he actually does it. Executing the right idea is the value, not the idea.
Everyone else is cush and doesn't want to rock the boat (SLS, legacy automaker leadership, etc). For those, innovation has no incentive. "Innovator's dilemma" [1] and all that jazz.
Disclaimer: Not a huge Musk fan anymore, "world's least worst billionaire", but credit where credit due. I don't endorse sweat shops, but he has teached a cohort to yearn for the vast and endless sea. He is a salesperson first, and he is selling the dream.
Let's be objective: SpaceX has radically lowered the cost to get to Earth orbit, and commands a majority of the worldwide launch market. They are simply more efficient than other launch providers, so much so that they are branching out into global communications (StarLink) to capture more revenue (global comms is 10x the market size of the launch market). They don't need to go to the Moon to succeed. There is no market to get to the Moon unless nation states make one. Musk is building his own market for Mars logistics demand.
> SpaceX's market share increased rapidly. In 2016, SpaceX had 30% global market share for newly awarded commercial launch contracts, in 2017 the market share reached 45%, and 65% in 2018.
Tesla, according to the opinion of a single auto industry expert, has a 7 year head start against legacy automakers [1]. They sell every vehicle they build, and combined, legacy automakers have shed almost 40k jobs in the last year as they attempt to transition to building EVs to compete against Tesla.
A reasonable person might conclude these metrics indicate an achievement of some level of success.
I don't disagree that SpaceX has been successful. I just disagree that this trial and error methodology is some radical new thinking from Musk—it's not.
Nobody says he has invented trial end error, but every single person that worked in aerospace would tell you that they simply don't work like the other companies in many ways.
And with the Starship, they have gone to an extreme that even for SpaceX seems crazy.
There are tons and tons of interviews and podcasts with experienced people who worked there. Listing to those that came over from NASA is specially interesting.
"Just iterate, baby"
"I’ve spoken with plenty of the earliest engineers who worked at SpaceX, and almost all of them have noted that Musk tackles the hardest engineering problems first. For Mars, there will be so many logistical things to make it all work, from power on the surface to scratching out a living to adapting to its extreme climate. But Musk believes that the initial, hardest step is building a reusable, orbital Starship to get people and tons of stuff to Mars. So he is focused on that.
He knows he won’t get Starship right at first. He employs some of the smartest engineers on this planet, and they’re still, in many ways, fumbling toward solutions for the extremely hard problem of getting a super-large vehicle out of Earth’s gravity well into orbit—then to land it and fly it again. Musk has come to believe the only way to realistically achieve this is through trial and error, by iterating closer and closer to the right design."