I’ve said this before, and I’ll say it again: if you’re using a 3D printer to make a few hundred identical plastic parts, you’re doing it wrong. That’s the place for traditional manufacturing methods such as injection molding or resin casting. If, however, you’re looking at printing a few dozen identical plastic parts, or even running a script to optimize your machine time, the current open source 3D printer world leaves one thing to be desired.
An Automated Build Platform
An Automated Build Platform is a fairly simple idea: put a conveyor belt on your heated bed, and when the print is done, send a command to drive a motor, dumping the newly printed part into a bin, The printer then begins the next part with a clean bed, and the days of doting over a 3D printer soon fade into the past.
For such a simple and useful idea, it’s surprising there hasn’t been much done with this idea in open source circles. There are, of course, problems both technical and legal, but hopefully nothing that should indefinitely derail anyone who would want to create the first open source automated build platform.
The Problem Space
For anyone who has been following 3D printers for a while now, the idea of an automated build platform will sound very familiar. Makerbot sold one for their first generation 3D printer, the Cupcake. Since then it’s been taken off the Makerbot store, and apparently expunged from all Makerbot literature; even the Makerbot Wiki page for the automated build plate has been removed. That doesn’t mean you still can’t buy the parts for an ABP, you’re just not getting any support or documentation from Makerbot.
“But Makerbot is terribly horrible and other ad hominem attacks!” you say, but let’s not forget that early Makerbot stuff was actually pretty innovative. They came up with a great pastestruder before anyone else, and created the ABP to fabricate their own printed parts. Just because there’s a Makerbot logo on a machine doesn’t mean it’s automatically bad….
It just means it’s automatically patented
Yep, there are two patents for the Automated Build Platform, apparently one for the apparatus and one for the method. Call it what you will, but an inventor deserves to profit off their inventions. The ABP is a great idea and we’d prefer to see it out in the wild by now. There is, however, one problem with this train of thought: no one is profiting off the Makerbot Automated Build Plate. It’s not being sold by Makerbot. That leaves us with one other option.
We’ll make our own ABP. With blackjack.
Here’s a challenge for you, oh Hackaday reader. The world of 3D printers has pretty much settled on a single build plate. There is, if I’m reading the suggestions for this 3D Printering column correctly, a demand for an open source automated build plate. Let’s make this happen.
There are a few things to consider before attempting to build your own ABP. The most important of which is the flexible bed material. The Makerbot ABP uses – according to the patents – a mylar or kapton material in the range of three to seven thousandths of an inch thick. Next is a way to remove the finished part from the bed. This can be done with a conveyor or by dispensing a single sheet of material that is consumed with each build. All this is in the patents, it’s amazing that they thought of just about everything..
Yes, the Makerbot ABP is locked down under a few patents, but think of this as just another interesting engineering challenge. Just as Converse All Stars aren’t taxed as sneakers, the goal of an open source automated build plate is to create something that doesn’t infringe on the Makerbot patents. Design something, put it up on Thingiverse, and send it in to the Hackaday tips line. We’ll put it up.
79 thoughts on “3D Printering: A Call For An Open Source Automated Build Platform”
Cool idea for sure, but I’d much rather see a call for automatic bed leveling, because it’s going to require firmware, hardware, and host software collaboration. That’s the REAL pain point in reprap 3d printing. Anyone agree?
hm, I tried a long time ago with an infrared reflex coupler. The project failed, as dust and old tracks on the bed change the IR reflection properties a lot.
I implemented a NxN grid dynamic compensation for non-flat beds. There was a lot of discussions whats better, NxN or 3d rotational matrix. I am still of the opinion that both are useful.
Theres sort-of that.
I have heard of bed-messurement-and-compensation in software. Moving the head to compensate for the unevenness.
Hardly ideal, but was a mostly software solution so it worked with “normal” setups.
Johann has a solution on a Kossel Mini . I’ve got it working awesome on mine, accurate to within +/- less than the height of a sheet of paper.
Marlin now has it: https://github.com/ErikZalm/Marlin#instructions-for-configuring-bed-auto-leveling.
Doesn’t require any “host software” support other than sending a G code, which most software (e.g. Pronterface) allows you to do directly.
From what I could see, it looked like auto bed leveling only started out a few weeks ago. If it’s actually made it’s way into Marlin, then awesome. Can’t wait to see where it goes from here!
Printrbot solved this a long time ago, with an inductive probe.
Here’s an option in the works by New Valence Robotics.
Joel, automatic bed leveling is already available through the Marlin firmware. Search Youtube for “Kossel automated bed leveling”.
its realy easy for those ppl that are printing directly on glass without brim and skirt, the only thing that i would need is a whiper for the parts when the bed is cooled down
Awesome! I’m glad to see enthusiasm for the Automated Build Platform. As the ABP inventor I’m excited to see my baby out in the wild. I wouldn’t worry too much about patents. MakerBot’s position has always been to use patents only defensively. It’s the position given in Bre’s comments two years ago.
“If we decide to invest in patents as a defensive measure, we’ll have to figure out how to license them so we protect ourselves but don’t block innovation in the open hardware community.”  – Bre Pettis
Here’s some of my original posts on the the ABP. Turn down you speakers; the CupCakes were loud back in the day.
I’d like to clarify a point on patents. I made the ABP long before I worked at MakerBot and with no expectation of profiting from a patent. Patents stifle innovation and harm those who would otherwise benefit from products suppressed by the patent system.
That’s awesome that you don’t mind us copying your work! Thanks!
I’d still be nervous about anything comercial though. Even if the leadership of Makerbot has no interest in ABP and no interest in enforcing the patent that can change. They might change their minds or… different people altogether may come to control Makerbot.
Just look at Caldera and SCO! Anything might be different tomorrow.
This. Makerbot is now a subsidiary of Stratasys, and even if Bre has the same views on patents that he professed two years ago, he is not now in control of what happens with Makerbot’s patents, and they have shown some interest in getting a lot more aggressive than they have been previously.
Full disclosure: I worked there for 2 years, at the same time as Pax.
From what I remember hearing at the time, the reason they’re no longer selling it is because they never got it to work reliably and it’d probably be quite a lot of work to improve the design to the point it’d work. What’s more, now Makerbot have lost interest in creating a working automated build platform there’s no incentive for anyone else to put the work in because Makerbot’s patents mean they wouldn’t be allowed to sell it or even make one for personal use. That’s why people were so pissed about Makerbot doing it – they have no interest in actually designing and selling a decent automated build platform, but they’ve made sure that if anyone else does in the next 20 years only Makerbot will be legally allowed to make it.
Oh, and an additional note – if you’re planning on working around Makerbot’s patents, look up the doctrine of equivalents and then remember that patent trolls don’t have to have a winning case, they just have to cost you enough in lawyer’s fees that you have no option but to settle the lawsuit with them. Now consider the fact that Makerbot are part of Stratasys who have much deeper pockets than you and every reason to try and shut down competition.
Patents actually don’t restrict personal or innovation use – only commercial.
The whole idea of a patent is you get a license to be the only one to sell your invention for a time, in exchange for describing the invention in detail sufficient to reproduce it, for R&D and personal purposes at first, but after the expiration of the patent, as a new member of the public domain.
firstly… Yes, there are some people who are “innovating ideas” in and around rep-rap projects because they love the technology and support “ideals of open source hardware”,
but… a lot of the people that are developing the ideas are developing saleable products, whether they are starting a business, (like Adrian the founder who has a business selling rep-rap machines and parts) or whether they are only selling things on the side to help fund their hobby…
anything that can’t be sold is rarely designed or at least if not rarely designed then rarely adopted by the community., the whole movement tends to stay away from anything patent encumbered.
whether that is because the people are out to make money, or whether that is because people realise that not everyone has the time or money to make their own machine entirely. or maybe it’s because the traditional way that machines come into existence is you buy a set of printed parts, then use your printer to make and sell parts like the ones you bought, (which just can’t happen if they are patented). I don’t know. I suspect it’s a mixture.
go for semi-automation, by starting with a scara design 3d printer.
My used scara Adept 800 has an area equivalent to 8 or so of the rep-rap style printers. So you could print all night and the next day before you have to tend it again. That might be automation enough for most endeavors, even if not fully automated.
Continuous automation only works if you automate filament spool changing too.
I didn’t look at the patents super close, but what if you put the 3D printer on a circular track and keep the bed fixed and move the printer? each bed section could have a seperate heater and turn on a bit before the printer gets to it and urns off when the printer is finished.
And when the printer gets around to the starting spot on the track, it “accidentally” plows the existing piece off the track into a bin.
As to my idea – The science fiction writer George O Smith used the concept of a traveling matter duplicator to make a fence by moving the duplicator in one of the storeis in “The Compleate Venus Equlateral” [A great set of engineering problems in a tube era science fiction enviroment]
just a sidenote on the automated build platform:
was posted one year earlier than Makerbots patent.
So, whoever states that makerbot invented it, go stand in a corner…
Prior art doesn’t really matter. Can you afford the laywer time and other expenses to make use of that prior art?
Yes, it really very simple and obvious idea. To its introduction can stir only exclusively idiotic reasons.
Ok, Nevermind, this world anyway going to sh%@ty sh$t, and this is not surprisingly :) These are a lot of simple enougth solutions which working on another principals.
I wonder if we can get the EFF to invalidate the patent.
I think that using a flexible belt has been mathematically proven to be very, very difficult if not impossible. In order to resist curling, the tension that the belt must be under has to be ridiculously high.
Turning more positive. I can think of a few interesting ideas.
#1 set up a bed system that loads from one side and ejects out the other side when printed. You could laser cut sprockets into an acrylic bed and have some sort of automatic clamping system.
#2 create a machine that has integrated x,y,and Z so that the bed is fully static so that it could crawl over to a new piece of table and then start printing. This obv wouldn’t work very well with heated beds though.
I also +1000 the automatic leveling -or- software compensation route though. I’d much prefer that over the ability to crank out 3 dozen bottle openers. If the machine limited the ‘leveling’ layers to the first ~5 then most parts would still be fairly dimensionally accurate.
“I think that using a flexible belt has been mathematically proven to be very, very difficult if not impossible.”
Cmon. The film should be just thick enough to avoid deformations and provide enough flexibility to be rolled.
Rather than trying to use the flexible belt conveyer design that never really worked well, I’d go with a removable bed. Have a robot arm or something that picks up the old bed and puts a new one in. The bed with the printed part on it can be turned upside down and chilled, which will usually cause the parts to fall right off. Optionally, a metal scraper can be used to make sure the old bed is clear before using it again.
If you’re willing to wait for the bed to heat up from ‘cold’, you could use Peltier effect element(s) to heat one side and chill the other, and just flip the bed along one of it’s axis between jobs?
After posting, I realized that both of my ideas above will still rely on having an automatic leveling compensation system in place. I have seen attempts at making machinery that will adjust the actual bed height, but I haven’t seen any feasible results. The important consideration is to keep the swing weight down. Adding 3 servos + linkages + cabling to a bed will probably do more harm than good.
One interesting idea is that you could possibly use the printer to print shims on the top of a bed, then flip the bed with the shims still attached, and have it be perfectly level. That wouldn’t be automatic, but it’s still be cool!
Many printers do this already. for example: http://www.kickstarter.com/projects/blueeaglelabs/kossel-clear-lets-build-a-full-sized-delta-3d-prin
3D printers require a fair amount of manual intervention and babysitting. If part of the print fails to stick to the printbed correctly, it can result in hours of wasted time printing unrecognizable blobs. That’s why everyone sets up a print, watches the first couple layers, and then wanders off if the print seems to be going well. Still, there can be other problems like kinked filament, a pulley that works loose, an extruder gear that clogs with bits of filament, and so on. Getting a cheap 3D printer to run for 24 hours straight with no intervention is a pretty impressive feat. Adding some kind of part removal system will only add one additional failure point to the many that already exist to be solved.
glad someone said it.
you have to learn to make a 3d printer that can print flawlessly every time with manual intervention to remove the part before you can run.
I don’t mean to be negative but at the speed my printer works, it’s not a production machine, it’s a prototyping machine. I don’t need an automated build platform. What’s more the fact that the author has decided to go the ‘commercial’ route with patents convinces me he hasn’t really picked up on what this is all about. I do wish him luck in his quest for money though, these are hard times.
Four small and cheap 3D-printers will make things done 4 time quickly (if project divided to parts) than one big. If they will be automated it will make 3D-printers cluster almost production tool.
Printable automated build platform:
A small side departure: It seems that for a lot of these patents, many people in the community were experimenting with similar (identical?) contrivances at around the same time. Now, I have only a basic understanding of patents, and this is just a mild curiosity that I’ve been thinking about lately, but if someone can prove that they invented something first, can’t a patent transfer from the current holder to those who “originally” invented it? If this is true (and I am rather skeptical) some of the patents makerbot holds may not stay theirs… Anyone know more, or is this just wishful thinking?
I believe that the American law says where two competing companies independently design the same thing it is the first to file that wins…
the only way around this would be showing the prior art to be public domain, (like designs released under CC licenses on maker bots very own Thingyverse before they even filed the patent, for a machine that is remarkably similar. or the open designs on the reprap community…
10 min later…
Arrgh, if someone wants to challenge this, they have to, among other things, not have “abandoned it” all this time…
I don’t suppose someone continued to experiment with this stuff in their garages?
I guess things arn’t that simple.
http://www.youtube.com/watch?v=TRJRJJxBUIo Here is my solution. Super cheap to make, and has been working for five months. Please contact me for plans.
I like it, could be done with one. stepper if you have a reliable about of friction.
I used a Tamiya motor and a couple of sense switches…. I’ll sell you one for $30 if you want. I will give out the designs and code (it uses a Picaxe8 micro) out, email me.
I am currious, remember the pin boards from when you were young, if one were to roll a wheel across the bottom it would create a wave to push the print off the bed.I am not saying it is a good idea this would increase the weight of the bed and also make the bed not smooth, but hey just snow balling here
Conveyer belts, rollers? ugh sounds messy, problems with things sticking the belt, wipers etc etc.
If I was trying to solve this problem I’d just rig up a diverter to push the piece off the side of the table into a catch bin at the side with a chute shaped to slow the piece if its fragile. Its a commonly available part in pneumatic automation houses. If you dont have compressed air to hand, use a linear actuator to do the same job.
Control would be whatever could trigger a relay to either open a solenoid for pneumatic or energise the actuator for the linear actuator. A standard “G” code to actuate the table eject to be placed at the end of the cycle, and when deactivated springs or other mechanism retracts the pusher automatically.
Link to a pneumatic diverter on ebay if anyone doesn’t know what they look like.
Have you actually run a 3D printer? On my bot, parts are adhered so well that I can hit it with a hammer without dislodging the print, sometimes.
The conveyor was developed as a way to peel the part off the build platform – which is exactly what I have to do when the hammer method doesn’t work (on parts with a small footprint, the hammer tap will generally dislodge the entire part at once)
I don’t get it, whats the problem? A 3D-printer is a cartesian robot with a 3D-printing tool, its not limited to just 3D-printing. Why not just let the extruder push the part away? Build a gripper next to the extruder that lift the part away? Bolt the whole printer on sloped surface, heat the bead so much the bottom layer melts and it slides away?
There’s a couple problems with what you suggest.
On using a gripper to dislodge the part – parts in 3D printers come under a lot of forces while being printed, the grip to the print bed has to be stronger than that or you get a failed print (lifted corners, air printing, etc) – I don’t think the axes on my printer are strong enough to overcome that grip.
On heating a bed past the melt point of the plastic – potentially feasible, but most heated platforms I’ve seen struggle to get up to 100 degrees. You’d have to get around 230-250 to “slide away” the print. Then you’d be left with a bunch of thin melted plastic that then potentially has to be removed – unless you just give up on that one and start with a printed shim, that just gets replenished and reused over time.
A lot of forces in a 3D-printer? Not really..and if so, don’t build machines out of paper and expect them to be reliable and precise.
Have a heated bed coated with silicone rubber, teflon or any self-lubricated plastic so the parts don’t stick too much.
Yes, there are quite a lot of forces, for example the adhesion to the bed needs to exceed the force applied the the plastic material contracting as it shrinks.
as for removal from the bed, it’s not unknown for people with glass beds to find that a shard of glass is pulled from a sheet rendering the glass bed broken…
this idea that a little flipper from a pin ball machine or a quick tap with the print head, (which is powered by relatively tiny motors in an effort to reduce the moving weight is just fantasy.
So how are you supposed to remove the parts from the conveyor belt if they adhere so strong?
Simple, the conveyor gradually peels itself from the bottom.
Think Velcro, put on a suit and you can suspend your whole weight, then you peel yourself off gradually.
Yeah that would work and flipping that principle up side down; print a large radius on the bottom of the part and push the part with the extruder as a lever. Your argument would probably be that you shouldn’t print in air and I agree but many hobby 3D-printers can and does, because precision doesn’t seem to be the main goal.
I’m just trying to give a few ideas on how to do an ABP without any major modifications or rebuilds to existing printers. With some testing of different adhere strength between part and bed with surface materials I’m sure its possible to push the part of, without braking the machine. I like the saying ‘It’s not the tools that makes the craftsman’ and I think its a more beautiful and HAD’ish solution to solve problems with features built in the printed part or using the nature of the process. 3D-printers are one of the most powerful tools ever made. With some imagination and knowledge you don’t need any other tool.
Anyway you seem to have your mind set on the conveyor belt, have fun with it.
Most people at some point find themselves printing a brim on their parts, basically a large skirt to help the part stick.
Where the plastic shrinks where cooling it lifts. Deforming solid cubes etc at corners.
I doubt I’ll be making a build platform anytime soon. But if I did I’d go with a scraper system to wedge under and dislodge a part. Kind of like what I currently manually do with a paint scraper.
The problem with pushing the print off is the stress it puts on the gantry (unless you have a separate mechanism for it). It also puts a lot of stress on the part. How stuck a print is to a bed very from printer to printer, some are loose enough to push off, others (like mine) is really stuck on and attempting to push it off would ether destroy my printer or the parts i make. The other issue is what if the part is low and wide? you have to fight all that surface area.
The way I would go about designing a system would be how do you as the operator remove a part then implement that method mechanically. I have no luck pushing pulling or lifting the part off. I have to use a razer blade to push under the part to separate it. I think a design like Ian Lee, Sr. showed: http://makezine.com/2013/09/26/worlds-first-fully-automated-3d-printer-for-schools/ would be most effective. (October 23, 2013 at 10:21 am) However the issue you might run into, and i do with my razer is that sometimes the blade cuts or rips into the tape under the part and i would have to replace that. You would need to print on a solid material that is resistant to being damaged by a blade sweeping across it.
Sounds like a pusher furnace design.
This would also allow for some additional processing stages. A bed preheat zone. A cool down zone. Heck Pick and place.
I always thought that this was open source.
I mean after all it was discussed on the reprap forums well before the patents for the exact same idea was applied for.
and open source versions even appear on the thingiverse before the patents were granted. and before Makerbot put up their version.
kudos to the guy who made it work and developed the idea from a simple sketch.
I would think that anybody who desperately wanted to make one of these not only could, but should quite probably should… (yes even commercially)
then if and when the patent holders challenged in court the validity of the patent would be called into question.
(note I’m not saying steal their ideas, I’m saying develop your own apparatus, and challenge the patents that way)
as I said there is enough prior art even on Makerbots own thingyverse site.
Here’s a thought. From my experience parts are easier to remove from the build platform once things have cooled down. What if we used a thermal electric cooler (aka a TEC or Peltier Cooler) to actively chill the build platform after a print is done. By chill I mean take it down to close to freezing. The quick change in temperature and corresponding difference the coefficient of thermal expansion between the build platform and part may cause it to pop off. Then the extruder head can be used to push the part off. The TEC could even be used as a suppliment to help the build platform heat up faster by reversing the voltage applied to it.
That thermal expansion difference between glass and plastic may cause the glass to crack.
Then I guess this would be limited to a metal print bed or a glass that can handle the thermal cycling.
I have put the glass plate often into the freezer and it never cracked. But a metal plate is still better. Cooling could also be done by a big fan.
Is it possible to invalidate the patent, based on any prior non-makerbot stuff from forums and the rep rap wiki, withought actually challenging it? More of a petition or notice or something to say hey, these guys diden’t deserve this patent? or is that not how it works?
AFAIK, no you can’t just make a claim that someone doesn’t deserve a patent and expect that company to spend time and money defending said patent.
I think you can. The last time I visited the patent office the told me that is possible. But the major problem as you already said is that they will probably fight it.
What about a dual sided heat bed that automatically flips upside down when the print is finished… The part would presumably have time to cool and unstick then fall off upside down before the next part is finished…. No conveyor belt required…
jugging from the patent, the conveyer isn’t the important part. From what I can tell, this lays claim to any mechanism for removing prints, such that another can begin. The only loophole I can think of is something that doesn’t SPECIFICALY do that, like a robot arm that assembles printed parts and could *theoretically* be used to extract them.
(Also, really? I would have thought that there would be a way to refute and invalidate a patent on something widespread and “widely adopted”…)
I do huge amounts of automated ejection printing – leaving printers running 24/7 or over a weekend in the E3D industrial unit. All ABS printed fan-ducts for the E3Dv5 are printed on an automated system where the machine prints 4 ducts at once, detaches them from the bed, and sweeps them off the build surface down into a box where they are collected. The printer then starts the next print. I’ve printed kilo after kilo of these, without human interaction apart from changing the spool every 36 hours or so.
I’ve posted about it on the forums here:
And you can find a whole slew of videos about it here:
See spiritplumber’s comments above. Already done. :)
Our idea would be to print on a sloping conveyor belt. http://qremote.org/private/bilder/3d_printing.jpg So it would be possible to print endless objects. (That would also solve the problem of recreating the printer which uses parts that are bigger than the printing plate.) So the main use of this conveyor belt would be to print “endless” objects and the patent would not be infringed. However, side-effect would be that the objects could be removed automatically with the same principle as the ABP worked. This would need a new printer design.
When looking at the patent I saw that there are several other patents from Makerbot Industries like this one for networked 3d printing: https://www.google.com/patents/US20120113473
Aren’t these patents too general? In fact that would mean that nobody is able to sell anything related to 3D printing at all without doing patent infringement.
What the hell, I’ll post something. First thought is pretty simply everyone already uses the LM8UU bearings and 608Z bearings and 8 mm rods along with electric magnets. Use the rods with either type of bearings to create a run for the build platforms. Have three plus build platforms all metal with buildtak on top. The rods are exactly three build platforms long, first reason for build platform, second for new platform going in, and three for the old finished build platform going out. Since its exactly three build platforms long, Magnets release the first build platform once print job is finished, it gets pushed along by the new build platform (gravity going the work here) and when it stops the second build platform is now in position and magnets grab it and new print job starts and a new build platform must be placed in position and the old build platform removed. Endstops are engaged by old platform and alarms or noise sounded to let someone know to grab finished print and to ready the next build platform. Since no motors, no gears or new equipment needed I would think this would be easy to do. The problem here being that most beds are heated (wires) and locked in their positions but the other issue is leveling of bed but with new calibration devices not as much of issue as used to be. Just putting this out there, just a random thought.
I just found this blog post. My toughts about this:
1. The 3D-build plattform ist not a international patent and there is no patent for that in europe. (https://register.epo.org/smartSearch?searchMode=smart&query=makerbot)
2. In Europe you are not allowed to get a patend on anything, that has allready been published. This means, when you find a picture of a printer with a conveyor belt form before that time –> not a new thing –> no patent
Keep the build platform the same. Just put a conveyor over the platform and let the belt slide across. Only two rollers, one on either side.
Any patents battles can be avoided by making and selling the part through a Chinese business entity. Pay yourself a salary as a consultant. Or just keep the money in China.
Check out thing http://www.thingiverse.com/thing:872617! Just punch the last print out.
I am working on a system for automatic print removal with a scraper. Printing bed is fixed and it do not move. So this patent mean I can not sell it, but I can upload it to Thingiverse.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)