A Simpler 3D Printer design

The one thing that was bugging me about existing 3D printers was how overly complex and bulky they are. So I thought of a design that would be simpler.

Have a turntable that descends and rotates and have one belt driven arm. The only problem is that it could not use existing software.

Here is an animation of my design.

[youtu.be]

Very cool, I have made plans for something similar myself, it would be incredibly good at making circles and gears fast. Although I was thinking about having the entire gantry rotate, and have the table only move in Z.

The same software can be used, though the firmware would have to be changed.

---

--
-Nudel
Blog with RepRap Comic

Look at the Polar robots forum

Very interesting idea, Aardvark, although what Nudel suggests has me intriged. I was just asking another couple of people about the very same idea. This would reduce the amount of movement in the printed part. But the software and firmware issue is bothering me. Not knowing sqwat about programming, I am curious how existing programs could be applied to exactly the design you are talking about.

Does this really reduce complexity?

Still need at least 4 steppers (Theta, R, E, and Z)

The R travel would be half of a standard machine, but the total width would be about the same since you still need to accommodate the bed

Still need either 2 gears (backlash) or a belt and 2 pulleys for the theta drive system, and the R drive system would be identical in principle to the X of a normal machine.

In addition, the polar design makes fast lines near the center of the platform very difficult (since Theta would have to change very quickly to maintain a line)

The software is less developed and harder to grasp (I think in rectangular coordinated far more easily than polar)

Not saying don't pursue it, just think very carefully about whether it actually is as beneficial as it may look at first glance.

---

www.Fablicator.com

I think software and firmware are the largest issues. The slicing algorithm has to be very different.

Also, parts won't have a consistent resolution. A "step" of the turntable is going to be different distances depending on how close to the perimeter the print head is.

Just a supporting point, an extruder carriage could be mounted on each radius allowing each one full coverage of the table. Continuous path lines could be written with one expression for each change in direction instead of two(ie. length in relative direction, speed vs x,y, speed).

Nudel Wrote:
-------------------------------------------------------
> Very cool, I have made plans for something similar
> myself, it would be incredibly good at making
> circles and gears fast. Although I was thinking
> about having the entire gantry rotate, and have
> the table only move in Z.
>
> The same software can be used, though the firmware
> would have to be changed.

Unless I am misunderstanding you, this seems like a much more complicated solution. Rotating the bed is relatively trivial. Even making the bed move in Z and R is simple. Rotating the gantry is hard, Just one obvious thing you need to consider: How do you keep all the wires that go to your rotating gantry from getting tangled as it rotates? That is just one of several problems I see with rotating the gantry... Overall it would seem to add a lot of complexity for minimal if any benefit.

wallester Wrote:
-------------------------------------------------------
> Just a supporting point, an extruder carriage
> could be mounted on each radius allowing each one
> full coverage of the table. Continuous path lines
> could be written with one expression for each
> change in direction instead of two(ie. length in
> relative direction, speed vs x,y, speed).

That is only true if you are only moving on one axis. So for an arc centered on the origin of the printer and of the same radius as the current radius of the print head, that would be true. But for any other movement, you would still need to provide one or more additional axis movements. And of course all that is true of a traditional cartesian coordinate printer. If you are only moving in X, you only provide an X value in the gcode... The biggest difference I see is that moves only in X or Y are far more common than moves purely in the rotational axis likely would be, even if your firmware was optimized for a polar bot.

Andrew Diehl Wrote:
-------------------------------------------------------
> Does this really reduce complexity?
>
> Still need at least 4 steppers (Theta, R, E, and
> Z)
>
> The R travel would be half of a standard machine,
> but the total width would be about the same since
> you still need to accommodate the bed
>
> Still need either 2 gears (backlash) or a belt and
> 2 pulleys for the theta drive system, and the R
> drive system would be identical in principle to
> the X of a normal machine.
>
> In addition, the polar design makes fast lines
> near the center of the platform very difficult
> (since Theta would have to change very quickly to
> maintain a line)


Having the table go from 0 to 359 degree rotation back and fourth to print but going beyond 360 degrees causes the z screw to unlock and wind down. Inner axle is the table rotation the outer axle is the z screw.

So the total is 3 steppers including the feeder.

And yes it would have to go faster in the middle for pour consistency but nothing a stepper could not handle.

kludgineer Wrote:
> Unless I am misunderstanding you, this seems like
> a much more complicated solution.

My main goal was to be able to drop the printer onto a barrel or similar, to gain a lot of z height and benefit from the stiffness of the cylindrical form.

> How
> do you keep all the wires that go to your rotating
> gantry from getting tangled as it rotates?

That is precisely where I stopped developing the idea further.

---

--
-Nudel
Blog with RepRap Comic

Quote

And yes it would have to go faster in the middle for pour consistency but nothing a stepper could not handle.

Really, the infinite angular acceleration and velocity required for any vertex on the centre of the disc?

---

[www.hydraraptor.blogspot.com]

I am working on a similar design but with very low resolution. Like a 1 cm grid.
I am not a wiz at programming but I thought I'd just take points from a cartesian grid map and conveting them to polar and printing them one by one. It seems simple to me but I am sure I am missing something.