I am interested in developing/building a printer specifically for printing sockets for prosthetic limbs in a robust and timely manner. Current "state-of-the-art" printers use a single large nozzle in vase mode to print rapidly, but this creates limitations on print thickness and other challenges associated with excessively large nozzles.

I am pondering using two independent print heads on a rotating (polar) print bed to allow for simultaneous vase-mode printing (printing at 180 degrees from each other), which could allow for many advantages such as larger thickness, reduced time, and/or multi-material sockets (soft inner, hard outer). Picture something like a Flashforge but with two independent carriages and a "4th axis" rotating bed sitting on the Z carriage.

Assuming the part is properly centerred, the independent print heads would compensate for the difference in printing speed that results from being a different radial difference away from the center of the rotating bed, but could share some linear componentry to reduce cost.

Anyway, just spitballing here before I get too deep into the design. Obviously two hurdles--firmware and mechanical performance. Any thoughts? Can make a CAD mock-up in the next few days if it would help to visualize the concept, but wanted to gut-check anything obvious before going too far.

Edited 1 time(s). Last edit at 05/18/2019 03:44PM by ChayDuuLetDuu.

Interesting concept. Another hurdle is going to be writing a slicer that would be able make toolpaths, unless the 3d printer can figure out what goes where. Im no expert, but I'm not aware of any existing slicer that works with this motion system. I'd create a poc slicer and a simulation of the motion system first to iron out the most obvious problems. How are the two extruded paths joined? There'll be two seams to deal with if not using the vase mode.

I've been experimenting with using perimeter only prints instead of vase mode. The thing with vase mode, at least with a .4mm nozzle is you can't use a higher velocity or the extruded filament will take a shortcut, not adhering to the layer below.
When printing 3 shell perimeters I can print faster, making the model stronger and allowing to print more complex objects, not being limited to a continuous line. It prints in about the same speed if you dont print bottom or top infill. It's easier to make the walls water tight so you would be able to make lost plastic molds or use resin reinforcement (stick the open bottom in clay or silicone).

---

--
Kind regards
Imqqmi

NFAN CoreXY printer:
[reprap.org]

Appreciate the feedback!

Definitely understand the software will be a significant part of the battle, but willing to take the plunge on that end if the mechanics make sense. From my initial figuring, in 'dual vase mode' could be accomplished just by generating two vase toolpaths 180 degrees apart (+ a small radial offset for inner/outer perimeter). HOWEVER, when you start trying to balance print speed due to the difference in radial distance from the center of the print bed, things get a little messy. Might have to play around with it a bit more as it seems it will be a bigger hurdle than the mechanics.

I'll have to look into the vase mode vs perimeter print thing. My feeling is that vase mode would be faster and the continuous printing might be stronger in the Z direction (which is critical for prosthetics). But not sure on that.

Edited 3 time(s). Last edit at 05/19/2019 04:17PM by ChayDuuLetDuu.

Since the two toolheads are basically printing the same thing (except 1 wall thickness difference on the radius), and if you print in vase mode (i.e. continous, unidirectional rotation), you could just make the second toolhead echo the gcode of the first one, with a 180* delay and a 1 extrusion width offset. This could be done in the firmware possibly (i.e. every time a command is sent for the rotary axis, make the first toolhead do it's thing and the second toolhead do what the first was ordered to do 180* ago, with the wall thickness offset) or with a post processing script that adds new gcode for the second toolhead. This sounds easier to me than making a new slicer to handle this sort of dual toolhead synchronized printing, but I could be wrong. If each layer of the part has a line of symetry (I doubt it would, but I don't know much about prosthetics), you could even mechanically couple the two toolheads and skip any software work.

Edit: With a bit more effort coding wise (not much I think), you could get rid of the "unidirectional rotation" constraint. Echoing the gcode would create some slightly incorrect extrusion on the second toolhead, but seeing as the difference in the radius it is printing will be quite small, it may not be noticable and you could probably fudge fix it by increasing the extrusion multiplier. If that fails, the math is simple anyways. This solution wouldn't work in cases where you need to print near the center, as the toolheads would crash into each other, but there are a few ways to get around that mechanically I think.

Edited 1 time(s). Last edit at 05/23/2019 04:06AM by Trakyan.

This is great feedback. My plan for "printing near the center" is to use a single toolhead to print the base, and then move to dual head extrusion for the sidewalls, hoping proper alignment of the model can keep any of the walls from crossing (or nearing) the center of the print bed to prevent crashing of toolheads.

Design does hinge on how much radial variation there may be at 180 degree printing points, but I hope your right, if deviation generally is low, I can probably use the extruder to compensate by having the rotating bed bed the average of the two required speeds and then slowing or speeding up extrusion to compennsate. This would probably be more likely in leg sockets than arm sockets due to average size.

You could angle the toolheads (like angled belt printers do) with the nozzles pointing towards each other, that way they're have less interference. With a large enough angle/thin enough nozzle you could get the minimum gap between the extrusion lines down to a few mm. Would a small hole in the top of the print cause any issues for the functionality? I think varying extruder speed should work fine, extruders have a fairly wide band with regards to print speeds that they can operate at, you'd just need to make sure the table rotates at the slower of the two speeds each toolhead requires (shouldn't be too hard, firmware already slows down different axis to prevent the one with a higher acceleration value or less acceleration needed of getting ahead of and out of synch with the other axis). Assuming a reasonable difference in print speeds between the two toolheads I think it's best to leave it to the extruder to compensate.

This next bit is me making a few assumptions, please correct the ones that are wrong.

Wouldn't the base just be the first layer of the side walls, and the top be the end of the socket? (i.e. you insert the limb through the part that was touching the print surface). This seems like the appropriate print orientation to me, seeing as the socket would taper (I think?) as it goes towards the closed off end, and having a print that tapers rather than grows as it goes up seems like a better idea. That being said, the tilted toolhead design would allow you to print a close off base if needed, as long as a small hole near the center is acceptable.