Hacker News Reader

Comment by metal_am | Hacker News Reader

While I'm not convinced on the utility of 3D printing a cylinder (I think they have plans for more complicated lifting body type shapes in the future), it's great to see more players in the game!

As an aside, I'd love to know what alloy they're using. I know NASA had problems with fusion welding 2195 and had to switch to friction stir welding.

View on HN

In my opinion printing the whole (±) rocket is a fantastic way to build your first protoypes and iterate your designs with high pace. This will probably not scale very well though, especially for geometrically simple parts like cylinders, as you said. But for novel applications where until recently a part like this couldn't have been manufactured realistically, 3D printing is a very exciting new technology in the field. Take a look at their tank caps, which are designed exactly with printing in mind. They have shapes that before their new processes couldn't have been dreamt of. Very excited to see where this goes! I'm currently doing my aerospace engineering bachelors in Germany, hoping to see similar concepts here or at least in Europe soon!

That's a good point about scaling. They're not trying to build 50 of them right now, just a handful of prototypes. And AM is particularly well suited to that.

I don't think they ever planning on building 50 of them. If they can get reusability sorted out quickly, they're unlikely to ever need more than a dozen or so.

I thought the whole point was that they wouldn't need to reuse them if they can print a new one every week for just the cost of the raw metal.

I think the idea is that if they can get the process down and prove it's viable (today was a huge step in that direction, especially surviving MAX-Q and not blowing up), it will scale because instead of spending TONS of money on tooling that is typically not reusable past a singe iteration, they can iterate and produce quickly and efficiently. Curious, though, why you say it won't scale? Why not?

This is such a glorious plan.

It's a wonderful un-process hack. You talk about tooling, but there's also just so much less process, so much less variability, so many less joints & non-uniformities. It may be slow, but you only really have a limited set of lessons about printing to learn.

By compare, look at videos comparing Starship nosecone over time. So much rocket building is a manual process, and it's hard work figuring out how each thing has to happen. I think even the cylinder fab has a lot of lessons that have gone into it.

At first this kind of felt like a gimmick, but being able to go deep on one ultra-flexible thing & reapply those lessons again and again seems divine. Being constrained by chiefly imagination, asking only where you want to put material not how you want to get it there is such an enormous liberator.

And once it's working well, they have such deep parameterization tweaks they can make, to optimize hone & refine. Thin this in this area, try some different reinforcement patterns on this wall... having a totally abstract way of building feels like cheating, it's such an obviously easier better freer constraint-less way of making real.

The potential savings here seem to more than make up for a potentially slightly slower build speed.

And, if the build speed it that much slower, get another 3d printer! Because it takes so much less manual labour, you can parallelize it in a very cost effective manner and ultimately end up with a faster production rate.

3D printing a rocket would be a lot more convincing if it hadn't taken Relativity ~7 years and billions of dollars to make it to a first launch attempt. They'll probably do alright, but that's much more capital and much longer than others (SpaceX, Rocket Lab, ABL)

Less than the time it is taking and money Blue origin or for that matter SLS took.

Falcon 1 and electron are smaller vehicles in addition to being not 3D printed.

They had the runway and investor confidence to attempt a larger vehicle as their first

I love to make fun of Blue as much as the next guy. But New Shepard is a more complex program even if it doesn't go to Orbit.

I agree, why bother with the body cylinder. Do welded rolled steel with cutouts like SpaceX, standard bolts, etc..

Fwiw, SpaceX also 3d prints quite a bit.

A bit of convergent evolution here.

Why not try something different? Just because spacex does something well doesn't mean everyone has to copy spacex.

I'm not making that argument.

SX picked rolled steel (really, a custom variant) as a good tradeoff for performance, durability and cost. That was trying something different, and Relativity's goals for low-cost and reuse look much more like these goals

Just because most cars use round wheels doesn't mean trying different wheel shapes is necessarily a good idea. Sometimes there's just a clear best way of doing things and there is no benefit in diverging from that.

>SpaceX also 3d prints quite a bit.

Pretty much every modern rocket engine manufacturer does. The engines have a lot of unconventionally shaped pipes that could be manufactured as a single structural element using 3d printing to avoid the need for extra welding and potential points of failure.

the only use case I can think of for the rocket body 3d printing, combined with methane engines, is a something weird: fly to mars with a 3d printer and some technology to extract iron from the sand and methane from the atmosphere and build multiple rockets there to fly something useful back to Earth. But that's such a big leap of imagination.

Reply on news.ycombinator.com