The IP Of The Infinite Build Volume 3D Printer

Last week, the Blackbelt 3D printer launched on Kickstarter. What makes the Blackbelt 3D printer different than any other 3D printer on Kickstarter? This printer has an infinite build volume. It’s built for continuous production. As long as you have a large enough spool of filament, this printer will keep producing plastic parts with no downtime in between. The Blackbelt is a truly remarkable and innovative machine. Yes, it’s a bit expensive, but it’s designed for production and manufacturing, not some guy tinkering in his garage.

However, the Blackbelt 3D website includes two words that have sent the 3D printer community into an uproar. ‘Patent Pending’ is something no one in the community wants to see given the history of the industry and a few poor decisions from the first movers during the great 3D printer awakening of 2010. The idea of an infinite build volume printer that allows for continuous production is not new; we saw one last March at the Midwest RepRap Festival. The question, therefore, is what is covered by the upcoming Blackbelt patents, what is the prior art, and is it still possible to build an Open Source printer that uses these innovative techniques?

MakerBot’s Automated Build Plate. Once available as Open Source Hardware, the Automated Build Plate has been expunged from MakerBot literature, patented, and apparently forgotten. [Makerbot CC-BY]

Lessons of the Automated Build Platform

Questions about the Blackbelt printer arose shortly after its soft launch last month. It was, simply, the second printer demonstrated in a few months that uses a tilted print plane and a conveyor to allow continuous production in an infinite build volume.

The first such public implementation of this design was at Rapid 2016 and at the Midwest RepRap Festival in March, a product of [Bill Steele]. [Steele] wasn’t releasing a product, this was just the culmination of an idea that began as a mechatronic middle finger to MakerBot and their Automated Build Platform (ABP).

The ABP was quite clever when it was released and was used in production by MakerBot in their salad days to manufacture parts for the Thing-O-Matic. Unfortunately, the APB was patented by MakerBot, all references to the ABP were expunged from the MakerBot site, and all development on an Open Source conveyor-based production machine stopped.

In short, the 3D printer community has seen something like this before. First, the community creates an innovative device that makes printers better, then a patent is issued. Regardless of the success of the company holding the patent, Open development on this sort of device simply stops, and we’re left waiting decades until the patent expires. This has happened before, and it’ll happen again.

Prior Art to Prior Art

[Bill Steele] first demonstrated his unnamed infinite build volume printer at Rapid 2016 and at the Midwest RepRap Festival in March of 2017. However, unbeknownst to everyone, [Andreas Bastian] of Autodesk has been working on a similar device for years. The Lum Printer is effectively the same machine as demonstrated by [Steele]; a conveyor belt bed over a tilted XY extrusion plane allows for prints of unbounded length. There are videos of this printer working, and while the Lum printer was never used in its full capacity — most demonstrations are very long single layer panels — it’s effectively the same printer.

The Blackbelt Patent

The Blackbelt 3D is still patent pending, and we don’t have any idea of what is claimed by these patents. However, Blackbelt was kind enough to share that they are only claiming, “the belt material, an adjustable angle for the extrusion plane, and G-Code manipulation.” For an Open Source implementer of the infinite build volume printer, everything else is fair game.

An Open Challenge To The RepRap Community

Although the Blackbelt patent will cover a variably tilted bed, the belt material, and a method to transform G-Code so any slicer can use this printer, that doesn’t mean the idea of an Open Source, infinite volume printer is out of the question. The only thing anyone needs to do is simply build one with a permissive license.

The product of a G-code shifter. From wjsteele.

This is a challenge to the entire 3D printing community. Come up with a printer design that uses a bed tilted 45 degrees to the print plane, and find a suitable belt material. The rewards will be enormous. To get you started, [Bill Steele]’s MRRF build used Kapton and paper. I’ve done a bit of research on this, and I suspect there may be a very strange source of belt material: the pancake robots at every Holiday Inn Express use a Teflon-coated silicone belt that’s just the right size for a 3D printer. The manufacturer of these pancake robots sells the belts as replacement parts.

Apart from the belt material, the only other bit of tech needed to create a tilted bed 3D printer is a G-code manipulator. For this type of printer, a different method of slicing is not needed; the Blackbelt patent will describe a ‘warp engine’ that manipulates raw G-code. Luckily for us, [Steele] has uploaded his own G-Code Shifter. The G-Code problem for a tilted bed printer is solved, and it’s Open Source.

Really, the only thing needed for an Open Source 3D printer is a bed material and a design. Of course, buying more than 5kg of filament on a single roll will also be a problem, but if that’s the biggest problem we’re all in a great place.

48 thoughts on “The IP Of The Infinite Build Volume 3D Printer

  1. “As long as you have a large enough spool of filament, this printer will keep producing plastic parts with no downtime in between.”

    Sounds like when I screw up in Infinifactory.

  2. Maybe I’m naive, but it seems to me that putting a belt under a printer would not be patentable because it is obvious to a person familiar with the art. It’s called a production line. We’ve been using it for a century. Some of the candy and food manufacturing is using the exact same concept, extruding material onto a moving mat.

  3. Maybe I’m naive, but it seems to me that putting a belt under a printer would not be patentable because it is obvious to a person familiar with the art. It’s called a production line. We’ve been using it for a century. Some of the candy and food manufacturing is using the exact same concept, extruding material onto a moving mat.

      1. Yeah, basically.

        I’m not lawyer, or judge, or patent officer. However in my experience this will get initially rejected because someone at the patent office will say “Well yeah, duh, people have been doing this with candy forever”. However, the applicant will then respond with “That is outside the scope of our art. What we are doing pertains to the development of plastic parts using this process”. They will probably get some sort of utility patent covering a specific use of a belt in a 3d printer.

        Just my thoughts.

  4. I am not a lawyer, but I seem to recall that once you publish something (hence making it “open”) you can no longer pattent it… Can anyone with a real law degree comment on this?

    1. (ianal but…)
      Different countries (read: patent systems) have different rules – the USPTO gives a year to file following disclosure whereas the UK & Europe operate whereby if you disclose something, it can become ‘prior-art’ so filing must be done first.
      It should also be said that being awarded the patent typically happens several years after filing but the critical bit is the filing date (“patent pending” means it has been filed), not the final grant date.

    2. I am not a licensed US attorney but my understanding is that in the US at least, they have moved to a first to file system a few years back, instead of a first to invent. Prior art still counts in that system but it is no longer “prove you invented it first” and you can go patent it like it used to be for many years. If you invent it first, don’t patent it or publish it and then somebody else patents it instead (even years later), they get the patent (and pay the patent application fees). Such a system incentivizes people to pay for more patents as a patent land grab of sorts.

      1. This is correct, it was becoming to difficult and lengthy a process to prove “I invented this first”. It often came down to “Well my engineer brought up the idea of using a touch pad to control a phone years ago, here are the minutes of the meeting” against someone who three years later actually developed something. Apple did this to a friend of mine a couple years ago.

        There is a law of gray area in the “first to invent” process. It also leads to lengthy legal proceedings. Even if you can show you actually “invented” first, sometimes the cost and time of court isn’t worth the effort for an individual or small company, whereas for a company with a large stockpile of resources and a product revolving around an obscure existing patent it certainly is.

  5. The Automated Build Plate i seem to remember being available officially from UK suppliers, where it was advertised as open hardware (i could be wrong). If that is the case then it should be as advertised. I wonder if this would be a legal argument in the UK courts under our advertising/trading standards laws to use against patent laws. I don’t believe patents have been challenged in this way before but would make for interesting test cases.

  6. I’m sure they will get it anyway but I really think they shouldn’t get the patent on the adjustable slope. Maybe they are the first to do it… I don’t know but.. it’s a 3d printer. EVERYTHING is adjustable… finicky, prone to needing re-adjusting, etc… It’s just the nature of the technology. Surely making something adjustable can’t be innovative!

  7. Lucky me, I am immune to patents!



    Step 1, see an idea
    Step 2, decide I want to build it
    Step 3, start collecting parts
    Step 4, build

    1 and 2 I do almost instantaneously…
    Actually getting to step 4… well.. by then the patent is expired!

    HA! Take that Blackbelt!!!!
    With my awesome powers of procrastination I own you!

      1. I can’t speak for other countries but in the US.. technically this is not true.

        If you only build one for yourself and you don’t brag about it online… What they don’t know can’t hurt you.
        And if they did find out… is a corporation going to even bother or care about your one-off? Well.. if you showed it off online I suppose they might since it could be seen as encouraging others to do the same.

        Anyway.. I’d be surprised if anything came of it but the law would not be on your side.

  8. It takes a few days to finish a 5kg spool of filament. If you have two printheads you can load the next spool into the second head and continue with that once the first head is empty. A simple microswitch can already detect “no filament”.

    The G-CODE manipulation is not essential. I think you lose some resolution with that. I’d rather adjust the slicer to do that. At least… the only think I can think of is that you reconstruct a solid from the G-code and reslice. All paths in the g-code are invalid in the 45 degree situation.

    1. The GCode manipulation is essential, as it’s what causes the part to “walk” down the conveyor. The Blackbelt team used a “warper” to modify the object STL itself, but then it gets sliced to produce gcode that is already shifted. If you change angles or layer heights, you will always need to reslice it regardless of the method.

      My Gcode Shifter simply is a post processor that occurs as soon as the slicer completes it job. It modifies the gcode to account for the forward motion and downward motion of the belt. It doesn’t actually modify the geometry of the object itself.

    1. We’re solving two entirely different problems here. A five axis head on a conveyor would be very cool… as it would allow for both problems to be solved in one machine. Conveyor’s allow for continuous building (or infinite build volumes,) five axis allows for support free printing (but within a limited build volume.)

  9. Few problems I see.

    1) I imagine first layer adhesion is a pain, not being able to squish the part into the platform at all to get good adherence.

    2) I seriously laughed out loud at the “This example shows a txpe of overhang, which requires support material on regular 3D printers. With the BLACKELT, this is not nessecary anymore.” statement. All it does it add an offset to the unsupported print area. If you mirror their example, you get screwed on needing support material.

    3) The thing-o-matic had a conveyor belt on it. If you get good first layer adhesion, the part still warps because it is greater then the belt tension. I would imagine thats going to happen no matter what material you make the belt out of on this printer.

    1. 1) The layer adhesion on the first layer isn’t really an issue at all. In fact, I’ve found out that it doesn’t need anything special done to just get it to stick as long as the height is calibrated correctly.

      2) Yes, when I saw their claims about not needing supports, I laughed as well. You do need supports… it’s just that they are shifted by the amount of the build angle.

      3) That’s why they’re using Carbon Fiber and I’m using Stainless Steel… you need excellent tensile strength on the belt to prevent it from warping. Also, don’t forget, that the old Makerbot machines were only use ABS, which warps under normal circumstances… we’re using PLA which is way more temperature stable and easier to keep straight.

  10. Maybe I’m overly thinking this, but why do you need to tilt the bed approx 45 degrees? Why not have the back end of the printer (with adjustable feet for example), or the bed/print head portion only, tilted (with adjustable threaded knobs)? Then you wouldn’t have to manipulate the G-Code, right?

    1. I would say that 1) it puts a little gravity back into the print plane because you would otherwise be printing on a vertical wall (assuming the belt doesn’t reverse) and 2) it allows the nozzle to get all the way to the print surface without having to make the rest of the extrude overhang the end roller. Imagine trying to get the tip of a screw driver into a corner while holding it perfectly horizontal to the one of the surfaces. If parallel, the handle will hit first.

  11. hey brian, have you not had your coffee yet?

    the difference withthe blackbelt printer is that the belt moves horizontally parallel with the table surface or ground and the angle of the extrusion plane is what they modify, hence it being variable. This means that in the blackbelt the belt stays flat and the x/y plane that the print head moves on is hinged at a point and can vary the angle with respect to the bed.

    this also explains why they need a patent on the Gcode warper.. with a variable angle on the print head plane you would be doing some interesting manipulation of the Gcode which wouldn’t be covered by prior art as everything to this date has used a fixed angle between the bed and the extrusion head movement plane and thus is not the same printer at all.

      1. i stand corrected, but does your code allow for the angle to be changed while printing? i think that is what the people at blackbelt are implying.. and from all the pictures ive seen it looks as if thats how the machine operates as in the plane that the extruder head is on changes as the print is in progress. i could be wrong, its just my interpretation of their statement combined with looking at available pictures of their machine.

        1. No, the Blackbelt’s angle for a particular machine is fixed… it’s bolted in place. They have several different versions of the machine at different angles. Andreas’ machine could be adjusted dynamically, however. Mine is fixed, but the slicer can adjust.

          As for shifting dynamically, you can’t really do that unless you reslice the part, because as soon as you rotate the bed, you need to recompute all the vertices and shift at a different angle. Now, between objects, that’s not a problem, but on the same object, it wouldn’t work.

  12. Hey Brian, the article has the CEO of Blackbelt showing off the Lum printer… but that’s not correct. The Lum was designed by Andreas Bastian of Autodesk. Andreas’ design supersedes all other prior art that I am aware of.

  13. I’d be very interested in writing/modifying software to create the g-code in support of a rotational print bed or even a static one at 45 degree angle, what OSS software would everyone prefer?

  14. I would like to say two things.

    1. The recently enacted “America Invents Act” makes their patent pointless. The USA now has a first-to-file system which treats public disclosure of a pending patent as prior art… if not disclosed by the inventor. So if anyone other than the inventor details a patent pending concept while not attributing the inventor….it becomes prior art and the examiner just has to be notified of it to halt the patent indefinitely. It’s a stupid law that completely screws over the previous one year grace period for investing or academic research papers, but meh, they’re already acting on it.

    2. The 3D printer community RARELY creates an innovative device that makes printing better. Nearly EVERY HaD ‘innovation’ can be found on expensive industrial 3D printers years before. Like the inevitable winch-bot-printer, this concept was done by the big players years ago. Maybe, like past community innovations, they simply found a way to make it super cheap and flaky…

  15. Wow, their website is TERRIBLE!

    Any effort by them to patent this is objectionable IMO. Nothing here should be patentable (not saying you couldn’t get one issued, but it shouldn’t be) and anybody who would want to doesn’t deserve my money.

    Not to say that I don’t purchase patented things, but this is just stupid…

  16. The infinite build volume claims are BS. These can’t build anything longer than the print bed because the print can’t be rolled off either end without coming unstuck. Even if the print is only rolled partway off the bottom end, the weight of the overhang would pry the top end up.

    1. Not sure where you getting that from. I’ve printed many things much larger than the print bed. In my case, during a convention, I printed a four foot part. As long as the bed’s tensile strength is strong enough and the adhesion is fine, the part can easily extend past the length the bed. It’s even easier to make long chain like objects, which one could argue are a bunch small parts, but printed as a single object that is infinitely long.

Leave a Reply

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

You are commenting using your 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