hexegrity

machine:design:hexapod:hexegrity

hexapod tensegrity  

As far as I know, this is a new idea. It's like a robocrane, except the six cables are arranged both above and below the platform, allowing force to be applied in any direction. The hexegrity can be made out of steel cables or timing belts anchored to the three orange "towers", and a central "spool" or octahedral compression frame which provides rotational leverage. Adjusting the length of the cables moves the frame in translation and rotation. Although an octahedral frame is more rigid and convenient, the spool version is shown because it's easier to visualize in a simple drawing.

Miguel here. Yes, it is a new idea, and I believe it holds more future than any other in parallel machining. None of its components is expensive and requieres minimum accuracy in the assembly to deliver good results. I bet a decent machine (capable of milling large pieces of steel) can be made for less than $1,000 USD.

Click for an interactive Java applet
(note: this has been reported to crash some browsers)

A simple tripod crane does not constrain the platform rotationally, and also cannot apply force downwards. A double tripod crane is overconstrained, allowing one to vary the tension or preload on a wire connecting the two tripods, but if your control algorithm is wrong the machine will rip itself apart. This preloading is normally a result of gravity and pulls the whole platform down, but we would like a little more control, so I've been playing with various arrangements of springs to get the right preload vector.

The timing belt swivel anchor. There would be one of these at each corner of the octahedron.

a mockup made of straws, wire, and legos

Problem:
gravity causes the cables to hang in a hyperbolic curve, regardless how much preload. Any change in the tension (due to changes in acceleration) will cause the cable to "twang" or vibrate perpendicular to its length at its natural frequency.

Solution: instead of a timing belt, use a measuring tape, which has an integral stiffening mechanism. Another benefit of measuring tape is that they can be used as a linear encoder, probably best accomplished with a moire mask. This also reduces cost of pulleys and belt significantly.

Another (silly) solution:
Run cables down one side of an inflatable cylinder, like a long balloon. Of course you have to fill the balloon with air somehow... (probably an air line going to the platform)

Another (really silly) solution:
instead of motors driving winches, use Airmuscles as the cables.

cheap gearbox substitute:
http://en.wikipedia.org/wiki/Differential_pulley

Miguel here. Six cables, each with an actuator, is the bare minimum to give the machine complete stiffness and liberty of motion in all axes. The good thing is that cables are dead cheap and can be actuated by regular acme screws actuated by directly coupled stepper motors. This is because, unlike a regular mill table, the tension of the cables takes up all the backlash. It is perfectly possible to excert down force (and to drill by extension) with six actuators.

Also, gravity is a non-issue for "serious" machines, as it is negligible as a force compared to the loads of milling, for example. For anything above the softest materials we need tons of preload in the cables. We're looking to build a machine weighting at least 100 Kg if we want it to be practical.