Non planar layers
-
very nice.
most printers will need a cooling fan adjustment for this to work on them
-
@veti
I don't think so.
The video shows the setup process where vertical clearance and nozzle angle are measured. I assume this is for the algorithm to decide which parts of the surface are reachable by the nozzle tip when doing non-planar moves without other parts of the hotend crashing into the print surface. -
yes i know, but look at the fan shroud of the mk3 for example. there is not much clearance between the nozzle and the fan shroud.
so not much room for manoeuvring -
This would be very cool, and It'd be relatively easy to create a longer thin "heated" nozzle perhaps out of a wire wound hypodermic needle. The heating would just maintain temperature from the heat block to the tip...
Either that or a 5 axis setup with a nozzle that could be sent alt azimuth demands... I could envisage 5 axis mill surfacing toolpaths could be wrapped around the exterior of a part!
Definitely one for the future...
-
This approach together with https://hackaday.com/2017/01/19/3d-printer-with-tilted-bed/ would allow to print everything without support material. It all depends on an intelligent slicer/kinematics and some additional actuators.
-
I have an adapter that allows an airbrush nozzle to be used as a 3D print nozzle. I've never tried it. I may dig it out, as it projects down several MM, and would improve the angles for this idea.
Hmmm... Where DID I last see that thing?
-
I don´t think this is very useful since the nozzle is not perpendicular to the printed layer.
There are a few problems.- This will result in concave lines on a downward printing path, and convex lines on an upward printing path, and therefore different surfaces on a downward/upward printing direction.
- uneven nozzle wear- this will wear out the nozzles one- or two sided, which will result in bad "normal" planar layers.
- this can´t be used with dual extruders that have both nozzles on the head. The second one would collide in most cases.
A true simultaneous 5 axis would be better. The mechanics is not the problem, it´s just that the software is lacking.
Some 5 axis printers:
https://www.youtube.com/watch?v=zhO2b0QnTao
https://www.youtube.com/watch?v=w8Fl8L4yk8M -
@nitrofreak said in Non planar layers:
The mechanics is not the problem, it´s just that the software is lacking.
specifically the slicer to generate the tool paths.
-
@t3p3tony said in Non planar layers:
@nitrofreak said in Non planar layers:
The mechanics is not the problem, it´s just that the software is lacking.
specifically the slicer to generate the tool paths.
Yes exactly, Sorry that is what i meant. For the Duet it would be just another axis, although for simultaneous 5 axis you need a pretty powerful processor that can calculate the Dynamic Work offsets realtime. As long as that is not a given, a 3+2 positioning is probably more likely. Also, as long as the duet can´t handle true sinusoidal G2/G3 computing, the simultaneous 5 axis is off the table.
I found a video of a 3+2 axis 3D printer, even running duet electronics:
https://www.youtube.com/watch?v=JdsR7nNDN3w
Anyways, slicers need to step it up to generate the toolpaths for a 5 axis printer.
First of all, they would need to implement slicing from .step files. This would allow a feature selection, especially for round holes. You could then also easily select planar surfaces to print features perpendicularly from.
There are CAM module suppliers that sell the toolpath generation algorithms to the CAM software suppliers. Just like moduleworks:https://www.moduleworks.com/de/cam/
I bet they have some 5 axis generation algorithms that could be adapted to 3d printing. However this is not likely to happen in the open source world, as these CAM modules are $$$.
When i was at the Formnext 2018, i was at the HAGE sondermaschinenbau booth and they had a 5 axis 3d printer. It was this one:
https://www.youtube.com/watch?v=qYE-WDh3NQM
They showed me their workflow and they used Siemens NX for the toolpath generation. However they could only use this for pipes and the sort. They partnered with siemens to adapt a "simple" 5 axis swarf to output the gcode for a 3d printer.
-
@nitrofreak Thank you for the video links, those are interesting samples what is possible. I think the sinusoidal discussion is only the "tip of the iceberg", the problem is, that the drive movement planning should be flexible for all drives, not only for the extruder. Maybe the current firmware extended by the possibility to define algorithms between the driver movements as dynamic parameters (e.g. G-Code at the beginning). E. g. if the heatbed is to be pivoted and is a hexpod, the movements will be nonlinear and complex.
Problems aside, I am sure an open software community can develop the same quality as commercial firms. There must be a driving group of people (or one person) however who are motivated enough to implement the solution. Maybe we can start with such an effort here with the enhanced Duet 3 capabilities.
If you look into the kernel of algorithms, they are in general very short, e.g. for Delta, for sinusoidal etc. in the different firmwares, so implementing is not a big effort. A correct model of behaviour is the most important point to start with.
-
I just found another example of non planar printing all the way back from 2014. It seems to do the same thing
-
I am thinking about whether Delta has advantages printing rapid changing Z. Cartesian, CoreXY, Prusa, Ultimaker are designed to print layer after layer, whereas Delta prints Z with 3 actuators, this may be an advantage in the case of non-planar prints. Especially because some of the Z layer printers have backlash when changing from move-up to move-down and vice versa.
-
If David implements a 6 axis delta firmware, it will definitly be my next 3D printer! For now, I'm working on my 2 axis tilt bed in order to print portions of sphere (https://forum.duet3d.com/topic/10934/5-axis-3d-printing).
-
@NitroFreak Thanks for the videos. I can add diabase to the list (https://www.diabasemachines.com/home) they use Duet to achieve 5 axis but in a different way.
More generally if the slicer was able to output X Y Z U V W, where U, V and W are the other three axis (e.g. a tilting/rotating bed) then there is no change required or special kinematics. If you want X Y Z A B C where A,B and C are the tool head rotations, then also no issues in RRF.
What do you mean by "sinusoidal G2/G3 computing", in this case? surely to implement G2/G3 in this case you would need to define an arc center in each axis that the arc would move on?
In general tool path generation is going to be an issue. Not just the path in 5 or 6 axis but also a path that avoids collisions with parts of the object already printed.
-
This is really cool!
Wonder if it wil make it to the home/hobbyist scene. -
Did anyone tried this non-planar-Slic3r edition? I can't get something out of it, just messy G-Code...
-
@dc42... Do you think this can be implemented in Duet?
These movement, requires especial slicer and firmware?
-
@peirof said in Non planar layers:
@dc42... Do you think this can be implemented in Duet?
These movement, requires especial slicer and firmware?
It only needs special slicer. No changes in firmware needed
-
And that's what is exciting about it, as vs. prior efforts: It is available in a "mainstream" slicer.
-
I have found another video/guide
