Relativity Space raises $35M for a new process to build and fly rockets
arstechnica.com
arstechnica.com
But going to their website[1] this launch market does not seem to be the company's main focus. They want to develop the technology to build currently very complicated objects with two orders of magnitude lower parts count. This is desired so that the products can be more easily built on Mars. I can see that as a long term profitable outcome with applications on Earth along the way.
From their mission statement[2]:
In the early days of settlement, there will be few people living on Mars. Intelligent automation and lightweight, compact 3D printing are fundamental technologies needed to quickly build a new society with scarce resources - and the most scalable means to get back home.
An ambitious goal (with an incongruous phrase tacked on the end??).
[1]https://www.relativityspace.com/home [2]https://www.relativityspace.com/mission/
Furthermore, to develop rockets on Mars you'd also need a substantial supporting industry (metallurgy, testing, etc.).
It's getting to that 100 million population with supporting infrastructure to build things like CPU's that's the problem. Not simply building rocket engines which would be relatively easy and relatively cheap to ship.
A country like the US is not even self-sufficient, although we could be. The consequence would be a reduction in productivity and quality of life. Now, Mars also has much higher living costs, although the costs are definitely prone to economies of scale. So a smaller population would have to work more and have a lower quality of life to allow life on Mars. At some point, you just can't keep up with the life support and maintenance required to live on Mars, so the population below some point cannot be self-sustaining (even if you have all the tools). You want as many people as possible to allow economies of scale and productivity improvements from specialization to take hold.
For this reason, I think Mars isn't going to prioritise self-dependence. A civilisation which could thrive if the Earth stopped existing is a romantic ideal, but like most places, the people will prioritise their prosperity. Producing CPUs on Mars will likely be very low down the list; they're light enough to cheaply bring from Earth, and have a huge supply chain. Elon Musk likes to joke about pizzerias on Mars, but I think they'd be much higher up the list, if people want to eat anything other than canned food.
And as far as Mars being cut off, certainly Mars will want to trade with Earth for as long as possible. Self-sustaining will hopefully remain theoretical. That doesn't mean they won't have the capability to do so if necessary.
Add to that police, a justice system, politicians and lawyers. Not just miners, farmers, plumbers, researchers, and architects. Basically, everyone needed to make every individual component physical and social of society, plus all the people to support them and support the people that support them. And beyond that people to support children born with physical and mental handicaps.
PS: In the very short term you can get away with an unbalanced population and hyper educated generalists, but that means an inherent dependency on Earth. AKA you need to build a society supportable by average people not just select highly educated, healthy, high IQ, elite. Further, this also removes things like entertainers.
PPS: Now if you want some form of Science fiction dystopia where less productive people are pushed out of an airlock then the numbers do shrink somewhat. But, that's going to make recruitment a lot harder.
I'm reminded of software rewrite efforts. Big software projects never get designed from scratch and implemented, they evolve by accretion. Somebody builds something halfway useful, they get capital (or volunteers) to help improve it, and eventually you've organized the planet's information and made it universally accessible and useful. In the process you acquire a shitload of technical debt so that it becomes very difficult to understand or change the lowest levels of the system.
Trying to rewrite large software projects with bug-for-bug compatibility and feature completeness always fails. You run into the same issues as trying to build a complex system from scratch.
But trying to rewrite a large software system where you're allowed to move the goalposts and sacrifice feature completeness often results in a system with an order of magnitude less code that's simpler for the majority of users as well. You get to prune out all the features you thought would be useful but turn out to be more trouble than they're worth, and you can often take simpler paths once you know where you're going.
Are there similar shortcuts available for human society? For example, could you race from farming villages to electricity by building wooden windwills that turn dynamos of simple copper wire loops? What would that change in how we approach industrialization, if you could start building machines without needing the labor of thousands of people and their supporting society? Are there similar shortcuts for semiconductors & integrated circuits, or do you really need the full might of a modern industrial society before you even have the raw materials?
The downsides are things like pest control becoming more important. On mars you can avoid pesticides by either not bringing them or killing them off by dropping the air pressure in a field to effectively zero and or dropping the temperature to -100C.
I can just see the “Mars First” and “Buy Martian” slogans now.
Complete printing of our rocket, Terran 1, reduces
vehicle part count from nearly 100,000 to under 1,000
components - and is the first step toward an entirely
autonomous factory
More info at their website: http://www.relativityspace.com
Now, printing might work for some other materials than carbon with similar or even higher strength to weight ratio. But AFAIK unproven so far. Would be a whole separate science project with lots of other applications than rockets.
Also, looks like the engine had some machining work done, maybe after the laser sintering aka 3D printing part, since the surface was smooth and shiny.
A larger challenge would probably be to cheaply produce the raw carbon fibre on Mars.
But yes, if that's what's meant, it's completely a no brainer. Nobody's hand winding tanks or airliner fuselages.
You'll need way too many rockets to cause an even measurable change on Earth's orbit in a million years. Try changing it to a few billions.
One concern:
> The Terran booster will hit what the company believes is a sweet spot between smaller rockets under development by Rocket Lab (and others) and the much larger Falcon 9 built by SpaceX.
The market is moving very quickly right now. I'm not sure the Falcon 9 will even be flying by the time this launches. Maybe that will mean more opportunity for them, it is hard to say. But it seems to me that all these small-rocket companies are banking on one specific idea: that the convenience or dedicated customer service they provide will exceed the pros of going with a more established low-cost provider like SpaceX. With BFR flying regularly in the same timeframe as this rocket gets its first launch, their only hope is that customers will not want to ride along as a 90th payload in a BFR but will want a dedicated ride themselves.
Or perhaps I am being close minded, perhaps the market will grow sufficiently such that SpaceX will not even be able to suck up all the new launches and new demand, and the small-rocket companies will thrive.
I just don't want to see them left in the dust. A $10M price tag is incredible (I mean, wow, that's low! for a dedicated launch!). But will that be the expensive option by the time their rocket is ready for use?
Two concerns, it turns out. Will these rockets be reusable? I'm guessing not, since the article didn't mention it and these days in the rocket world, if you're not reusable, then who cares? You will be left in the dust.
Edit: Wait a minute. I just went to their website. It says nothing about reusability either. But it exclaims how great it would be to build and fly rockets in days instead of years. Are they very specifically making a disposable rocket? If so, fuck this company and fuck everything about it. It's not the way of the future to build large objects and throw them away. It's evil in fact, to research something like this in the modern era with an attitude of 'fuck you' to the materials' sources where they are mined from the Earth.
I'll hold off on an actual opinion until I know more. But I sure hope Y Combinator didn't invest in a fake-new-space-actually-oldspace company. I hope this is better than it looks because it looks awful.
Edit 2: What does the success case of this startup look like? Will they ever research reusability? Or are they aiming to make 1,000 new rockets per year? That's the success case based on the financials (=$10B revenue/year) and it sounds gross. Without reusable rockets that whole thing is extremely wasteful.
The efficiencies of 3D printing things quickly are not correctly accounted for in the loss that the planet takes due to increased mining and wasteful dumping that would have to occur (and which the investors and company do not pay for).
Just before Russian rouble crashed, Soyuz 2 was estimated to cost USD $5.5 in materials and labour. With current rouble fx rate, it should be well lower. If they will get squeezed much, they may well remarket disposed ICBMs again. And at that point, I don't see any Western competition having a chance. Maybe Chinese will deliver bearable launch service on solids by then.
Pure material costs play a rather small part of considerations for a launch service provider. This is why Space X is not booked for years ahead, and launch services companies still line up for Proton that had 4 successive failures, and launch on ULA vehicles for twice the Space X cost.
Only companies for which launch cost is a genuine issue or which simply want to get to space by any means possible (startups, 3rd world countries) will be going for them.
The metal in a rocket isn't the most expensive part (it has to be well under 5% the weight of fuel). The expensive part is fabrication. So reusability is important iff fabrication is expensive. Reusability also requires adding extra launch weight for landing fuel and struts and guidance fins.
So it's too early to conclude that reusability is mandatory. For heavy repeated lifts to a standard orbit, maybe. For small lifts to custom orbits, maybe not. It all depends on the costs of fabrication.
3D printing buys you faster iteration and design freedom. It's also convenient for making a regeneratively cooled rocket engine (which makes reusable launch easier, by the way!).
Making a rocket a little bit bigger to enable reuse is not actually that expensive. Costs don't scale linearly with rocket payload, so making it larger has a smaller marginal cost than you might think. Besides, now you ALSO have a larger rocket that you can fly in expendable mode if you need to.
So if anything, 3D printing allows you to do reuse, as it allows you to still afford making a small number of vehicles per year with similar per-unit fabrication costs as if you were making a large number. But because each airframe can be used multiple times, you can do just as many flights per year as if you invested in a large conventional tooling capability but at a much lower cost.
The near-utter-lack of reusable smallsat launchers is baffling to me. Especially now that reuse is effectively proven.
And I completely agree about the financials. There's some pretty strong Kool-Aid being passed around in the smallsat industry. Requiring dozens or hundreds of launches per year to meet revenue requirements? The main argument against reuse has always been lack of demand, but if you're already assuming 100 launches per year (or thereabouts), reuse clearly makes more sense.
And 3D printing is not likely to reduce costs per unit aerospace dry mass. 3D printing is fantastic for increasing the iteration frequency and reducing tooling costs. But if you're entering high-rate production (as all these revenue assumptions imply), then that advantage of 3D printing rapidly disappears, and conventional manufacturing is almost always cheaper per unit.
And for performance, this is even more true. Conventional forgings have much higher specific strength than any 3D printing process with a similar alloy. The alignment in grain flow just can't be beat. And forgings allow mass production as well, so benefit a lot from series production, allowing low unit costs at high material performance. (The soda can is the ultimate manifestation of this: extremely good mechanical performance with very little weight and almost no cost.) 3D printing generally produces parts somewhere between wrought and cast. So if you're making the "mass produced expendable" argument for low launch cost, you should avoid 3D printing!
I have a deep (but autodidact) knowledge of space launch, and I professional do metal 3D printing, including of aerospace components.
And there are a whole bunch of these smallsat launchers. All of them seem to think they will be the winners with ~100 launches per year (in a market where Pegasus gets about 1 launch every two or three years??). They also all seem to think they can compete against reuse by just really trying hard to make expendable costs lower. I simply don't see how this is feasible.
In a world where there are hundreds of smallsat launches per year, there's no way you get to charge $10m per launch to do so.
EDIT:I forgot to mention: 3D printing is not a cheaper manufacturing option than existing conventional manufacturing on a per-dry-unit basis. You're still talking on the order of $1000/kg, which is the same as finished aerospace products. PLUS, there's still an enormous amount of touch labor involved in 3D printing anything. The main advantage of 3D printing is the minimum order is 1, which is good for testing and iteration.
But there is a big BUT - being complex does not preclude them being produced like sausages with big enough manufacturing complex, big and competent labour force, and very good rigid procedures perfected over 60 years.
In comparison to the cost of the whole space program, launch vehicle costs are almost microscopic.
Every rocket is a project, not a product.
Every rocket launch is not a service, it is also effectively a project.
And it was an achievement of the crippled post-Soviet space industry in making launch vehicles and launches more close to being a product and a service. Prior to that, a commercial satellite launch was hard to do without spending an eternity doing talkshops with likes of Boeing, Lockheed or French government. When ex-Soviet rockets popped up on the market, procuring a satellite launch service turned, figuratively, into a grocery shopping.
I think the trend may well go in the other direction, and become more like shipbuilding. Due to reusability, we can afford to make costlier rockets, as they're useful for many launches. This may for instance mean the use of relatively expensive composite materials, such as in the BFR, which prioritises fuel cost above unit cost. Also, the efficiency of rockets generally increases with size.
Carbon fiber composites are more expensive than stamped metal parts like automobile bodies, but if you're talking CNC machined or 3D printed metal parts, then carbon fiber can indeed be cheaper.