3D Printing with a Robot Arm - Computerphile
ฝัง
- เผยแพร่เมื่อ 20 พ.ค. 2024
- With dozens of affordable 3D printers out there, why re-invent the (filament) wheel? Feng Zhou has a plan, to improve the strength of 3D printed objects.
/ computerphile
/ computer_phile
This video was filmed and edited by Sean Riley.
Computer Science at the University of Nottingham: bit.ly/nottscomputer
Computerphile is a sister project to Brady Haran's Numberphile. More at www.bradyharan.com
![🔴 [TRỰC TIẾP] U20 VIỆT NAM VS U20 THÁI LAN | Cúp VTV9 - Bình Điền 2024 | JET STUDIO](http://i.ytimg.com/vi/tdmZlw-XJJI/mqdefault.jpg)
Can we please have another, more in depth video about the mathematics behind it? I'm curious to know how you achieve cross hatching on subsequent double curved surfaces.
Looking forward to see your work appear in cura!
There's quite some researchers looking into non-planar slicing. Rather than following some predetermined wave pattern the real challenge is to determine the layer geometry based on the particular of the 3D model. There's some people in my lab working on that, but I know of quite some labs working with robot arms and even companies which are (planning on) bringing this type of stuff to market. Check out Chengkai Dai, Jimmy Etienne, Vshaper.
Sadly the chances of 6-axis non-planar slicing becoming a part of Cura are approximately zero. CuraEngine is simply not made for that atm.
@@angelorf I don't know why youtube didn't inform me of your comment, but thanks for the info!
I thought of this video again as the guy from the Roetz 4.0 youtube channel just bought a 2nd hand KUKA, and asked for ideas on what to do with it.
I wish we'd been able to see a close-up of the printer printing for a lot longer.
In the short snipped we saw in the video there were a lot of vibrations in the printhead. The robotic arm seems mechanically not stiff enough to achieve positional accuracy for speeds that would be interesting to look at. My guess is, that his project is more focussed on the software side of the problem, but the mechanical side is a little bit underdeveloped at the moment. Maybe he'll do a follow up video some time I the future in which the arm is more rigid and more tightly controlled.
@@Backbeardjack99 i suppose the control system accounts for mechanical drift. you're right in that a 6 axis arm like that compounds errors by 6 if the thing was naive. I'm just not sold on the necessity of the project, unless he actually demonstrated improved strength from having the print head essentially pivot to the object plane while extruding.
@@sixdigit the whole concept is in itself a very good idea. Because with composite fiber materials the direction of the printlines matters a lot. Carbon fiver for example only exhibits it's enormous strength if loaded lengthwise. If you could print a part with different print line orientations in specific areas of calculated stress concentration, you would increase the strength by a lot, compared to a conventionally layered 3d print.
The problem with the current robot arm is not one of drift or backlash, it is one of mechanical rigidity. Thats not something you can just compensate for in code. At least not if you rule out going really really slow. Because the problem is, that when the arm moves it has to accelerate its own mass to the desired speed. This creates a force and subsequently a torque at the joints of the robot arm. If those joints are either of high tolerance or it's material is simply not rigid enough this creates deflection. One could try to compensate in software by doing very elaborate calculations of torque prediction with precise analysis of the robot joints, but that's not trivial. The best choice would be to use a bigger robot arm designed for greater loads and with tighter tolerances.
@@sixdigit
It could achieve alot of thing. Like more printing arms to accelerate the process.
Planar frame could only add so much printing head in tight space.
You should display Zhou's name more prominently in the video, It took me a while to find it hiding in the video description. It's a really amazing project!
One of those ideas that after seeing it, you wonder why it wasn't invented years ago. Pretty cool!
I believe it has. Namely, I remember seeing a video using an industrial 6dof robotic arm with a welder at the end being used to 3D print in metal. But having more actuators makes the 3D printers more expensive. And in order for 3D printing to become more mainstream I think can only happen when it becomes more affordable.
It was, the problem is that eventhough the concept looks simple the implementation is anything but. To get to the precision of mid-range 3d printer you would need a veeeeery precise robot. Same thing with software. Sure thing, you can create something that shows offf your machine can print non planary but it's not the same as making a fully capable 6-axis slicer which would slice any input model in any desired way.
I've thought of it years ago, but just assumed it had difficulties that would be solved with time.
Also I'm not a mechanical engineer so I couldn't solve it anyway.
Multi Axis 3D printing is as old as an idea as multi axis CNC machining is.
The main problem is simply the programming task behind it.
"Slicing", which converts models into those paths is the main reason there is nothing out there yet outside of those small tests like you even see here after 3 years.
Planar slicing is already hard enough to program and get right, adding dimensions just adds even more complexity.
difficult to implement and much more pricy than simple fdm printers
Loving all the roboticists being featured on your channel! You should cover inverse kinematics using Matlab or something free on Numberphile!
ugh MATLAB
Very cool! More from him please
It would be interesting to compare the speed and accuracy of this 6-axis versus a normal planar cartesian 3D printer. Those additional joints probably also bring more backlash and less stability with them.
MORE!!!! This guy is a genius.
You can do nonplanar 3D printing with a regular 3D printer, if you have the capabilities to produce a gcode that does. You have to be aware of the size of your print head though, so there is no collision. This six axis arm helps with that
This is an awesome child of the ideas of consumer 3d printing and a 6-axis CNC. Super cool!
"Can we see it do some printing?"
And we don't get to see it.
Very impressive ! Congratulations !
This is very interesting topic.
I immediately thought of this very new Startup in Germany in my area that is working on this at the moment.
It is called FreeD printing.
I tried posting this twice before, but the process failed. Please excuse any duplicates.
Based on the company name, I'd have thought it would be Bri' ish
This is an amazing idea, it's incredible it hasn't been tried commercially yet
Probably in part because you need an extremely precise robotic arm to make it worth while, and a highly precise robotic arm is very expensive relative to the benefit. Standard 3D printers are designed in such a way that precision is pretty well optimized for the price, and the majority of people using 3D printers either don't need the extra strength printing at different angles could provide, are willing to make the sacrifice of printing with supports to achieve a certain angle, or are willing to use a different method of producing the product and are only using 3D printing to rapidly prototype.
The software would also have to be more complicated, particularly if you're trying to optimize to specific types of stress, and the R&D necessary might seem prohibitively expensive for most.
This is incredibly impressive, wow
You can also add multi-axis positioning to the printing bed, getting 6 axes above and below, permitting printing in ways that can appear to defy gravity.
BTW, the PLA may be full of moisture: What was coming out the nozzle looked a bit bubbly.
extremely cool, would love to learn more about the challenges of writing this software
6DOF industrial robot arm 3d printing is super fun! allows for a huge workspace and easily printable furniture! You should look into Pellet extruders as the toolhead, it makes for an infinitely easier setup and much larger potential print size.
This is really cool. I'd really love to see it paired with a FEA simulation to find the optimal printing path, so that the strongest part for any set of stresses is produced. How does the precision compare to something like a 7 axis CNC? Traditional setups are definitely stiffer.
That talented guy just saved me three years. How thoughtful!
That's cool. It would have been good to get more on how the program paths a 3d model but maybe that's for part 2. Pls?
Awesome topic.
Would love to see a follow-up on the design of the nozzle and discussion of non-planar slicing techniques.
Amazing!!!
Super cool!
im interested in the coat rack used as a spool holder ... i have a 3d printer and im struggling to find a spool holder !!! i will look no further !!!
very interesting project ... but very expensive
I feel like I’ve seen the father of the next evolution of 3D printing. Such a major leap forward and should be a game changer for sure.
I did this exact 5 years ago in 4 weeks. These arms specifically are a pain to work with, but the ability to work along curves and screw/unscrew heads is really exciting
When I see technologies like this, I imagine taking this process to the microscopic scale and after mass repetition and optimization we will see extreme other technologies.
Computerphile
Any link to find project online?
just when I was getting the hand of conventional FDM....
fdm printers can do some limited nonplaner printing
I would like to have seen a finished part of this! D=
What was used for the path generation? Could you please give more information about it?
Hi! Very nice! How is this project going?
Amazing!
Very cool. I wonder if he has tested multiple arms for printing the same thing faster.
I wanted to see it finishing the piece! We never got to see the curvature this technology allows for 3d printing!
Conventional 3D printer can do more complex tool paths instead of printing just horizontal layers. The software that converts 3D geometry into G code is a simple slicer but it doesn't need to be.
Yep. I don't think I saw anything that requires more than 3 axis in this video. "Non planar" printing has been a thing for awhile now.
Similar effects are achievable with CoreXY printers with 3 axle print bed. Also look up MX3D. Interesting project none the less :)
I was thinking developing a 3d printer from a robotic arm, until I saw this. Someone is ahead of me
This type of 3D Printing is amazing! lookup MX3D here on youtube they basically have a 3D welding robot arm.
Hello,
I am currently working on a project that involves 3D printing using a KUKA KR3 R540 robot, a REVO HEMERA XS extruder, and a ramps 1.4 control board. However, I am facing synchronization issues between the robot and the extruder, and I'm not proficient in programming.
I am seeking assistance on how to synchronize the movements of the robot and the extruder in real-time. Can anyone provide guidance or sample code to help me achieve this synchronization? Any relevant information or resources would be greatly appreciated.
Thank you.
While 3D printing (planar layers) can do a lot of things, it's plenty of inefficient at times. There is also stuff you can't print.
Going 6 axis will allow you to print more stuff and get better performance. But I doubt it will make it to hobbyist anytime soon.
The spool holder and no part cooling fan is a little odd to see.
But what can I say - my 3D printer isn't operational for two years.
arm-type printers are a really cool option for larger scale printing. just slap that arm on a larger lower-precision carriage. or if ur going for arbitrarily large areas u can slap that on a rolling, climbing, or spacecraft drone & ud have a way easier time than trying to adap traditional FDM. give that thing a quick swap mill head & a wire welder n uv got an all porpose construction drone bruh
This makes me wonder what other possible ways there are to print
Might be a dumb question but what are the other 3 axes besides moving in the x, y, and z directions?
rotational axes maybe?
Brillant 6 degrees of freedom
please tell me you are running g-code file for printing and if you have any helping material about this project mean (code) pleaseeeeeeeeeeeeee share with us.
Cheers
I love you, cold unfeeling robot arm!
I love how he calls peek normal matetial :) i have 2 such robots i have an extruder... but i lack the brains and the time :) very well done!
Idk if it can get better than LCD
Nice
It does look like it's printing a little shaky.
When this is a finished product, it will be expensive.
Not for us, simple people.
Well, actually, maybe use 3 axis arm and 3 axis print surface. You don't have to rotate the arm if the bed does, no?
I started watching this and the first thing that came to mind is this needs to be developed for using carbon fiber materials or incorporating carbon fiber in to 3D print and at 3:23 Feng Zhou said he was looking in to this. The guy will be on to a winner if carbon fiber materials can be included in to the process.
Look up "Continuous composites." They very much are doing what you are talking about.
@@SirSpence99 Thank you for the heads-up, I will check them out.
Did he make that 3d printing software as well or is someone aware of the software he is using?
He created the software himself using Rhino3D (nurbs and more geometry modelling software and framework) and Grasshopper (a node programming tool for rhino). Have fun digging in, it's a great Program!
@@peterkrattenmacher9075 I couldn't have imagined a more in depth and thoughtful reply to my question. Thank you very much. Have a blessed day today.
As seen in the title sequence of "Westworld".
Go Feng!
hasnt this already been done with traditional 3d printers by using modified slicing software?
Its crazy what kinds of things u can achieve with an arduino
That would be handy for welding the body sections of Space-X Starships.
I think they are already using robotic welders to weld the stainless steel rings
Science fiction made real; this is the future!
I'd love to see this implemented using a soft-bodied robot. Like a spider-snake. A little bit scary!
CS2 called for a robotic arm to print inside any kind of aircraft structure... and no one dare. Search HLFC on Wing,
cant believe they did subtitles to my boy
i'm glad they did; they were faithful to the intent of his words and it can be fairly hard to understand if you're not accustomed to this accent
Uh. In fact it's not that special. E3D heat element. .4mm like all other standard printers, and there are slicers mods for doing curved slicing already.
no results were shown tho ...
It was printing and you could clearly see that it was severely underperforming a conventional 3D-printer. I bet it has something to do with the sensor resolution.
"But Harry, maybe the arm had bad PID parameters", maybe it did, but if he showcased it then I guess he set everything to be as good as it possible could be. Surely he would optimize them before the video, right?
for some reason any nonplanar printing video is super short and doesn't explain the non planar research
This kind of technology could be key to organ printing. I don't care about making widgets, i do care about being able to make replacement cloned organs for organs that have either failed or are about to fail.
Love that he got a freakin RGB fan 😂
🥰🥰
This will change the game for carbon fiber reinforced polymers!
An actual example would have been nice
Hopefully I can print out a 3d printed body and xfer my mind over before covid takes me out.
Surely I cannot be alone in thinking that these people have no clue whatsoever about the production techniques that have been in used since before they where born. Robots have been welding reinforcements to carbodies for decades.
And as for using a robotarm to print... my dude, in Amsterdam's they used a welding robot to "3d print" a metal bridge across a canal. Builders use printers that print entire walls in concrete.
Ijection molding machines use strengthening inserts, some 3d printerts can print multiple filaments at once...
I know, this guy isn't stupid, but this is re-inventing the re-invention of a re-invention.
3-D robot arm printer? Cool.
3-D robot arm printer running on Arduino? Even cooler.
3-D robot arm printer with three possible heads connected to 3 different forms of ink? Wait, don't tell me it's going to complain about being out of ink if any of the heads are out of ink...
I'm glad there was subtitles, I didn't understand when he said "aluminum"... * Reads subtitles *... Oh, aluminium! 🤪
Would have been better if we could see one of his completed prints.
The idea isn't new. I was watching a video from 4 years ago that talked about this, and it had a demonstration of what it could do that a traditional printer could not.
This guy only showed the computer mock up, and not the completed build.
There is literally a company selling these machines for industrial use in my hometown. Hardly a new invention.
5:50 that start of print looked horrible ngl
Cool project poorly presented.
No finished result shown, no head change, no filament change, no strength comparison vs flat printed objects, only hardcoded subtitles :(
Kojima????
I'm not native English speaker, I wasn't looking at the video and when I did, I noticed subtitles.
They are really not necessary.
Maybe there are viewers that are worse at english than you are. Or maybe they are deaf. Having subtitles never hurt.
@@harrysvensson2610 It's possible, but as the deaf thing, TH-cam does it automatically :P
@@IIARROWS TH-cam does it automatically? I can't activate any subtitles on this video, not even computer generated ones.
@@harrysvensson2610 I don't know, maybe you must enable them for your video, it makes sense.
thanks for the subtitles, otherwise it would be nearly impossible to understand
Yes you can clearly see how they are not native speakers
I wonder how people would react to my german accent
Why is there a "- Computerphile" in every title? Redundant and useless. Add the presenters name instead.
Can someone explain me the advantages? I can't see them
The print can have stronger structure, because it doesnt just stack layers
I think the benefits are two-fold.
First of all, the more manoeuvrable robot arm stays orthogonal to the surface being printed, which means that it can lay down the extruded fibres more evenly and the result is stronger.
Second (and I'm guessing here because I don't have a 3D printer), because this robot arm already needs to be able to locate itself in space better than the less manoeuvrable normal printhead, it's easier to switch between different heads and have them line up properly with the print in progress.
It also seems that he is able to print with different materials actually during the print process, so blending them together: is that an improvement over the usual process?
I would have liked to see a finished print showing the advantages of this method. Great idea!
You can print overhangs without support. Planar limits you to a plane, meaning any overhang (over ~50°) gets worse in quality and impossible at 90°.
Non-planar printing doesn't has actual overhangs as the nozzle can always be "planar" to its current path.
Another benefit is a better look on the print. If you print round top planar you have the steps on the "round part", with non-planar you can make smooth lines that follow the roundness creating a smoother surface.
More like 3D-painter...