Hey guys, anyone knows of an 4-axis-DIY-3D-printer project?
This one is the only one I’m aware of: http://www.dezeen.com/2013/05/17/mataerial-3d-printer-by-petr-novikov-sasa-jokic-and-joris-laarman-studio/
Mataerial looks quite interesting. But as which reprap has more than 3 axis?
with a 4th or even 5th axis you might get more freedom for sculptures which would need supportive materials or splitting to more parts on usual 3-axis printers
This one too: http://www.engineering.com/3DPrinting/3DPrintingArticles/ArticleID/6081/Freeform-3D-Printing-with-an-Undo-Ability.aspx
@Mic_Possert many of the budget robot arms don’t have anywhere near the resolution and accuracy that would lead to an improvement in quality.
While i do like the general idea of five-axis printing, i fear that it would require a fairly expensive mechanism to properly implement the rotary axes; be a nightmare to calibrate and write software for and generally not improve quality enough to justify the bit of extra quality.
We’re getting closer to having usable water-soluble support material, which, in my opinion, is a much better solution for printing overhangs.
Does my recent dual x-carriage printer project count? (http://www.thingiverse.com/thing:115885) That’s 4 independently driven movement components - although its not really another degree of freedom in movement I guess.
@Robert_F-C As excellent as you’re modification is, the extra head is still only moving along the x axis so It dosen’t really class as a 4th axis.
When your building one layer at a time is a 4th axis at all helpful?
@Thomas_Ellison with four or five axes, layers wouldn’t have to be limited to being parallel to the XY plane anymore. Imagine a figurine with its arms sticking out - the printer could start by building the torso from regular XY layers and then add the arms with layers parallel to ZY. Since FDM doesn’t care much about which way gravity is pulling, the arms would come out perfectly without any need for support material.
The layers wouldn’t even have to be a constant thickness along planes, but could bend and change their thickness to smoothly wrap around the model.
Now, that got me thinking. A Delta printer typically has three arms, which are each actually two beams tied together. If we were able to control each beam individually, that would allow us to tilt and rotate the platform in any direction possible, right? Six motors would result in six degrees of freedom (along and about XYZ), so while rotating the hotend around its Z axis wouldn’t be terribly useful (but possibly allow for some kick-ass nozzle designs), the other two extra degrees of freedom should be exactly what we’re looking for.
@Thomas_Sanladerer I can see your point about how the extra axis could facilitate an entirely different approach to FDM but would it not be easier to move the workpiece as with and A/B axis on a CNC machine? that only requires 2 extra motors and uses a fairly well tried and tested mechanical solution. Granted, it might raise some issues about how stable a model is and how good it’s bed adhesion is when it’s rotated around X/Y axis. GCode would be interesting, I guess a couple of extra axial coordinates would be needed that would allow the object to be rotated so the workplane was aligned with the extruders plane of movement. Slicing would be even more interesting…
@Tim_Rastall adding more motors to a delta setup would make the motion control system quite a bit more complicated - but the main reason why Deltas seem like such a perfect platform for this is that all axes can already travel at pretty high speeds. As soon as you leave the XY plane, Z will have to be as fast as the other axes.
I was also thinking along the lines of leaving the workpiece stationary and moving the head around it, as moving object would call for huge printer volumes (just look at how large basic 5-axis mills are compared to their working area, which often have one or two axes integrated into the table).
Slicing would have to be based on totally new approaches and algorithms if it should handle many-axis layers. However, simply choosing a part of the model and telling it to print that parallel to a different plane would be fairly simple, as it’s just a 2-axis (and the extruder) operation, which has been translated to a different orientation.
The gcode itself should be no big deal, either. Instead of controlling XYZE, we’d have XYZABE coordinates. It is then up to the controller to properly move the bot around, which, for cartesian robots, is the same thing as a XYZE bot, but with two more independent motors.
I’d prefer a rotating bed for the object and an extra pivot for the extruder - that way we wouldn’t have to mess with the gravity as an 5-axis-millingmaching where the object is properly fixed 
But where to start at?
Currently I still have no 3D-Printer as I’m out of money… but for the beginning I would only need a heatable bed and an extruder, cause I already built a wooden prototype of cheap componente… And I need some hours to solder my stepper driver for tge third axis again
Than I could do my first steps in 3D printing. But as I need to spend some money on better components one day, I’d be very pleased to keep the option open to add a fourth and a fifth axis.
Btw. would a 5mm aluminum plate work well as a cheap heat bed? Or would it be too thin or a bad material for that purpose?
@Michael_Weber 5mm plate would be fine as long as you intend to keep it static on the x/y plane. If it’s going to move back and forth on the y, it’s momentum will likely overcome the holding torque of your y stepper unless it’s a beefy nema21 (or whatever) or you run very low acceleration values. I use a 3mm plate and still have to be careful with acceleration settings. You might want a glass sheet on top and a structure underneath to allow for attachment to z/y rails and levelling, so overall weight can be a challenge.
@Thomas_Sanladerer OK, you’ve convinced me that a delta approach would make more sense in this case. Because I’m a Project Manager by profession, I tend to see problems as opposed to opportunities so here are a few I can think of:
Collision detection: the firmware would need to be aware of not only the shape of model but also the shape of the nozzle/carriage and potentially the position in space of the 6 rods. I’m thinking of the example of a figurine with it’s arms out and how the tool head would need to deposit material on either side of the model for the arms (y-z plane) and therefore have to find a path above/around it to reposition the head to do the other side witout crashing into the model. Not impossible I know, as it’s what happens with industrial bots and CA/CB milling tools but a whole new area of development for 3d printers.
Maximum angle you could orient the nozzle to relative to the z axis: I assume that you’d want to be able to at least get the nozzle oriented perpendicular to the Z axis, taking the delta approach with 6 rods, I can’t see how you could achieve that without a rod fouling against a Bowden tube coming out to the back of the carriage, unless of course the rods were telescopic…
On the positive side, if you could get it working, the non FDM applications are quite interesting, painting with an airbrush for example.