I've always been interested in multi-axis printing but there is like no software support. Their approach to slicing is super interesting and makes a ton of sense!
I want to know more about the potential strength increases. This could be really big. I also want to see what happens when you try to build more complex parts.
As you said, mind blowing. I mostly use my printfarm to print nosecones for the model rocket kits I sell. Due to the method of layer adheasion they have a weakness for dropping hard surfaces. This method could change the way the layers form this making them stronger without needing more infill or walls.
Anything that helps in reducing the mechanical anisotropy of printed parts is welcome, that actually interests me more than the overhangs not needing supports because for that we have many solutions (good profile and printer tuning for easily removable supports, dissoluble supports) whereas we don't have a lot of ways to drastically increase inter-layer adhesion. Having used a Blackbelt printer in the past, just having a different strength distribution was already quite interesting, this new concept must be even more exciting ! What about doing some mechanical tests on these conical printing samples :D Also, kinda cool to see some open source 3D printing from my own country :p
Actually, the hardware was never the bottleneck, as a cnc post processor programmer, we know as a matter of fact that you can add more axes on the go but trajectory calculations for that axis is a much harder task. And it gets harder as we add non linear axes. But hopefully more developer's will join on the train.
Seeing techniques like this shows that we're still at the dawn of FDM printing, and I'm certain that we're going to see great advances in coming years to improve speed, quality, and function!
To me, it's like we're just getting into the Model T era of printing. We're going to places that we only dreamed of even a few years ago at this price point.
@@BigBlack81 It's like looking back at CD-R's when the first drives were hundreds of dollars, and wrote to $50 blanks at 150kb/s. Now we stuff GBs worth of data to SSDs in mere moments, and it costs next to nothing to do so!
We are not at the dawn of FDM. We are simply at the point where 2.5D printing is finally getting the 3 and up axis it should have had a decade ago! LinuxCNC is open source and has had multi-axis capabilities for well over a decade(EMC was from the 90's). Why has it taken this long for additive technology to catch up? Clearly it is all about managing angles with flow, that is rather simple volumetric math. You remember that robotic FDM (6 axis + rotating table!) FDM That was YEARS ago. Where is it now? It made non-planar prints with ease. Money. That is what the real hold-up is. People want to earn a living. I don't blame them, giving thousands of lines of code away to everyone for little to no return is not very appealing in these days. Not to mention all the complaints and bug reports ;-) Multi-axis printing has been around but has only slowly been trickled to the printing world we all reside in. Maybe it also has something to do with patents.
@@leofortey7561 yes it is all old 80s tech. people forget it was patents that kept it from being widely available, not tech. just wait for more newer patents to shut this down too...
i´m a machinist programming 9 axis mill turn on a daily basis and i can´t wait for 3d printing getting rotational axis, there is so much potential in 5 axis 3d printing. No supports, parts with reinforcing layers that are going three dimensionally over the part, that help distributing forces, reinforcing the weakspots and eliminating layeradheasion as a limiting factor etc. ...
my thought was similiar "if we add another axis to be able to rotate the nozzle, we could smoothly switch between different print methods" but the approaches all have the same issue we need a slicer code able to execute it i think the mechanical challenge will always be the smallest part it will always be about the code that causes the slow development
I am no engineer, but I'm worried more axis might be a little two much stress and complication for the filament that we use today. And what if the new printers need different material and non of the producers catch up. Burring new printers from the beginning? Like others said we are still on the dawn of additive manufacturing. Nothing is set. I just don't trust the big industries. I will put my trust in us, the creators because this is really exciting
@@mad-0614 I can see this being a problem with some materials, with abs I don’t think that it would be much of an issue but I have had pla‘s that we’re pretty brittle wich definitely could be a problem
An IDEX printer which combines both technologies would be absolutely killer. One printhead could completely cover the weaknesses of the other, and it still eliminates significant waste because IDEX doesnt need a large purge tower.
The idea of pre-distorting the object before slicing it and then un-distorting the code afterwards to get the desired results from an existing slicer without writing a whole new slicer to do the task is genius. I did think that the extra axis would be to control head tilting, that way it can still behave like a regular FDM 3D printer, but the tilt ability would then allow the printer to do overhangs etc. without support.
This is what one of the rocket manufacturer does with its 3d welding machine to print the whole tank. Its slices and adds some pre-deformed calculations and so when it prints it is now a proper print. This is done because welding deforms the metal so without this technique it would be a very distorted tube.
Pre deforming the part and then running it through a regular slicer is freaking brilliant. I can't believe I never thought of that. There is so much you could do with that. Heck screw cones, use a parametric curve. No more shear planes that way...
If you can get the RotBot team to print you some testing dog bones and some test hooks I'd love to see a strength comparison of this conical slicing compared to planar slicing!
I would reall love to see this 6-axis printer, with tilting base. This should make printing overhangs with no support even better, but of course it is more complex and costly, so this modified prusa with tilted nozzle is an interesting cost-effective option.
@@greatestevar Given that WE would not be programming it, your claim applies to the select few who will do so, the rest of use would just be users of the software….but even then, not so much as 6 Axis machines have been around for awhile. So it’s not like the human race hasn’t had any experience with it. Also, there was one int he video. So it has already been done.
Wow, this looks huge! Simple, yet effective. Making it open source will make it possible for many people to try out this new technique and improve the design further. Maybe a design with an additional axis to tilt the printhead between -45 and 45 degrees could take care of the inwards facing overhangs.
I’m glad you finally got to go and see this. I was hoping you would since the it first hit the 3D printing news sites. Very interesting indeed. The conical slicing is a clever trick. With some effort, I could see it being somewhat more generalized with a tool that segments a more complex part in to blocks each having their own pre-warp and post slicing de-warp.
Wow that spinning head made my head spin, had to watch this twice & finally understood the basics about conical slicing, thanks for explaining that Stephen.
There is an chinese approach which does the following: 1. Define the object in a voxel grid. (Small box of size dxdxd, where d is the diameter of the nozzle) 2. Calculate the reverse print aka how a CNC machine would destroy our object in layers, one voxel at most per layer. 3. Calculate the moving convex hull of our object, while simulating the print. 4. Calculate reachability of voxels, while simulating the print. 5. Use 2/3/4 to define an order of voxels to print, while printing on existing voxels. 6. When done, combine the voxels of step 5 to groups of voxels. 7. Calculate the actual geometric layers, from step 6. => The result are 3D complex layers that can only be printed on a 4+ Axis system.
I think it is "next evolution of 3D printing", as this will help speed up CAD/CAM effort, material waste, more automation. Yes, will be slow and hard at first but it is an investment in the right place. I have been trying to do the same and seeing this is very encouraging , thank you for awesome content!
I’d almost be impressed if I didnt just see a 5 axis machine on a Prusa i3 earlier today. ERRF attendees talked about you a lot today Stefan! Hope we can see you at a future ERRF!
I've been looking forward to non-planar layer deposition for a long time, but mainly to improve top layer surface quality. Extending it through the part looks likely to significantly improve part strength due to the effect you mention.
This is a fantastic idea and what would make it even more versatile is to have a head that can tip to any angle instead of just a predefined 45 degree angle. I could see taking a dual filament machine, setting up one head on the 45 and the other on the standard and then creating the code so that the best head is used for each independent layer.
There is a 5-axis version of this somewhere around (I think even Rene Mueller built one) but this introduces even more complexity into slicing and flexibility into the tool head.
An approach to fixing the inward overhang would be morphing the stl non uniformly, basically instead of a 45° cone, you could make it so that along the y axis you are at 45° inward, and along the x 45° outward, making a saddle shape or even slicing inwards to 1 inch off of the center, and then outward, it's fairly free, you're already converting cartesian to polar and then scaling up or down, so you just have to apply a curve that takes into account the position along the polar coordinates and applies a pair of curves instead of a set angle from just the distance. Automatic overhang detection exists, so mapping overhang position vs the angle and distance should be doable, and automatic curves may be created.
You're exactly right. Came to the same conclusion as I watched. Some of the edge cases could be trickier to fix - detecting and modifying travel moves that would otherwise result in a crash, for example - but not impossible I'm sure.
@@bobthecannibal1 Good point. Sounds like an interesting problem though. No need to complete one "layer" before starting the next, necessarily, so perhaps there's an ordering problem buried in here.
This is essential for FDM to move forward. We have stalled on 2D printing + Z (a more accurate description) and weak layer adhesion. Would be interesting to see strength differences and see if we can get true anisotropy. The next step would be a rotating nozzle from 45-90 degrees.
I like the aproach better where you have a normal CoreXY printer with independent tripple-z axis (like many use for planar printing) because you utilize gravity in order to print overhangs. New ideas like this one here help us to get better 3d printers and slicers for makers. Thank you for showing us, Stefan :) Greetings from Germany ✌️
I want this on my printer yesterday. I hope Prusa3D is taking a good look at this and trying to implement it, because I would buy that printer instantly. I do mostly mechanical parts and not having to worry about overhang limitations is an absolute gamechanger.
I'm really excited for mainstream non-planar printing and slicing. I hate wasting material because I just never throw the stuff away. And from an engineering standpoint being able to compensate for the (material) non-linearity of 2.5D printing with non-planar printing is a huge improvement.
@@CNCKitchen You where. I was making fun of the fact that I am still so uneducated in many things. Watching your channel teaches me many things. So thank you for your hard work. Baie dankie, ek sien uit na die volgende video.
Hi Stefan this is amazing and I'll have a look for details for sure, I'm thinking to update my diy Prusa based 3d printer, as soon I have a time I'll start to implement it, thanks for sharing!!! Greetings from Cham/ BY
As someone who managed applications for a 5 axis 3d printer building company, this is huge! We've mostly used a modified version of Mastercam, which doesn't really make itself available for makers, let alone the custom plug in.
Agree, over hangs is a big limitation and conical printing while adds a layer of complexity is an ingenious step forward. Super interesting and curious to see where this is going to go.
Stephan never ceases to amaze with his ideas to cover the unique and fringe applications of 3D printing and the excellent quality of his videos, congratulations and thanks. This technique would be perfect on my TronXY X5SA-400 or ANY large format printer were you can imaging printing a section of a car bumper skin that no mater how you angle it in slicer still has considerable overhangs, print time, and waste plastic, With this printing technique, there would be essentially no supports (or very minimal) and print time would be dramatically reduced, and waste almost eliminated.
This seems really similar to your non-planar 3d printing video you made a little while ago! It's cool that so many tricks can be pulled off by modifying what the "plane" is in planar slicing, and it feels like we're just scratching the surface of what's possible.
Oh damn, I was imagining a system to no longer need supports!!! I thought of a 45° swivel head, but in my idea, the plate should also tilt. I see we're never the only ones thinking the same thing..
Good afternoon all, This is a phenomenal breakthrough, I have been looking in to all the files including the script files in Python, excellent work, nevertheless one thing is missing is all that, the marlin configuration and modification files, also how and where connect the Z-Rotation motor to, it would be nice to have some complete concise Opensource compilation of all this so that amateurs like me can make one and be part of the testing bodies. Thanks guys
The thing with development of slicing software for those machines is that you need to somehow bring together people with the necessary Backgrounds, that are willing to pour a ton of time into this project. I mean, you'd need people, who know exactly what this printer can do and how it does it mechanically, as well as people, who know how to write a slicing algorithm and someone, who knows the math to pull it all together. I'm really glad they open sourced their concept, so that everyone interested can put their ideas in it. I think this is the beginning of something awesome
This looks so cool! Of course when I see an extra stepper on the hotend, *my* first thought is “ooh! Move the second stepper to the other end of the X axis! CoreXθ!”
one thing touched on in the video I think is going to make a huge difference on current gen-fdm printers is a more complex slicing that can intelligently section off geometry that can be printed either in a specific order, or with a different slicing technique to improve performance and speed. It seems that the state of the art is a naive, uniform approach, but it can be significantly better. And an entirely new piece of hardware wouldn't even be necessary.
Or have the hotend only tilt to different angles and have a build plate that rotates. Then you could use a bigger stepper to accelerate the build plate rotationally more quickly, not have a cantilevered hotend, and less mass on the hotend as well. Maybe even a simple hotend swap with a hotend at a predefined 45 degree for simplicity/ weight reduction
would be intresting to try it on printer with kinematic beds, who would be able to mitigrate the digging of the nozzle into the already printed lines a bit
Sounds like the industry is going forward. Might be great as a separate tool head in the Prusa XL. There is also the the opportunity to and an additional axis for all angles for the extruders.
While Stefan is visiting universities, I got a vote for Austria's University of Applied Sciences Technikum Wien. Their robotics and 3d print farm automations are amazing
I could be wrong, but I'm pretty sure the "sample overhanging part" shown at 6:28 is the z-axis bearing holder from a Hypercube Evolution. Kinda cool to see a part I've printed for one of the first printers I ever made on a big channel like this.
Das Beste an der Metote zu drucken ist, das man keine Stützstruktur mehr benötigt! Deswegen fände ich es cool wenn diese Technik bald für Privatanwender verfügbar wäre! Zum Beispiel wäre es doch schön, einen Prusa Mk4 mit der 4 Achsentechnik kaufen zu können! Dieses Thema finde ich sehr interessant und werde es weiter verfolgen! :)
Having access to conical slicing would be really interesting for certain applications, I think especially so if you want to make a print watertight, so having it on most printers would be really cool
It still needs supports in some cases like bridge like things. Solution is both rotating table and hotend together to get rid of supports at all. I'm sure that kind of machines exist in 3D printing world but it should be far expensive than the ones we have in our homes..
I see nobody other than students and researchers using this "upgrade" as a means to study slicing profiles and developing printers capable of 5 DOF printing heads.