Incredible polystyrene 5DOF robotic arm

[letsmakerobots.com]
[hackaday.com]

The notes regarding suping up hobby servos seem invaluable.

I'm semi-torn on hobby servo arms. They seem like a good stop gap while other servo technologies are tied up in R&D, and allow for Inverse Kinematics work to move forward, but I don't want to get too attached to their mechanical designs because their design considerations are built around an entirely different animal than the R&D servos.

These robots are great for pick and place assembly, but I really don't think it is accurate enough to print its own parts.

Interesting issue you raise. Begs the question what parts of a polar arm are critical dimensions? If you combine 3 will the error average out? How small of a work envelope is necessary for acceptable accuracy?

If you used a jig to drill the pivot holes: I think the dimensions could be incredibly sloppy.

Another question being: could photo-lithography be applied for laminated structures?

I had been thinking along similar lines with hobby servo's. Particulalry looking at the Dynamixel.

[www.crustcrawler.com]

It is all a bit rumsfeld, knowns unknowns and what we don't know might be known etc.

The only real way forward is for a number of folks to experiment and see.

Something to add to your reading is :-

[openservo.com]

You can also get a lot more out of the standard analogue hobby servo.

[forums.trossenrobotics.com]

Gives positional feedback

Measuring the current to the servo gives force feedback.

There is also a lot to be learned from angle poise lamps. I nearly got thrown out of ikea for studying their range for too long. The joints and arms on them are organized mechanically such that the lamp always maintains it's angle to the worktop as you move the arms. Without clever calculation. It is a feature of the mechanics.

Redundant joints are not necessarily redundant they can add resolution. Notably so in the case where the positioning is discreet (ie digitally derived) as opposed to fully analogue, continuous or non discrete.

An finally on the kinematics, for many cases there are several solutions to each end effector position. Given that only one solution in each case gives the best resolution together with best accuracy and precision. Again particularly when considering digital or discreet positioning.

Take all of the above and it is worth building a quick and dirty arm that will do little or no actual work just to go through the research needed to get close enough to a good enough end solution. Then re-implement in something more durable and stronger.

Incidentally some of the newer generation of hitech digital servos have two way comms and are programmable.

[www.hitecrcd.com]

My 2 pence worth for what it is worth.

Cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

I'm beginning to think a SCARA is all the arm a reprap calls for. Now it's just a matter of building one so I can begin working on improving accuracy & calibration. The IK seem stupid simple to me.