3D Printering: Key Patents



Here’s a little tip about tech blogs, and journalism in general: absolutely everything you read is one hundred percent true, except in the cases where you – the reader – know anything about the story being discussed. Those stores on Wired and CNet where a device using an ARM Cortex-M3 is described as having, “the same CPU as a modern-day smart phone?” Totally legit, unless you know that running Android on such a chip is a virtual impossibility.

Such is the case with ‘key 3D printing patents set to expire in 2014′ – a phrase bandied about tech blogs with the fervency of news the seventh seal has been broken. If you believe everything you read on the Internet, we’re looking at a world of 3D printed lollipops, unicorns, and rainbows in just a few short months. Following the logic of journalistic veracity above, this obviously isn’t the case. What does the expiration of these patents actually mean, then?

Let’s Back Up A Bit Here

The current crop of 3D printers use fused deposition modelling, FDM, or the ‘squirting melted plastic’ method. This technique was patented in 1989 by [Scott Crump], co-founder of Stratasys, one of the largest manufacturers of 3D printers. This patent expired in 2009, and there’s no coincidence 3D printing really started to pick up around that time with the development of the Reprap Mendel and the founding of what was previously the Open Hardware community’s golden child, Makerbot.

If past results are any indication of future performance, the expiration of these key 3D printing patents will result in yet another boom in the field of one-off manufacturing, rapid prototyping, and some really cool projects coming out of hackerspaces in the next year or two.

And Here Are The Patents

The ‘key patents‘ (just search for [Carl R. Deckard] as the inventor if you want more) referenced by hundreds of articles spread out all over the Internet involve selective laser sintering. What is SLS, you ask? It’s actually pretty simple: take some powder, shoot it with a laser, let the powder melt, and put a dusting of new powder over the mess you just created. You can use a wide range of plastics with SLS compared to the FDM Repraps and Makerbots we have today; you can even print in metal and make yourself a rocket engine. If NASA is doing it, it has to be awesome, right?

So What Makes SLS So Great?

Even though the current lineup of ‘squirting plastic’ printers is fairly capable and can do a lot in the right hands, there’s some stuff an FDM machine such as a RepRap or Makerbot can’t do. Overhangs are possible, but for very intricate shapes – a one foot tall scale model of the Eiffel Tower, perhaps – you’re looking at a world of hurt. The only way an FDM machine could print something like that is with two filaments, using one material as a support and later dissolving it away.

The same goes with printing parts inside parts like the popular ‘ball in a cage’ carving project. No squirting plastic 3D printer can do this without supports, but an SLS machine makes it very, very easy.

SLS also allows for many, many different materials. While most FDM machines will not see a filament besides ABS and PLA, laser sintering machines can print in just about any powder that melts. Everything from nylon to polycarbonate to metals are possible with laser sintering.

Finally. lasers allow for much higher accuracy than the most common 3D printers. While very accurate FDM machines can print with an accuracy equal to that of a human hair, this isn’t the case for the majority of RepRappers out there. SLS simply doesn’t have the problems of oozing and misaligned layers so common in home-built printers.

Why You Won’t Have an SLS Printer in Your Garage

Oversimplifying everything a great deal, these printers are basically made of two parts: a laser cutter on top, and a plunger and roller system to build up parts layer by layer below.  Simple enough, right? Let’s just do some back-of-the envelope calculations on how much it would cost to build our own SLS printer.

First things first. We’re going to need something that moves a laser beam around on an XY plane. Here’s a fantastic Open Source laser cutter that does just that. The BOM lists the component costs (minus a laser tube) at about $850. Throw in an eBay CO2 laser tube from China, and you’re looking at a fully functioning laser cutter for just about $1100. That’s awesome, even though it’s right about in line with the cheap, smaller-capacity Chinese-made laser cutters you can get via the usual channels for about the same price.

That’s half of our build right there. Now all we need is some sort of roller to dispense the powder and a plunger mechanism to build a part layer by layer. This is where things get a little more difficult. You’re probably going to need some sort of sheet metal build tank to hold all that powder, and the plunger will need to work at some fairly tight tolerances. The roller is simple, but you’ll also need some way to (somewhat) evenly spread the powder in front of the roller. In the end, you’ll probably looking at around $2000-$3000 for a low-end, home built SLS printer.

Here’s the problem, though: we’re around the price point of a Makerbot or Ultimaker – both proven machines – and an SLS machine is not going to be that much better. You’re still basically only working with plastics, and while you don’t have to deal with support structures on our DIY laser 3D printer, you’re not doing anything that can’t already be done with a stereolithography printer like the resin-based Form 1 printer for the same price.

ElectronAs for printing in metals, that’s a pipe dream for any machine cheaper than a car. Sintering metal with a laser requires a vacuum chamber, diffusion pumps, and some very hard core equipment to do it right. Not to mention you won’t be able to melt any appreciable amount of metal with a 40 Watt laser. While there has been some progress with a similar project called MetallicaRap, the MetallicaRap team estimates their final kit will cost about €10,000, out of the range of just about every hobbyist or hackerspace.

By the way, the MetallicaRap project is really awesome and you should think about a small donation to the project.

Basically, if you don’t know how to build an electron microscope or fusion reactor in your basement, you don’t have the skill set to design a machine that can print a usable metal part.

If not homebrew, then what?

The reason everyone is so excited by the expiration of ‘key patents’ is the fact that other large companies besides 3D systems – Stratasys and Zcorp, for example – will be able to manufacture their own SLS printers. That’s great and all, but even 3D systems, the maker of these SLS printers only use them for their professional range. The bulk of professional printers produced by these companies use a method similar to SLS – using an inkjet to spray binder onto powder – but this isn’t covered by the SLS patents.

If anything, the expiration of these key patents will mean a reduction in cost for the very, very high end printers. The stuff NASA uses. Of course a few large companies will be using this tech for custom, one-off parts, but don’t expect to see any laser sintered parts show up in your car’s engine any time soon. Remember, 3D printing and rapid manufacturing is only ideal when you need to make a handful of parts. Anything more, and you’re better off with traditional manufacturing methods.

In Closing…

Will the expiration of key 3D printing patents in 2014 change anything in the arena of 3D printing?  Well, large, already established 3D printer manufacturers will be putting out cheaper printers that can print in metal. You’ll be able to buy a 3D printed rocket engine for a few hundred dollars instead of a few thousand. Once Shapeways gets their hands on one of these machines, you’ll probably see a few extremely tiny internal combustion engines built by hobbyists. There will probably be a flood of combination 3D printer / laser cutter machines on Kickstarter. Other than that? It’s going to be cool, but patent expirations aren’t going to change the world overnight.

60 thoughts on “3D Printering: Key Patents

  1. Put the laser on a printed scara arm,use a cylinder with a plate that drops down each layer, can’t be much more than a high end reprap?

      1. What waste/mess?
        The unused powder is simply put back in the supply for the next print. There is only mess/waste if you get it VERY wrong.
        All you need is a single arm that slides over the print surface with a slit along it and a hopper of powder on top. As it moves a thin layer of powder is deposited, when it gets to the other end it stops, Laser does its job and the print surface moves down a notch, slide the hopper back over and rinse/repeat.
        The print surface would have to move down for practicality so you would need a deep “bin” containing the z axis components and a good seal at the sides to prevent powder escaping.

        1. Actually the powder doesn’t recycle all that well. It has to be blended with a batch of virgin powder(I assure you this does not mean powdered virgins), due to the fact that thermal cycling degrades the powder.

          1. What about high-temperature printing, where the bottom layer is kept precisely at just below the melting point of the metal, so it’s a little tacky? Then it should not cost too much to keep the upper layer energized just enough to melt the new layers on, and while it would have limitations on thickness etc and tolerance of finished part, it might be cheap to make (maybe not so cheap to run).

    1. @Liam Jackson
      “Put the laser on a printed scara arm,use a cylinder with a plate that drops down each layer, can’t be much more than a high end reprap”
      I have a scara robot and want to tell you that the tolerances of the bearings in it are incredible, and mine has very-high-count encoders on each joint. It is not possible to replicate that with any “printed scara arm” that I know of. The final accuracy of less than .001″ true position over the whole range of the arm is no easy feat, and no amount of determination and willpower, will make a 3d printed arm to the specifications of the scara arms.
      Considering that the scara arms are serial, one mounted at the end of the other, The errors are compounded–the first arm is off radially a given amount, the second arm adds its own errors to the previous ones, which makes .001″ accuracy that much more impressive. Don’t forget that in a 3d printer, the Z axis has to be the same level of accuracy as the x and y axis, so similar encoders and vertical slide have to be employed. If you succeeded. the cost would be that more than a shelf full of repraps.

      1. 1st: you would need the same tolerances of the bearings you have on your commercial bot.
        2nd: You would need to use servo motors with the same repeatability as your commercial bot has.
        3rd and most important: you would need to run it at a much slower speed than what your commercial bot is capable of. As long as there is no deflection in the 3D printed arm, you could then get the same accuracy.

        FYI: I am assuming that your bot is a closed loop system in which the controller gets feed back from the servos when a particular step has been achieved. It then goes on to the next instruction set.

        As far as repeatability is concerned, as long as the speed does not surpass the ability of the 3D printed part to maintain it’s constant shape, the accuracy can be achieved.
        One advantage the 3D printed part would have over your commercial bot is less mass which would aid it in maintaining it’s form.

        It can be done, but not if the operator is trying to run it as fast as your commercial bot. I think that the limitation there should be obvious. On the other hand, we’ve all fallen for our own delusions now and then, and we would come across a user trying to use the 3D printed version on the same jobs that your bot was made for. As far as that working, it’s not gonna.

        1. I get what you are saying, that it can be done, it would just run slower.
          But that is not the case. The premise from liam is that this would cost not much more than a high-end reprap.

          You start by assuming you have the same bearings and hi-count encoders–magically, somehow–because each axis’ bearings and encoders cost more than the highend reprap liam is referring to. Nobody i s 3d printing to +/-.0001″ tolerances for bearing fits, and similar tolerances for straightness and squareness of inline bores.
          That is currently the domain of precision machining.

          Rigidity is not a matter of just go slower, and it will be OK.
          Starts with material selection and you cannot compare something made from cast iron to something make from abs and expect the same rigidity–but you are saying that you can do that, just go slower.

          Sorry, I love 3d printing, I have a printer and am building my own as well, but they don’t “do everything, just slower”.

          1. I agree with your points.
            In addition to choosing the correct material, you have to design it to be rigid.
            A 36 inch long 1 inch diameter rod will never be as strong or ridged as the
            same amount of material properly caste into a pipe.

            The pipe having natural braces within it’s design will hold more weight than the rod of the same mass.

  2. While this article pretty clearly states that the current round of expiring patents won’t change the world that much it kind of sounds like the expiration of the 1989 patent back in 2009 did.

    Does this mean that were it not for our 20-year patent system we could have had the RepRap much sooner? If so then where would we have been by now?

    Isn’t 20 years almost an eternity as far as technology is concerned?

    1. 20 years an eternity? In software, yes. In hardware, in the broad sense, I’d say no. The nominal cost of manufactured economic goods is much higher than software. Whereas the cost to crank out the millionth copy of Windows X rapidly tends to zero, not so much the millionth Makerbot.

      1. Personal computers became common in 1980. 33 years ago. Look at the leaps and bounds that were made in 33 years. From 1990 to 2011 (21 years), worldwide mobile phone subscriptions grew from 12.4 million to over 6 billion. Do you remember what the 1990 cellphone looked like? I would say 20 years for hardware technology is indeed an eternity.

        1. Not all technology is cell phones. Processing power, displays, and so on have made tremendous leaps, but manufacturing technology much less so. CNC machines and 3D printers don’t run on cutting-edge processors.

        2. Not all hardware is cell phones. Aircraft, radar, washing machines, cargo ships are also technology and 20 years (except maybe washing machines) is still usable life.

    2. The whole patent system truly sucks. It was set in place to allow small people to make money from things they invent and ended up being protection for big companies. No individual can afford to take the risk and challenge the likes of Sony and Cisco, the big boys just walk over any small enterprise or individual.There are also serious problems with what is allowed to be patented, Dyson’s cyclone that has been used for years to sort crap from seed, Heated chambers for containing things that need to be warm, get real, we would have no chickens if that hadn’t been invented. My favourite has to be one of mine that expires this year about putting high frequency signals on powerlines. It should never have been granted and it should also never have stood up to the challenge that Siemens put forward. The whole idea of patents needs a complete rethink. Having spouted all that verbage there is no reason why these patents should stop any hobbyist from making things, they do not prevent anyone from making stuff they only stop people making money from stuff without paying some reward to the owner of the patent. The owner of the patent cannot refuse to grant a license if requested and that license has to be “reasonable”. << Whatever that means. Personally I just hate them and would be quite happy for it all to be done away with. First to market is now how it works because things change so fast, keep it under your hat, make it and get it out there.

  3. Carl Deckard got his initial patents for SLS while at UT-Austin, and I worked with him some when I was getting my degree there. The SLS machines can be done DIY, after all, that’s what his first machine was. But it is very challenging- Carl mentioned that the laser can’t just trace the shape that you want in the powder, because as the powder melts, it will wick and puddle and create different shapes than intended. Carl had to develop the algorithms to determine where you wanted the lasers pointed to generate a particular shape. Also, the method of rolling a extremely thin, uniform layer of powder was a much bigger challenge than Carl had anticipated, and I’m sure a DIYer will go through many trial and errors before they get this right.

    Other cons for the machine is that the powder is/was expensive, especially the metal. The plastic is probably not that much more than the overpriced plastic spools for FDM machines, but the powder has to be a very specific diameter, and there probably won’t be cheap sources for it for awhile.

    Also, the machines are really big for the work area size you get- no table top machines here.

    Despite these downsides, these are the ultimate 3d printers. Think about- you can use anything that melts as your powder, and you can truly make any shape- overhangs, hollow parts, whatever. Also, these parts can be structural, they will not delaminate like FDM parts.

  4. You forgot about the temperature controlled inert gas build chamber, that’s another thing that makes Laser sintering(SLS is a trademark of 3d systems) really hard to do. The other is that the sintering equation dictates that laser sintering is VERY SENSITIVE to temperature! Which is why you have to have a heated bed and do all sorts of wizardry to keep your powder bed’s temperature uniform.

    Of course, the key patents on laser sintering have already expired. A company called EOS has been selling laser sintering machines that don’t use Carl Deckard’s patents in the US for quite some time now.

  5. not so nice.
    using “squirting” when speaking about fdm shows your lack of appreciation about the majority of 3d printers.

    overhangs are possible with only one material if you use a decent slicer.

    moreover makerbot is not a golder child but a traitor

    this trollticle really looks like it was written by a professional sls 3d printer seller.

    1. You know what. I agree with this. Sure there are problems with SLS – but surely the whole point of hacking is to find creative solutions to these problems.

    2. Yeah thats completly right, it was also my first thougth…
      I´m currently on calculating and thinking about the specs of a DIY-3D SLS Printer, it seems to be hard but there is no reason that this issue is claimed as “impossible”- especially on a page like hackaday!

  6. And besides pla and abs, one can use hdpe and nylon, with nice coloring effects.

    If you want to learn about serious reprap 3d printing, follow the richrap blog.

  7. Very good writeup.

    But I think being able to perform SLS with plastics on home machines would be a great advance, even if metals are still out of reach. The lasers would certainly be affordable. Plus a variety of powder coating media would likely work for the raw material, and is already cheap and widely available; so there’s probably no need to build up a hobbyist supply chain like was necessary for FDM filament.

    I’ve also seen some technique where a liquid was selectively printed onto each layer of powdered metal, in a pattern, using an inkjet printhead. The liquid prevents fusion where applied. Then a cheaper heat source like an infrared lamp was passed over the entire layer, fusing only the untreated metal powder. I can’t remember the name of the technique, but it’s food for thought; perhaps we don’t need lasers, much less massive ones, to produce metal parts after all.

    Are there other ways? Could you sinter small spots of metal powder with a tiny induction/spot/plasma welder instead of a laser? I don’t think it’s been explored, big industry has no reason to since they can afford the big lasers.

  8. I don’t necessarily need my very own metal sintering machine taking up space in my garage. But if there’s one nearby that can print a strong part and send it via UPS to my doorstep within a day or so, at a reasonable cost, I think I could find a use for it.

    So anything that reduces the cost of those machines is good, even if they still remain impractically expensive for typical hobbyists.

  9. SLS of metals might be possible in a DIY realm if a powder version of the Precious metal clays becomes available. PMC is a play dough like material composed of finely divided metal powder suspended in a plastic binder. When heated to 4-500 degrees the plastic burns out bringing the metal grains into contact and causing them to fuse. The result is a porous metal structure about 2% smaller than the original but perfectly to scale. Currently silver, gold and bronze are available. Just sayin’

  10. Ahem,


    Just a couple of things. Namely that link above. 1st the main cost of the buildlog.net 3d printer is the frame. A lot of money is spent on extruded aluminum. The second most expensive part is the laser, designed to cut things. So if you eliminate the expensive frame, and you reduce the power requirement of the laser, you might have a chance of developing a printer in 6 months at a price of $469? Granted you might be only be able to print in wax, and still have some development left.

    The facts in this article were fascinating, They would go a long way in helping someone understand the environment around SLS. The opinions and guesswork only serve to dissuade someone from trying and serve for nothing. Please stop it. For a blog that promotes creativity this post needs work.

  11. “absolutely everything you read is one hundred percent true, except in the cases where you – the reader – know anything about the story being discussed.”

    Best HaD quote to date.

      1. Bad part about that quote is that it is false.
        Regardless of whether the reader knows about the subject matter or not, if the article is false, it is false. The facts do not change by virtue of the readers knowledge and therefore the lies are lies regardless of whether or not the reader believes them.

        The only real difference between a reader that has knowledge of the subject and a reader that has no knowledge of the subject, is the potential for deception. It is more probable that a reader without knowledge of the subject will be deceived by a deceptive article than a reader with knowledge of the subject.

        Regardless of whether an author successfully deceives a reader with deceptive writing or not, the false statements in the article remain 100% lies. There is no such thing as a partial truth. A statement is either true or false. There is no partials in the matter.

        1. You need to take it one step further.

          “Everything in this article is true, because I don’t know anything about it.”
          “But someone reading this actually has first-hand experience with this”
          “Therefore, just about everything in this article is either false or a gross oversimplification.”

          It’s not a matter of deception, it’s just a function of the impossibility of truthfully conveying the whole of the reality discussed in print. I haven’t yet decided if this observation is due to the inability to reach objective truth in natural language, or the fact that journalists are really, really lazy.

  12. Someone could always try and build a fdm style system with a mig welder style head. You’d have to cut the build platform away from your print & you’d be working with some high voltages. it’s need to be built in a box that could be filled with inert gas to prevent spluttering. But fillament can come really fine for a mig. That might be a cheaper alternative…. you’d definately have to keep it outside or in a well ventalated area…

  13. SLS of metals is really hard. SLS of plastics is hard. Just laser a bed of UV resin. Preferably with a galvo laser and a single precision Z stage and a stepper. Hobbyist grade, cheap, ultra high quality, low cost. Seriously not that hard, it’s been done for 25+ years now.

  14. Here I was, thinking about layering 2mm steel, welding it, then milling the fine detail before adding the next layer. Then using that to make a cast, for the mass production line.
    Silly me.

  15. you can get makerbot comparable prints from a sub $1000 printer. Let’s see what happens with SLS in a few years time. Some materials can’t be used with an SLS printer, but can be used with an FDM. It goes two ways really.

  16. This article is naive. There are probably already politically backed or heavily funded entities filing on these patents so they can do vendor lock and expensive licensing..

    If people aren’t publishing SLS kit schematics then it’s probably because they don’t want to or they only care about the financial market aspect, where they can’t profit because of the patents..


    Where exactly do you get off posting intelligent, balanced, reasoned, well-researched and correctly-spelled articles on here?

    I demand an immediate apology in the form of a series of articles on how Ikea cutlery is a cheap and practical alternative to 3-pin mains plugs, and haters be damned.

    1. LOL I was waiting for the rant about how Hackaday sold out, blah blah. Well positioned sir.

      I for one am not unhappy at all with the potential refinement of something I read every day.

  18. SLS is way cool, but like you said, it’s out of our range, but you sure went on about it. Kind of a long article just to end up feeling more frustrated about not having an SLS than curious about the patent rumor. A lot of folks here have cheapo ‘squirter’ printers, maybe it would be more interesting to focus on what we can do with what we got than what we can’t do without something we can’t afford.

  19. I’m surprised no one has mentioned using SLS/FDM to create investment casting patterns. Multiple unique patterns can be attached to a tree and all cast at once. It would make quick turnaround, low volume production of almost any material practical. Startup costs would be significant, but there’s a large market for prototyping in materials other than plastic resins.

    This sort of thing is already being done w/ low volume metal stampings. A WWII style Liberator pistol design using stampings and spot welding would be much easier and more practical than the ABS version.

    Ruger pioneered investment casting of steel gun parts long ago. Some machining is generally required (e.g. precision holes, etc), but it’s much cheaper than traditional forging and machining.

    FWIW there’s a company selling SLS manufactured muzzle brakes for AR style rifles


    1. The tree huggers will be opposed to attaching _anything_ to their beloved trees, with exception to whatever they attach to their beloved trees… B^)

  20. This article is not very accurate and probably not enough research was done when.
    Let me remind you that SLS process can be applied to plastics also reducing a lot the price guessed. Sure metal is cool but lets start with the basics.
    I can say for sure that in 2 years there will be SLS printers (not metal) in the 500~1000$ range.

  21. I am looking forward to the day where I can design a high flow intake manifold in a 3D modeling program, send it off to a print shop and have it attached to the engine in less than a week. Although that might be a while by the sounds of it……

    1. Ooh yeah BABY! (Seriously…) putting fuel injectors on my brothers Farmall Super M
      (someone stole the carburetor) or replacing those lousy Hitachi carbs on my old Virago…
      Make it rain!!!

  22. This is powder deposition printing with fusing instead of binding.
    There will be many open sourced hobby level laser 3d printers out there in the next couple of years for plastics at least.

    Metals will be a bit further down the road but there is a good intermediate step.

    At the RepRap level of production and tolerances we should be looking to the traditional powder metal industry for clues. First the process is separated between shape and sinter. That allows for multiple points of control.

    A powder deposition printer with the appropriate binder in the head will get you a green part at least at tap density.
    Sintering could be accomplish at least at first in a kiln.

  23. The cost of the laser stated above is just for the frame. Add the tube, laser power supply, low voltage power supply, stepper drivers, DSP, optics, water pump, air pump, and other bits and pieces and you are looking at $2000-2500 easy. I know because I just finished mine last month.
    That being said I would be willing spend a couple grand building a SLS printer!
    Bring it on!

  24. Interesting article and comments. I built an SLS machine a couple years ago (http://reprap.org/wiki/SLS_wax_printer / http://andreasbastian.com/3dp/videos.html) and specifically targeted a low-energy material to side-step the higher energy material/processes that normally make this technology inaccessible to the DIY 3D printing community. While I had some success with that machine, the lasers that I used were simply too low of a power to be terribly practical. Additionally, the lack of a heated build chamber and more sophisticated SLS-specific scanning algorithms led to high amounts of warping.

    However, as mentioned above, SLS machines are basically a laser cutter and some hardware to manage powder. As the cost of 3D printers has dropped over the past few years, so have the costs of laser cutters. I’m currently developing an SLS experimentation platform as a drop-in addition to laser cutters. I started this project back in August as a fellow at the Advanced Manufacturing Research Institute (AMRI– see more at http://www.amrinstitute.org) and am documenting progress at http://opensls.tumblr.com. I think that using CO2 tubes is the way to go for this early stage of DIY SLS experimentation– they are used in commercial SLS systems and for the power range are the highest wattage per dollar.

  25. Hi guys, interesting thoughts,,
    Can (SLS) provide consistent quality in a large build envelops (ex. 4*4*4 meters X,Y,Z)?
    if Yes, where can I get more help to start building one as soon the patent expires we have the fund we lake the brains.
    If No, why?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s