A 3D Metal Printer Is Churning Out Rockets | Hacker News Reader

A 3D Metal Printer Is Churning Out Rockets

spectrum.ieee.org

154 pointssohkamyung81 comments

Too bad they don't mention the competition that is also using 3D printing, and is in fact quite far ahead of them. Rocket Lab has been flying 3D-printed engines for years.

I understand why a regular PR piece wouldn't mention the competition, but this is IEEE Spectrum which is supposed to be better than that.

> is in fact quite far ahead of them

This is like saying that AWS doesn't matter because people have been hosting servers for decades. It's technically correct but misses the entire point.

SpaceX and RocketLab do print parts, but they still need assembly lines with hard to source machines and skilled labor. RocketLab is working on reusability not because of cost savings but because they are assembly line constrained.

https://www.rocketlabusa.com/news/updates/rocket-lab-announc...

> we’re now establishing the reusability program to further increase launch frequency,” says Mr. Beck.

That's a jaw-dropping statement, it's easier and faster to develop reusability as a way to get more cores than to build a new assembly line.

Relativity is printing out entire rockets, only stacking required. New assembly line is entirely contained in a new printer, constrained on upper stages? Just retask a printer no multi-year long lead time to build up a new set of traditional tools. Don't need any more rockets right now? Print planes instead. The list of advantages goes on.

I'll suggest this interview with Relativity space. https://www.wemartians.com/episode062/

You may also be interested in this one about another company looking to print solar panel arrays on orbit. https://mainenginecutoff.com/podcast/131

Rocket lab though uses less metal and has carbon fibre chassis, fuel tanks, etc. They 3d print their engines which is the main metal part of the ship.

The constraint may be in autoclaves as carbon fibre is a more intensive product to make than just 3d printing or bending metal.

Those are good points. It's unfortunate that an article supposedly for a technical audience didn't make any of them. The "related work" section is a longstanding tradition.

I’m guessing that’s what you mean by “PR piece” ?

About the Author

Bryce Salmi is the lead avionics hardware engineer for the Los Angeles aerospace startup Relativity Space.

It is possible to work for a company and still describe clearly and fairly the differences between that company's approach and their competitors --- i.e., do better than a purely self-aggrandizing press release. I assume Salmi probably could, and it's just that the IEEE Spectrum editors didn't ask him to, which is disappointing.

He definitely could have - he worked at SpaceX as well. But saying things about what other companies are doing is a big no-no in the launch industry, and I doubt they would give permission. It's the kind of thing that can get you put in jail due to ITAR.

In contrast, here [0] is ULA doing it the more traditional way for their prototype. Don't think it can be compared 1to1, since small sat launcher play in a different league. But it certainly highlights the massive rooms involved.

[0] https://www.youtube.com/watch?v=Eb8QkORA3HA

Looks like an aluminum welder on a robotic arm.

Not discounting it, because I think the simplicity of doing it that way is brilliant for plenty of parts. Plus, I bet there was plenty of trial and error to get the timing and gas settings right for something like this. Got to give them credit there. Aluminum welding is a bitch (I suck at it). So, I'm not trying to rain on their parade... but a quick google search is showing they're not alone. To be fair, I think they might be doing it better though... especially when it comes to scale. A few minutes in, I don't see anything near to their scale, which is cool.

> Relativity has in-house metallurgists who are honing the process, ensuring that our components meet the strict quality standards for aerospace hardware.

My understanding is that much of the cost of flight materials is ensuring/maintaining the quality pedigree. I would have liked to see more detail into the ways quality changes with a new manufacturing process. E.g., are certain non-destructive evaluation techniques necessary? Are inspections capable of being automated?

Great Real Engineering video with a lot of discussion of the issues with material properties, crystalline structure, and fatigue cracks related to metalized 3d printing:

https://www.youtube.com/watch?v=fzBRYsiyxjI

Currently, one would expect to filament wind all the tanks, intertank and nose out of carbon fiber.

You can buy off the shelf machines for that.

It's used widely for solid rockets, where the internal pressure is very high.

I don't know why not for pump fed liquid rockets, maybe the design being buckling limited makes the approach worse.

It's oft said that Relativity Space is a 3D-printing company; that's the hard problem they're solving. Launching satellites into space is a side-show.

If the problem they're solving is harder than "just" launching satellites into space, then that doesn't bode well for their chances of success.

Just launching stuff into orbit, or attempting that, has swallowed many companies in failure.

Not necessarily. If they can solve the issues with 3d printing then they can potentially have a cheaper, faster, and more versatile production methodology which can result in faster iterations to solve the issues that they face with the actual launch process.

When this technology can exist in space, earth will have orbiting shipyards building space craft.

Is this assuming some of the materials are available there? This would make sense if you're getting fuel or metal from asteroids, moon's surface, etc and building ships near those. But in Earth orbit, you'll need to lift the shipyard, then the fuel, then the metal powder etc for sintering, so what would be the gain?

Yes, you'd probably want to go get some asteroids. Even small ones can contain huge amounts of metals, and icy ones could be turned into fuel. There's a lot of potential, but it's all a bit out of reach for the moment. https://en.wikipedia.org/wiki/Asteroid_mining

Isn't moon dust mostly bauxite? (oxidized aluminum)

Not exactly https://en.wikipedia.org/wiki/Geology_of_the_Moon. I think the problem with moon resources is the moon's gravity while forming was strong enough to pull metals like iron and nickel down into the centre where we can't get at them (they're on the surface a bit too, but not conveniently concentrated). Asteroids, on the other hand, can contain lots of accessible nickel, iron, cobalt, etc.

I'm no expert though!

Wikipedia looks like it says [0], by elemental concentration, roughly 42% oxygen, 21% silicon, 12% iron, 8% calcium, 7% aluminium, 6% magnesium, 4% other.

[0] Bar graph with insufficient labels: https://en.m.wikipedia.org/wiki/File:Composition_of_lunar_so...

Magnesium should be even better for some space apps. In fact, it would sinter better in a vacuum than air.

https://www.deepdyve.com/lp/wiley/sintering-of-magnesium-h06...

Still, the concept of space mining is slowly becoming a thing: https://space.mines.edu/

Lifting many small things is easier than lifting one Battlestar Galactica sized thing.

Also, if it goes wrong the damage is limited, and likely fixable.

You could construct things that wouldn't fit inside the nosecone of a launch rocket if you assemble them in orbit.

You could also repair/refit things without them having to go through re-entry and then be launched a second time E.g. replace damaged/worn out rocket cones or solar panels etc if you can build them in-situ. Imagine Earth-Mars shuttles that just go back and forth that would need maintenance

Orbiting shipyards seems like a likely bet (assuming we want to explore and interact with space).

If we want giant, heavy spacecraft that are capable of interacting with their environment — the space equivalent of a construction vehicle or a tank — earth's high escape velocity makes it impossible to launch from.

We'd instead need to construct it in space, even if all the materials were from earth.

As the cost of launching from earth continues to fall, it'll also become increasingly economically feasible to launch multiple payloads to build a larger, more complex spacecraft.

(Think of the assembly of the ISS as a proof-of-concept here [1]).

[1] https://en.wikipedia.org/wiki/Assembly_of_the_International_...

The entire point of construction in space is to avoid launching from earth. It takes 36 times less energy to get a ton of material from moon surface, and can be done without rockets at all - a space elevator can be built on the moon with 'normal' materials.

Why is Elon Musk aiming for Mars if building a solar panel array and all of the subsequent manufacturing industry on the moon could potentially be way more productive?

Is there something I am missing?

Musk has been talking more and more about the moon lately. Maybe he's thinking the same.

The fact that NASA has a clear directive to spend an absolutely humongous amount of money on the moon over a relatively short time frame may have something to do with Elon Musk's interest.

US military uses quite a ridiculously larger amount yearly. Nasa budget for going to the moon is insignificant compared to what is spent on war.

False dichotomy. Military budget that can be spent on Elon’s business is probably less than what NASA is prepared to spend on a proven Earth-Moon-Earth service.

It's not a perfect false dichotomy, between people getting to keep their wealth and all of the things it could be spent on, there actually is a sum that has to balance to zero.

Robert Zubrin argued that Mars has more lightweight bulk resources (water, carbon dioxide) twenty years ago, and at the time the moon was thought to be barren of such resources that are useful for starting a human colony.

Recently (in the last decade) water has been discovered on the moon, frozen in areas of cold permanent shadow in craters.

Also the current US vice president wants to go to the Moon. The previous president wanted to go to Mars.

Mars is a more hospitable, Earth-like environment for humans. It has an atmosphere, a normal-length day-night cycle, less extremes of temperature, and closer to normal gravity.

Mars and the moon both have pros and cons. Mars would be an easier place to survive, but the moon is easier to get to.

"Mars would be an easier place to survive, but the moon is easier to get to. "

Since it takes 6 months to get to mars and only 3 days to the moon and the only advantage for surviving on mars is the mini gravity, I really think we should start with the moon.

It's not just the gravity. On the moon, you have to deal with 2 weeks of day and 2 weeks of night, which means having to deal with wide temperature swings and if you rely on solar you need very large batteries.

Mars also has an atmosphere which contains CO2, which is useful for making methane. It also has a lot of water. (The moon has water too, though it may be less abundant/accessible.)

Could some of those downsides be mitigated by putting your base on a "pole"?

That is, are there locations on the Moon's surface where the sun is perpetually visible?

There are just a few crater rims where this happens, not the full pole - but those do indeed get sunn almost all the time & border areas in perpetual darkness, which is also pretty useful (cryogenics, temperature gradient, etc.).

Sort of; there are some spots that get sunlight almost all the time. (The moon has quite a bit of wobble, so it's hard to get to 100%.) That greatly restricts your choice of base locations, though. There may be other resources or geological features you might want to be close to that aren't convenient (lava tubes, water deposits, flat places to land, minerals that can be used as building materials, etc...).

I suppose if our base is extensive enough, we could use those locations primarily for power generation and then transmit that power to more desirable locations for other tasks. Though I do realize that I'm talking about a lot of effort/time/money/resources.

Day and night cycle is an issue yes, and sure, for a long term self sustainable settlement, mars looks much better.

But in terms of setting up a base soon and building rockets for example ... the very short travel distance to the moon offsets any advantage the mars has

The habitability difference is not relevant at this stage - both will kill you instantly. The ISS orbit has no atmosphere or gravity, but it's easier to keep people alive there than on Mars.

For a Moon base locations, look up peaks of eternal sunlight. We also have space-grade nuclear reactors and RTGs ready to go.

On the Moon you can have realtime communication, bring way more equipment, have robots remotely controlled from Earth, and get help or evacuate quickly. NASA and China will be able to land people on the Moon, other nations can get rovers/supplies there.

Anything goes wrong on Mars - you are dead. I would pick Moon any day.

Mars has more resources available to work with and is a less extreme environment than the moon or space. There's definitely a trade-off between quick and easy access to the moon versus an easier environment to work with.

No one can survive in any of these places without a space suit, but on Mars you have easier access to the various natural resources you'd need in order to survive and build additional infrastructure. Light for solar power is consistently available on a normal day/night cycle, as is CO2 and water, which can be used to make methane fuel and liquid oxygen. It has concentrations of sulfur, which can be used to make a form of concrete. The gravity is more conducive to long term human health. All this is relevant.

Surviving on the moon or in LEO is also quite possible, it's just that you'd need to bring more equipment and all the resources you can't find on-site.

Mars is absolutely a better place to live, it's an actual planet.

But for industrial hub, the fact that we have the moon is an absolute miracle. We can build and refuel ships that dwarf our current ones. And we will need moon based factories if we are to make any meaningful progress is settling the solar system. The only things that should ever be lifted off earth is People/ other life, and super high-tech goods like microchips.

The Moon is also close enough to Earth that you can get help, or go back in case you suddenly need surgery. Good luck with that on Mars.

Settling Mars with our current ships is like settling Australia with a canoe.

Indeed, yet there are various challenges inherent to Moon colonization, such as continuous direct surface impacts due to no atmosphere, extreme temperature surface environment due to 14 day days, lack of various elements (such as carbon or nitrogen), etc.

Still super useful & the short distance overrides much of the downsides.

> The entire point of construction in space is to avoid launching from earth

This is true, but avoiding launching from earth could be done for one of two reasons:

(1) to increase the efficiency of what can be built

(2) to increase the possibility of what can be built

Your point is (1), and I agree with it. The point mentioned above is (2) — even if all the materials for the ISS were from earth, it still could not be constructed on earth and then launched as a whole, and so construction in space is viable.

It would still be a good stepping stone to complete off-world industry; one of the reasons current space launches cost a lot is because you don't only want things in orbit but you want something in orbit intact and functional. A pallet of dumb steel ingots could be lopped up with something with much more generous tolerances ala the Big Dumb Booster[0], with any orbital maneuvering once up there accomplished by space tugs.

[0]https://en.wikipedia.org/wiki/Big_dumb_booster

How will the heat dissipate if there's nowhere for it to go?

Heat can still leave parts in space by radiating the heat away. We don't usually think about it because it's generally much slower than convective cooling here on Earth but it does happen and doesn't require a medium to cool. It's the method the ISS uses to cool its interior for example. [1]

[0] https://en.wikipedia.org/wiki/Radiative_cooling

[1] https://science.nasa.gov/science-news/science-at-nasa/2001/a...

right, but the rate of dissipation would be a huge problem for the additive technique described in this article. Plus the lack of gravity means that the molten metal would adhere to the existing component very very differently. Essentially it would mean designing the process from scratch - with almost no ability to test on earth during the design process. Not sure how effective computer modelling is for that type of thing :)

You can do small scale tests of each bit though: testing the sintered powder jet on a parabolic flight, test the heat dissipation in a vacuum chamber, etc. It wouldn't be easy but it's far from impossible.

Unfortunately it's too slow. I work with electron beam welders so this is an everyday problem as it takes place in a high vacuum (1e-3 Torr or lower). Though for welding it usually isn't a big issue as most parts are large enough to sink their own heat. For deep high power welds in material which can't sink the heat we attach copper heat sink blocks to soak up heat like a sponge or for extreme cases, circulated oil cooled heat sinks.

In space you can't vent to atmosphere or run the oil through a fan/water cooled radiator. So I'm thinking you'd have to build a heat reclaimer or high surface area radiator which uses an active cooling system like thermal pipes or circulated fluids like oil. But how does this adapt to different parts with odd shapes? Perhaps the printing would have to take place in a pressurized environment to facilitate cooling using gas and then figure out how to get rid of or recycle the waste heat.

Depending on the conductivity of the material being used could you not have it sink it's own heat through conduction into a pad connected to a beefier version of the ISS ammonia to radiator cycle or combine it with a phase change cooler that melts wax into liquid then slowly radiates that heat away? Even in an atmosphere you have to deal with all that heat eventually.

Truthfully I don't think the best way would even be to try to 3D print the whole structure though just generic pieces that are then assembled into larger assemblies. Just being able to mass manufacture extruded profiles and sheet metal would allow you to do a lot in space that would be hard to launch piecemeal from the ground. Then smaller specialized pieces to join them together in the specific way required for the current vehicle being built could be printed and assembled.

There's a lot of neat research going on in assembling structures like that atm.

>Unfortunately it's too slow.

well, it depends on the target temperature you want to maintain. Radiative power grows with the forth power of temperature (https://en.wikipedia.org/wiki/Stefan%E2%80%93Boltzmann_law). So that red glow of Merlin engines mentioned by the other comenter radiates so intensively that any other method of heat transfer can hardly come close. Another example - just 5700K of the Sun surface is enough to radiate away all the power the Sun's body generates.

The upper stage variant of the Merlin engine used on Falcon 9 uses radiative cooling as well - that's why it's nozzle extension glows red hot - it cools itself that way! :-)

Thought experiment: how does the heat of the sun reach the earth?

If you turn on a blowtorch in space, does it need some sort of kylo-ren double-action torch to keep it in place? I guess it won't move if it's anchored to something, but that something will still move as the force is unbalanced?

Probably safer to tether yourself to the thing you're blowtorching...

We are closer now that the ISS has a functional 3d printer (looks like PLA)

https://www.nasa.gov/content/international-space-station-s-3...

You could manufacture goods ( electronics, cars, etc. )in space, and then drop them onto Earth. Free shipping!

The reason relativity calls their rocket Terran 1 is they want to make “Ares 1/Luna 1“ someday :)

Rockets are really neat, but to me this says a lot more about how far robot manufacturing has come and where we're going.

The robotic technology involved here is decades old. What they're doing that's new is designing for manufacture using modern materials.

That robot arm is probably an off the shelf part.

It's off the shelf, it's a Kuka arm but the arm itself isn't the interesting bit. It's integrating the 3 arms, the print head, and the rotary table. I've programmed an older one of those arms and it was no cake walk and my program was to just do the Towers of Hanoi as a demo. [0]

[0] It's a pretty simple demo because it can just be done as a simple set of calculated linear motions. The only real trick was that the version of the language it was using didn't support recursion.

I'm eagerly anticipating the day when hobbyists can 3-D print engine blocks and cylinder heads. Imagine the engine swaps enthusiasts will come up with when they can make their own designs from scratch and share high-performance designs as open source hardware. It will also be a boon to people restoring antique cars, the ability to print a new period-correct engine block.

Wouldn't milling be a much better way of creating bespoke engine blocks than printing?

Printing engine blocks will never work without sintering. Metallurgy has limits.

Direct metal laser sintering (DMLS) 3D printers already exist. Is there any reason these aren't capable of building engine components?

Yes, tensile strength and metal composition. A modern engine block is far more complex then a simple milled piece of steel.

Sintering != milling.

It is not uncommon for engine components to be milled. I guess what I'm asking is if there are properties inherit in sintering (e.g., crystalline structure) that make it unsuitable for engine components. You mentioned tensile strength and composition, are you alluding to the fact that specific alloys can't be used in laser sintering or that the process itself creates inferior composition?

The latter, although many common sintering alloys simply couldn't meet the requirements to be an engine block. The crystalline structure caused by the unique thermal process of sintering creates a far less stable structure, especially against thermal stress, which is less than ideal for a combustion engine.

In my original comment, I meant to say forging. You could sinter and then forge, which could work.

Any background sources on that? I'm curious because SpaceX is using 3D printing for some of their engine components

It's getting there, but the dimensional tolerance and finish for something like an engine block isn't something 3D printing can do yet-- or ever.

It might be OK for many parts (like the manifolds) but some parts will _always_ need to be finished by traditional machining operations.

https://www.youtube.com/watch?v=nyYcomX7Lus

I'm not sure if that's what this one is doing specifically, but there are metal additive-manufacturing machines out there that are 'hybrids' where they can print and mill as they go to effectively make 3d-printed parts with very good tolerance / surface finish. So they _do_ get finished with traditional machining operations, it's just done as its being printed out.

Relativity Space is one of the aerospace companies I'm most excited about (especially in regards to whatever they pull off in the 3D printed launch vehicle space)!

I am unfamiliar with material sciences. Are 3d printed metals as structurally sound as other techniques for metallurgy?

DMLS is generally within 80% iirc. This appears to be a bit different in terms of technology, but I don't see why it should be significantly off. Both Rocket Lab and SpaceX have been flying 3D printed rocket parts for a while now.

3D printing is only one aspect of the solution. And that isn't the real problem yet (ha ha..)

How's the 3d printing landscape for metal overall? I hear there have been made great strides in 3d printers that rely on sintering. Very affordable relatively speaking.

Any other methods that have gained traction?

Reply on news.ycombinator.com