3D printed titanium structure shows supernatural strength | Hacker News Reader

3D printed titanium structure shows supernatural strength

rmit.edu.au

79 pointsgmays55 comments

This is step 0.

I suspect the real power will come in with composites. Even with metals, metal matrix can have very nice properties (https://en.wikipedia.org/wiki/Metal_matrix_composite) compared to a base metal.

The key thing of interest here, to me, are the holes. The holes can be filled with something other than air.

Wouldn't that just raise the density of the structure and defeat the purpose?

Depends on the material. Matrixes are certainly a proven value.

horse hair in plaster, carbon fiber or glass fiber in resin, rebar in concrete, are all better than either of their individual components.

A supporting filler could increase the toughness, supporting the filaments against buckling, or there could be other values like filling the space to prevent moisture or oxygen from getting in there, etc.

I mean, maybe you can fill it with something that changes the thermal conductivity of the matrix, maybe something that has minimal weight but improves the compressive resistance, etc.

Hydrogen or helium would decrease the density, but I digress.

In most cases, composites have complementary materials. You would need less titanium if it was reinforced with something else. Examples:

- Most common are materials like carbon fiber, ceramic, or aramid.

- Another failure mode is denting (or localized impact damage). If I make a tube, as here, the thinner the wall and wider the radius, the stronger it is as a unit. However, at some point, it's easy to crush it like an aluminium can. Cover the inside with a polyurethane foam, and that failure mode goes away, or at least dramatically decreases, for very little added weight.

- I might care about things like thermals.

Etc.

Are these properties unique to printing in titanium? Doesn't sound like it. So hopefully we will see this as an infill option soon.

Right. This should be helpful for all 3D printing. Finite element analysis on infill patterns looks like it helps.

This pattern looks like it's designed for an isotropic material. Sintered metal printers ought to produce solid metal. Extrusion printers, though, tend to have weaker inter-layer bonds than bonds within the layer. So this may not work as well in PLA. Might need a pattern that compensates for the weaker inter-layer bonds.

Also helpful for Stereolithographic printers (SLA/DLP/LCD), since they also don't have inter-layer adhesion issues like extrusion printers.

Recently saw an interesting infill/layer pattern that offset the layers and spaced them (think hexagonal pattern when viewed from the side) - might help with the single plane delaminating issue. Combined with this style of infill pattern we could be getting somewhere.

Makes me wish I had time to mess around with GCODE

This sounds a lot like "brick layers" by CNC Kitchen: https://www.youtube.com/watch?v=5hGm6cubFVs - I believe that something like that will be a must for future 3d printing slicers, as it increases mechanical strength with no extra hardware needed.

No, I know I've read an article about 3d printed steels being stronger if arranged just so.

Edit: it was probably about https://www.science.org/content/article/3d-printing-doubles-...

It looks like there might be enough difference in relevant scale that both techniques could be applied.

Not according to the article, the mechanics are structural. That said, I downloaded the 'cube' picture that was the banner and the cube does not use the structure described in the article. That makes me a bit curious.

The article goes out of the way to talk about the tubular structure that is bisected by a solid structure. And yet the cube has only tubes, and where the tubes are unfinished at the edges, the tubes appear to have an internal bisection.

One of the fun things is that you absolutely could not injection mold this :-).

Generating it as an infill would be interesting. I'm generating some cubes with CAD (since that is easy to do) that printed at 100% infill will (aka no guessing on the slicer's part) will let me try their cube structure, the paper's suggested structure, and others. I'm going to guess they patented it (could be wrong but most university research seems to go for patents) so it might not be possible? to put it into an infill library without dealing with that.

It would make sense that a complex structure that mimics the molecular structure would be stronger per unit of mass than a lump of the same material, in the same way that a simple i-beam or a truss is stronger that a solid rod of steel in specific design applications. In both cases, its strength where you need it; nothing where you don’t

Really neat stuff.

Do you think this team invented that idea? Im struggling to understand how large their contribution is.

We haven’t been able to fabricate relatively perfect metal crystals in most metals until recently.

I dont this the object they manufactured is monocrystaline. Its a part made with a commercial laser sintering 3d printer

The point they’re making is that since making a monocrystaline material is difficult, it can be emulated by using a metamaterial structure that confers similar properties.

Ah! I had started working on tetra-lattice supports several years ago, but got distracted and never went back to it. I figured it would give a relatively optimal strength to material ratio and could scale pretty well up or down.

This wasn’t an original idea. Between ‘00 and ‘03 I worked at MSOE’s Rapid Prototyping Center and one of the research teams there pioneered a method of using 3D printed tetra-lattice to make cooling channels in injection tooling. [0]

0 - http://utw10945.utweb.utexas.edu/Manuscripts/1999/1999-066-G...

At first with the kerning on my screen I was wondering why MIT had an Australian TLD.

They seem most interested in the thin band running through the lattice, my curiosity is how well can the 3D printer make that consistently? I would assume it's contribution to strength might be pretty significant, and might be quite non-linear with any variance.

A common frustration I've had with 3d printing is even keeping within specs on thin planes and walls can turn out suboptimally, but maybe thats some bike shedding, I'm ceratinly not doing anything in metal.

Those industrial titanium powder printers are not comparable at all what you know, I assume. They‘re used at scale to produce airplane parts. Usually also utilising bionic structures, e.g. to make brackets that are much lighter and sturdier, than their conventional counterparts.

Yeah most of what I do is in SLS, I don't really know what the nuances to how the sintered metal powder ones behave vs polyamide etc.

One of the bits SLS can struggle with is long plane or ribbon-like objects. And then as you get thin enough, the surface finish starts being a non-trivial % of the overall thickness in that area. Just was curious how sensitive the strength is to uncertainty, and as a second part of the overall structure aids it by making ribs to mitigate any weirdness in the flat bit.

This video by breaking taps is a good one for anyone interested in the process or the challenges of producing these sorts of metal lattices

https://www.youtube.com/watch?v=tpO5DgiCgyM

The author seems to confuse what an alloy is with the shape of a metal part. The alloy is the chemical composition, the shape is just the shape. The basis of comparison is likewise flawed. It seems like they compare their shape printed in titanium to a solid block of magnesium, but they should really compare it to the best known titanium lattice as that would show how much of an improvement their design is. For all we know their design is worse than the existing state of the art

The article has whole section called "2.3 Comparison with Existing Metal Lattices"

Yes, not the article but I found that section in the paper and it does appear to be better than existing techniques. I wish the author of the paper could have told us that lol. Thankfully the paper is not paywalled

This is awesome, but did they really need to use "supernatural"? There's an explanation for the strength using the laws of physics (or, dare I say, Nature) ergo it is not "supernatural"

If the outcome is stronger than what would be found in nature, I would find this to be "supernatural". Of course the term could get watered down as anything manipulated and altered by human or machine could be deemed supernatural I guess.

But elemental titanium doesn't occur in nature in the first place, so the straining of terminology seems ill-founded unless we want to call my titanium spoon supernatural as well.

I do.

https://www.youtube.com/watch?v=XO0pcWxcROI

It does occur in nature. Every chunk of it that exists, is currently existing in nature. The fact that usually, on earth, it's always in some compound and we have to extract it by some process is irrelevant. That does not make distilled titanium "supernatural", merely organized. We took some mixed stuff and seperated it into unmixed stuff. A pile of graded pebbles all the same size is not spernatural.

>That does not make distilled titanium "supernatural", merely organized

Supernaturally organized and supernaturally spoon-shaped if we're to accept the parent's terminology. Which, you'll note, I have not suggested we should do.

Right, my mistake, the definition of supernatural I would like to take is actually referring to the strength of the material, and not the material itself. In my original comment I made a reference to both of those things, but side by side the definitions are confusing.

In your real life then, would you tell someone to their face that an aluminum can is "supernaturally" strong?

I do not see myself doing that in a conversation, on a scale of 0-100 my involvement in material science is 0, however I could see a scientist somewhere having that conversation.

supernatural: attributed to some force beyond scientific understanding or the laws of nature.

https://www.merriam-webster.com/dictionary/supernatural

1: of or relating to an order of existence beyond the visible observable universe especially : of or relating to God or a god, demigod, spirit, or devil

2: departing from what is usual or normal especially so as to appear to transcend the laws of nature

If we put this all together, you think scientists are saying to each other that aluminum cans are beyond scientific understanding and they transcend the observable universe. Do I have all this right?

Do you think maybe you misunderstood what supernatural means?

We already have 'artificial' for this.

Artificial strength does almost nothing to progress my understanding of what the strength level is, supernatural strength gives me a sense that whatever was made is very, very, very strong.

“Incredible” or “exceptional” does that without implying that the cause is magical or otherwise outside of the domain of science.

It IS found in nature. They fabricated them and they actually exist. Existing in reality in the universe by whatever means, is existing in nature.

Supernatural is imaginary.

Aside from that, even taking the layman idea of "natural" meaning merely not artificial, the structures are literally based on structures found in "nature" in the first place, as in both organic fractal structures and inorganic things like volcanic foams.

I think supernatural is describing strength, not the pattern itself. A water lily that has a similar lattice structure does not have supernatural strength.

Supernatural is imaginary in the way many concepts of language are imaginary, they simply convey ideas. You could explain to me in detailed technical terms what the discovery was, but you would probably lose me 4 words in because I am not in that field.

I see my error in my comment though...my idea of supernatural as something that is much stronger than what is typically found in nature vs my idea of supernatural as something that is found in nature but has been modified. Those two understandings are at odds with each other and confusing.

The hollow struts are interesting, I've worked with lattices but that's not something I saw before. I guess they are stronger than solid struts with a smaller diameter so they have the same amount of material.

I’m very curious if they can actually empty the un-fused material out of those tiny struts. They said they printed it with a powder bed printer, so voids would be filled with powder unless they can be emptied. I’m curious how that affects weight. I wonder if they used the theoretical empty weight or the actual manufactured weight in their calculations, and how different that makes things.

Huh I didn't see the powder part. If the voids are filled with powder anyway, is there a point in not making them solid? I don't see how you would go about emptying that?

Rockets of the future

cool now perform a repair on something inside of it.

Can you repair your hairdryer? It’s literally a Chinese motor with a Chinese resistor, the two most frequent objects in the world.

Objects in the industrial world can’t be repaired. But Titanium can be melted.

> Objects in the industrial world can’t be repaired.

That makes no sense to me, as industrial things commonly are repaired. Sometimes many times across decades as one part or other wears.

From a certain point of view repairing is fixing things that are literally broken, which involves glueing, welding, screwing, ...

Replacing parts is not repairing from that point of view, even if at a macroscopic level it returns something back to working condition.

This is really not making sense.

Also, why did "replacing parts" suddenly spring up?

"Repairing" something is generally considered to be taking whatever action is needed to get it back to the required condition. If some parts need fully replacing, then so be it.

But it could just also be things needing to be re-seated, maybe something needs to be bent back into shape, etc. The list of possibilities is probably endless.

> Can you repair your hairdryer? It’s literally a Chinese motor with a Chinese resistor, the two most frequent objects in the world.

Most hair dryers don’t come with parts manuals so we wouldn’t know exactly which replacement part to get. And there’s more to it like a switch, a cord, etc. These things could be designed for repair but they would have to provide reliable part numbers for the internals, which they currently do not do.

> Objects in the industrial world can’t be repaired.

Objects in the industrial world most certainly can be repaired and it happens all the time. For example here is how a giant excavator tire is repaired:

https://youtu.be/p0R-YWQCfdQ?si=iGagfyttgO5vTvDr

How do you perform repair on something inside a solid steel, titanium or aluminum block, which is what this structure replaces?

These are structural components that don't have anything "inside" but are used as material for carrying/distributing loads, and their strength and weight is the thing that matters. If their integrity is compromised, you generally don't repair structural components if the repaired seam wouldn't have the same strength as the original material, but replace them.

Do you think there will be a market for typical household electronics items such as the hairdryer specifically made with repairability in the near future? Perhaps driven (unlikely) by consumer demand or more likely by resource prices or legislative pressure.

I think a real-life example that might illustrate the difference is a car that's been in an accident. A good body shop can do a lot as far as straightening components, doing new spot-welding, etc. in a car that's made with contemporary techniques, but imagine if it had a chassis or unibody components made out of this material. They would need to be replaced, because straightening won't repair the micro-structure, and they probably can't be welded. Replacement is going to be far costlier as well, because you'd have to disassemble the whole car and put it back together. plus the replacement part has to be meticulously 3D printed instead of stamped/forged/welded.

That said, I can imagine it being good for applications like fighter jet components where it's probably considered scrap after the first accident.

I am not an engineer but i don't think car repair techniques would work on traditional materials used in this niche to the strength required for anything that needs to use titanium alloy. We don't currently make cards out of titanium, we probably aren't going to start making them out of fancy titanium. Car repair shops work because cars don't actually need to be that strong.

That's like saying, cool, now make a repair of a small Crack in the middle of the center later of a sheet metal panel.

There is no need to do so. You would replace the part.

Reply on news.ycombinator.com