Six-Arm Simpson

I've been thinking about 6-dof bots and the Sextupteron. The toolhead rigidity in Sextupteron concerns me, so I've been thinking about other geometries.

What about a six-arm Simpson with a hexagonal toolhead? Actually, a hexagonal toolhead isn't ideal, it needs to have three long sides and three short sides to prevent twisting in the toolhead. This would still give a very wide work range, but it has the advantage that all six arms work together to bolster the rigidity, rather than having each axis's rigidity determined only by a pair of arms. Some might be concerned that Simpson's torsion springs will be a limiting factor on milling force (if used as a mill instead of a printer) , but remember that with six arms, three are pulling (driven) while three are pushing (un-driven).

Simpson's apex even makes a natural toolhead switching area.

Here's how a commercial CNC mill with 6-arms works:
[www.youtube.com]

Six-Arm Simpson would just be this inverted; the Annirak Drive instead of overhead arms.

I can't figure this out in my head but say we have a standard Maggie with 6 arms instead of 3. We change the hub to a hexagon with 6 ball joints (or other 3DOF joint) at each of the vertices. Will the hub be fully constrained?

It's not fully constrained if you have an equilateral hexagon. It is fully constrained if you have a non-equilateral hexagon. In that respect, it has a lot in common with the Stewart Platform. In fact, you could use the Stewart Platform as a model for a Six-arm Simpson

There are two ways to constrain it: either make the base non-equilateral, or make the hub non-equilateral, or, even better, both. The closer you get to triangular on either the base or the hub, the more rigid the hub's position becomes. You can see that in the video I linked, they made both the hub and the base non-equilateral. That is, the arms are in pairs on the base, but they are in opposite pairs on the hub.

Okay so I have 6 arms that connect in the alternating triangle pattern you mentions. What do I need for hub and shoulder joints? Do the bottom ones need to be 2DOF and the top ones need to be 3DOF? The Stewart platforms I have looked at have a u-joint on the bottom and u-joints on a swivel up top.

Separate Note: A Stewart platform usually has a very limited range of motion. I guess I can get around some of that by having actuators that have a larger range of motion than what they usually use. Does that track?

I believe spider / six arm machines are promising:
[youtu.be]
[youtu.be]
[3dprintingindustry.com]

Also, this is inspiring:
[youtu.be]

Edited 1 time(s). Last edit at 08/17/2013 03:46PM by Buytaert.

That 5DOF machine in the last one is amazing. It is a cross between Simpson and Wally. The only thing I don't like is the one linear component. Thanks so much for the link!

I think their bulky design is limited by weight and super precision need. So there're rooms that you can improved.
Thank you!

It's a bit off topic, but how would you create toolpaths for such a device? Commercial 5-axis CAM software for CNC milling looks to run thousands of US$. I don't even know if there's a concept of 5-axis FDM to find software for.


I'm comfortable enough with the concept of 3-axis fdm that I could probably muddle through writing a slicer. I'm not sure where I'd start with 5DOF.

Also, and maybe I'm missing something here, but I think 5DOF is the most you can use in a device like this: 3DOF for position and 2DOF for orientation. Controlling tool axial rotation seems rather silly. Could you get away with 5 arms?

I have experience with the algorithms and the necessary math. However, at first I plan to make one-off gcode programs to demonstrate one advantage at a time which will hopefully drive others crazy until they write a general solution. :-)

Edit: 6DOF is for symmetry. Besides you can add a light duty gripper for assembly.

Edited 1 time(s). Last edit at 08/16/2013 09:25AM by nicholas.seward.