Originally shared by Nicholas Seward
I had this idea but couldn’t find a video of someone doing it. I then made a motion study to see if it would work and then promptly found 3 videos of real machines. I hope you guys/gals enjoy it anyways. I think this would be one of the most sane ways to approach a 6DOF printer.
How would it work in terms of FDM? What would be the benefits compared to a Delta or Cartesian printer? Although it looks cool, those additional DOF wouldn’t actually add anything to how FDM works. Why would “tilting” the print head be a benefit? FDM still can’t print without support beneath the filament and since it works layer by layer it never has to print “underneath” anything. Why would rotating the head be a benefit? Unless you’re also going to produce a non-circular nozzle, head rotation is irrelevant.
There are applications for this sort of movable head, but I don’t think 3D printing is one of them…
+Jon Gritton Aww. You have fallen into the trap of thinking gravity is important. Gravity forces are much much less than plastic forces. This means that if you can tilt the head 45 degrees then you can print 45+45=90 degree overhangs. But increased overhangs is only one use case.
- There is no rule that says you have to slice a model with planes. Nonplanar slicing has been done on regular printers but would be extra awesome on a 6D printer. https://plus.google.com/+NicholasSeward/posts/bcV99nWRE32
- Top surfaces that aren’t perfectly planar and aligned with the bed can now be perfectly smooth with zero stepping. https://plus.google.com/+NicholasSeward/posts/FFzY1EcFgkn
- Reduced support structures.
- Increased part rigidity.
- Print on curved surfaces.
- Add a gripper and do assembly after printing or part ejection.
The list goes on. The trick is you need a 5D slicing program. Luckily, I am slowly working on one.
@Jon_Gritton Side note: you are absolutely right that the head rotation is mostly useless. It is useful in that this 6D version is simpler than any 5D version I can think of.
@Nicholas_Seward hmm, not sure I fell into any trap, I don’t recall thinking about gravity at all (it’s far too early still it was more of an under caffeinated knee jerk response, sorry. Yes, I understand now, so effectively, if you printed at 90 degrees, it’d be virtually the same as if the nozzle was printing at 0 degrees because the plastic adhesion is greater than the downward pull.
That said, the slicing is going to be hideous, isn’t it (from a user perspective)? It can be hard enough getting the settings right for planar slicing, I can imagine needing lots of caffeine to correctly set up non-planar, curved surfaces and so on.
@Jon_Gritton Yep. The software is the linch pin here. Think of all the settings Slic3r has vs Cura. They both do about the same job but Cura has a fraction of the settings. The approach I have in mind should be just about as automated as Cura is.
I think what you are really saying is that I am nuts for thinking about this. Can’t argue there.
@Nicholas_Seward not at all. Oh, yes, it’s completely crazy, but it’s the sort of crazy that we might all be printing with in a few years time
5-axis CNC is a REALLY big deal, and has a lot of super important uses like generating complex contours and optimizing cutter performance by changing the rotation axis relative to the workpiece. 5-axis 3d printing is probably less meaningful, but it should offer some nice benefits in surface finish and strength/orientation optimization.
I personally question whether the “print any geometry” benefit (eg 90 degree overhangs) is really worth the added mechanical and toolchain complexity compared to using one of the various soluble support solutions.
I have seen a similar setup in 5 axis cnc a few years back, It worked ok. 5 axis cnc is very complex work. The cad-cam packages to work in 5 axis are extremely expensive. Good luck trying to program something useful! I do believe nonplaner slicing will make better parts.
I agree 100% that the lynch pin is in the software. I designed a 6 axis delta printer in a DIY friendly construction previous to your sextupteron work. See pictures here:
The prototype currently functions as a large impractical paper weight. I encountered two major problems:
- deriving reverse kinematic equations.
- slicing software that takes advantage of the additional degrees of freedom.
If you can solve 1 and 2, I would be thrilled to upload and open source the design.
@Adam_Thorp Sweet!! There are some really slick design choices. I really like the x carriages.
Yeah. I really want to build it but that is the easy part. I keep delaying a 6D build because it is mostly pointless without the software. The inverse and forward kinematics are well understood in literature so mark off item 1. At my current rate I should have a 5D slicer in 2 years which means probably never.
I LOVE hexapods, but my practical side says putting two rotary stages on a Cartesian printer is going to be a lot more practical. Column hexapods are quite limited in achievable tilt angle with good stability throughout the useful working volume. Maybe 20 degrees? You have to be careful to avoid entering singularities and inadvertently changing assembly mode at high tilt, which is an additional layer of software complexity if you don’t just set a conservative flat angle limit. Whereas serial rotary stages are really only limited by collisions.
@Ryan_Carlyle . Yeah. I was worried about that. I was going to look at what was allowable without crossing singularities. My gut says not much like you said above.
I think I would pick a cubic Sextupteron over a serial. It is very well behaved.