Big 3D printing machine prints metal objects | Hacker News Reader

Big 3D printing machine prints metal objects

theage.com.au

58 pointsreirob45 comments

Originally, I read the first few paragraphs, noticed it was mostly fluff journalism and immediately jumped to the HN comments and posted a question asking about the use of lasers to melt the metal either pre- or post-deposition to fuse it to already deposited metal. I then went back to the original article and read to the end of it just to make sure it didn't mention that this was the technique used, since I have made this mistake before. Lo and behold it was exactly the technique used. I quickly went back and deleted the question I posted before I was downvoted with a comment like "didn't you read the article. it said that the process used lasers as you described".

The reason this is interesting is because this is exactly the type of process that ends up being patented. Chances are that the scientists that perfected this stereolithography technique are going to apply for a patent and probably get it, even though it was an obvious approach to use to me, someone who has zero experience with stereolithography. If it is obvious to me, then it is certainly obvious to someone with familiar with the field of stereolithography. Too often people say "Ohh, everything is obvious after the fact", but like this approach using lasers and stereolithography it is very obvious before the fact too. Most things that are obvious in hindsight are just as obvious before sight.

Anyways, sorry to be OT with this comment. It's not meant to be flamebaity, it's just relevant because my original naive question was made on this post.

There's a world of difference between saying "I'd use a laser to melt the metal" and actually perfecting the process to the extent that (a) the metal being deposited is adequately fused, (b) the metal underneath is not disturbed every time a new layer is added, (c) the over-shoot of the laser beams doesn't randomly melt holes elsewhere in the piece, (d) it works at all angles, and (e) concerns I can't even think of yet.

There are lots of scholarly articles about the laser process dating back to at least the 1990s. No patent would be granted today on the general idea. But if this research team invented a unique way to prevent any of the problems listed above (that nobody had thought of yet and solved a real problem --- maybe they flick the beams on and off in a particular pattern, or use different wavelengths in a novel way, or have a unique mounting system that swings the beams off the work more quickly, etc.) then it would be perfectly appropriate for them to obtain patent protection on that improvement.

As somebody said on a different HN patent thread, it's the difference between trying to patent a Star Trek "warp drive" today versus patenting the individual inventions that add up to an actual working warp drive. You'll be laughed out of the PTO if you file on "Step 1: Build a warp drive. Step 2: Profit." But if you can invent the technology to actually make it happen, you can get a patent even though the idea of "warp drive" dates back to 1960s (or earlier) SciFi.

It's also important to point out that this is a manufacturing process patent. Those don't work like software patents.

Fundamentally this is an improvement in the ability to make metal stuff. It might be faster than CNC machining for low volume production, and as less material is wasted it's probably cheaper. Likewise it's almost certainly faster than making casting tooling for low volume.

But fundamentally, it's just making stuff. We already have plenty of ways to make stuff from metal. So anyone who wants to make metal stuff isn't meaningfully impacted: at worst they might be priced out of the market for the new toy, but most likely they just have a new option to investigate. The patent inefficiency impacts only people who want to improve on this obscure technique. What won't happen is that the public will lose access to metal junk.

Software isn't like that at all. Everything is built on everything else. If you get a patent on something like "pinch zoom" and refuse to license it, no other devices can use it. The public loses, because all of a sudden important features can't be combined with other important features.

>maybe they flick the beams on and off in a particular pattern, or use different wavelengths in a novel way, or have a unique mounting system that swings the beams off the work more quickly, etc.) then it would be perfectly appropriate for them to obtain patent protection on that improvement.

If it turns out that they use one of those methods, don't you think that would be obvious?

I think if the patent was specific down to the actual wavelength, pulse shape, and pulse size, well... maybe then that should be patentable, unless it turns out that those parameters are chosen according to a natural law (parameterized based on material density, melting point, etc...)

I think certain things, especially in regards to manufacturing, are more than okay at being patented. But I do think it should be implementation specific, and if an implementation is dependent on a natural law with no alternatives, I'm not sure something like that should be patented, it would really depend on some things.

Yes, the "If it was obvious, someone else would have done it!" argument is a frustrating one.

I have a hard time wrapping my head around it.

There are an awful lot of supporting developments that go into any sufficiently advanced technology.

Patents as they relate to some technologies feel like an odd kind of race, one in which the winner isn't defined by finishing first exactly.

Rather it's the act of crossing a non-specific finish line within a very specific window of time then proceeding to tell everyone else that their times will be invalidated if they happened to take the same route.

It seems plausible to me that a new distribution model could be borne from this tech. With a few large-scale 3D printing sites deployed geographically, delivery times for products drop to next-day or so. Alternatively, a community-based or government-sourced set of facilities allows for arbitrary production of items (replacement parts for washing machines, light-rail components, etc.). Manufacturing might not be as tied to volume-of-scale issues, so that bespoke articles are still inexpensive. It's conceivable as well that parts for obsolesced machines could still be produced (my previous oven need just a heating unit, but the whole thing had to be replaced since none were available).

With knowing admittedly little about the field, I'm hopeful that the issue of producing multi-material or bonded objects is tractable. Current 3D printers are akin to black-and-white dot-matrix machines. Nowadays, we have photo-realistic full-color desktop printers (which are given away for free at this point).

I'm very excited about this technology, as it seems to have very practical (and marketable) uses that we're only beginning to touch on.

I think for at least a while the fab technology is going to require quite a few integration/finishing steps, which will still require specialised machinery which prevents them from being easily deployed in neighbourhood Fab4U scenarios.

I have a vague recollection that the biggest industrial use of 3D printing is in producing the moulds and dies which are then used to mass-produce.

In the longer term, we'll need either more accurate initial positive machining processes, or more integrated systems that include negative processes like CNC milling/EDM/waterjet for cleanup and hard-to-print forms like screw threads.

Yes, that's always been my nagging concern about 3D printing. Sure, it can print any single thing, but nearly every machine is made up of multiple parts. So either there's a set of construction machines down the line from the 3D printer, or 'some assembly required'. The related consideration is that each part of a constructed item has to be printed separately (and probably serially).

Another issue that occurred to me when typing the earlier comment was, what about the inside of a hollow piece. I know there's camera+software setups that can generate a CAD-like model of anything based on a set of photos, but how would they determine internal structure. Seems like those systems would have to be augmented with an additional, small flexible camera that can image the inside of the object.

Is there a vast decrease in time/waste product/investment when this is compared to something like a CNC mill? I've seen those things turn down 3-dimensional objects inside of other 3-dimension objects with only the slightest holes, so shapability (is that a word?) is not an argument.

So you do get part complexity for free with 3D Printing, but you sacrifice speed, surface quality, material properties, and tolerance - at least currently.

As far as waste, laser 3DP probably wastes a lot of energy compared with CNC milling. The laser is probably several kW for many hours - where's a similar, but less complex, part in a CNC would take minutes - like everything it really depends on the application.

Although you can make most shapes with a CNC machining center, there are definitely optimal shapes and non-optimal shapes. Seemingly simple things like the ratio between inside corner radius and depth of a pocket can drive the cost up. (eg. a small pocket radius forces the machine to use a smaller/weaker cutter, at depths the tool deflects/chatters and becomes less accurate). It's basically the core of design for manufacturability/lean design thought process.

3D printing avoids a lot of those concerns, but for most metal things I still think it's more trouble than it's worth.

I know modern CNC toolpath generators can be pretty complicated beasts, but I'm not sure to what extent they can use the sorts of constraints or tradeoffs you mention to optimise their flow.

The optimiser for a deposit-then-machine type workflow sounds even more interesting (and difficult). Is it worth printing those extra structures you're going to need to remove later? Maybe, if it means you can do some finicky machining operations in a different order later...

The bigger problem is the vast decrease in mechanical properties.

This is nothing new. I had very precise, laser sintered, steel objects in my hands during exhibitions, at least 5 years ago.

I never understood though, there seems to be no laser sinter aluminum. The most common answer is oxidation. But how hard can it be to do it in a vacum?

The breakthrough here is in speed, but the headline makes it sound like metal printing is new. Here are some old (and much more interesting) examples of metal-printed objects. https://www.bathsheba.com/

Her stuff is also produced without lasers - it's the ProMetal process which requires a bulk firing of parts to sinter the metal/adhesive composite the printer deposits.

Metal alloys are typically manufactured in a specific fashion to achieve the needed toughness, etc. For example, two identical steels with different heat treatments can mean the difference between a hard, brittle part and a "springy", malleable part. tl;dr: don't expect good knives from this tech.

No reason post-creation treatment can't be applied.

The whole point of 3d printing is to reduce labor and decentralize manufacturing. Also, some of those post-creation treatments are not possible.

I didn't see in the article how that one works but most of the metal 3D printers use metal powders mixed with a bonding agent which is activated by heating it with a laser.

I therefore don't think you could really heat treat the product as the metal isn't in its normal crystalline structure.

I think the main value for 3d printing is its easier to create the casts for molds. (that still raises the matter of whether or not casting is suitable in those application domains)

There are lots of 'print bureaus' now that render metal. Anyone can do it. The raw parts can be polished, machined, etc.

And the reaction is - abundant negative comments. Everything is amazing, and nobody is happy. http://www.youtube.com/watch?v=8r1CZTLk-Gk

Indeed.

Here are some nice titanium drops 3D printed for a bicycle in mass production: http://road.cc/content/feature/63359-exclusive-printing-tita...

More of this happens than people realise. I remember seeing those drops and wondered how they did it given the temperatures that you need to have to work titanium.

Naysaying is what HN is all about! We just prefer to call it 'criticism'. We like to say 'HA! I see a FLAW in your work here, that means I'm smarter than you!'

While this is awesome when application is help to mankind. What happens when people start printing guns, grenades and bullets?

Edit: No this is not flame-bait. I really feel that this type of innovation will end up within the ambit of "Government licensing and regulation model". Which also means goal to persuade "...manufacturing ... to household garages or backyards and places where you want quick production of parts in a very short time" will take a hit.

This is flamebait (perhaps not intentionally on your part), as this question usually flares up in the 3D printer mailing lists every so often with everyone weighing in.

Mainframe Computers was first used to crack codes and calculate ballistic missile trajectories. If you were around in 1979, when personalized computers came along, would you have asked asked "what happens when people start cracking secret codes to read wires and calculating ballistic missile trajectories and selling them to foreign gov'ts?" And if so, do you think it's an important question given what you know about computer usage in 2012? Do you ask yourself that question about computers now?

Perhaps pertinently, the government absolutely did ask that question, and restricted computer export for precisely the reasons you give. It might not be a valid concern in and of itself, but it's definitely going to be interesting to see how the government handle it.

Poor metaphor. If anything, the risks of hackers and proliferation of Wikileaks like sites have made this danger more of a reality.

Killing people is a trivial effort when it comes down to it; most people in the world would rather not kill people nor go to jail.

You've obviously never heard of the Sten submachine gun then?

You can knock one of those up with a pillar drill and a mid-range lathe/mill and some easy to get hold of raw materials:

http://en.wikipedia.org/wiki/Sten

Licensing and regulation of mills and lathes has not happened yet...

That is thoughtless flamebait. You can apply that fear to anything. Paper (what if people start planning crimes on paper), wooden sticks (what if people hit each other), cars (what if someone drives into someone else on purpose), computers (what if they are used to commit crime).

Not to mention lathes. Lathes have been able to make guns for decades/centuries.

Many of you have down voted, but here is what came to my mind: This technology, like it or not, will end up being regulated easily. Which also means that the goal to get manufacturing in household garages or backyards in a very short time will take a hit.

Isn't it unwarranted to push this call into something as negative as flame-bait?

What this technology can achieve is one thing, but will it be able achieve that as freely as it could is a question that I sought to ask. IMHO, it is not easy for technology people to convince and help other people (read decision makers) overcome their fear of misuse/abuse of any new technology. These people dislike the word "disruption".

Lathe machines, paper, Internet, mobile networks have all gone through this vetting at their time, and I believe that same barrier-to-explain will apply on 3D printing too. Only thing is, that it is going to be really tough to convince the authorities and allow "free usage in the backyards" as claimed in the article.

But it is completely irrelevant how things are manufactured. The end product is what needs regulation/laws. For example if you build a gun at home, then you own a gun and are bound to whatever laws exist for that in your jurisdiction. The means of creation are irrevelant.

"End product is what needs regulation/laws" is, I am afraid, a very narrow view. The degree of difficulty in manufacturing or obtaining (along with operational training) has a significant impact on use (or misuse) of weapons. Printing such objects in your household backyard, for example guns, will surely be a matter of "concern" for authorities who not only have to make laws and design regulation, but also enforce it.

Already being done:

http://defensedistributed.com/

People already create guns and bullets on their own with commonly available equipment.

Guns are low-tech.

John Browning, who produced most of the "modern" gun mechanisms, died in 1926, about the time that electric motors were becoming commonplace in factories.

Today there are folks making very functional guns with hand tools in very primitive conditions.

Yep. Afghan gunsmiths can recreate an AK47 from raw materials.

That's interesting. I would imagine you would need precision to the hundredth of the millimeter when making guns? I hung around some people who reloaded brass once and they were using these measuring gauges with that kind of precision.

It depends on what you want to do. Long-range or other forms of precision shooting are going to require more precision in the manufacture of the firearm and ammunition.

An AK47, on the other hand, well the whole design is based around very loose tolerances for ease of manufacture and for reliability - not as affected by dirt, mud, sand, etc as other designs.

> I would imagine you would need precision to the hundredth of the millimeter when making guns?

You do for some types of shooting, but in general no.

And, even if you did, what makes you think that 0.01mm requires modern equipment?

Moreover, tight assembled tolerances don't require tight part manufacturing accuracy. You can hand-fit parts.

Here's a an older but still accurate article:

http://www.armscollectors.com/darra/darra.htm

and a more recent one:

http://news.bbc.co.uk/2/hi/south_asia/5066860.stm

John Malkovich's character in "In the Line of Fire" made a wooden gun capable of firing two bullets (before becoming worthless), and the Mythbusters made a duct-tape cannon good for a few shots.

I believe mass-produced street guns will be more cost effective and reliable for the foreseeable future. However, the likelihood of this being used as a reason to regulate 3D printing or 3D design files shouldn't be ignored.

Probably just end up like color scanner and copiers that have "anti counterfeit" measures baked into the hardware.

"Error, I see your printing a spiral bored tube...."

Maybe future gun designs will include a EURion[1] mark?

The problem is you're building a utility item that can be constructed in literally an infinite number of ways, and defining some sort of 'Is this a gun?' test seems to me like verging on hard AI.

Of course, simple rules like 'this is 85% the same as that knockoff AK47 pattern distributed in the "Anarchists 3D Printing Handbook"' will probably catch the dumb ones.

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

Reply on news.ycombinator.com