Robot arm with interchangable head?

Instead of using a 3d cartesian robot for the printer is there any reason why a 'simple' robot arm couldn't be used. A robot arm could also easily change heads (or carry multiple heads).
they would also be useful for CNC machines. Im assuming that there are 2 main reasons they haven't been used: 1: they were expensive and complex back when CNC machines were being developed so noone really thought of using them. 2: Accuracy is low.


Accuaracy is something that can be worked on though.
A robot arm based printer could also pick and place electronic components with the right tool.

Just a thought. I'm still going to build a regular reprap!

that is an excellent idea and one i personally support and one that would be interesting to experiment with.

however, we went with a cartesian robot because of the simplicity. once we get the machine out into the wild, you should definitely try and develop the parts in thermoplastic. i'd imagine a easy way to produce robotic arms would be very exciting.

I love the robot-arm idea. But it is really a 2.0 or later type idea. It is a complexity thing - but I'm sure that after 1.0 is out and there are many people experimenting we will see plenty of work towards that goal. With that said - the important thing is getting 1.0 done.

It would definitely be a cool thing to make one day. I believe the main reason for not using them (which you already suggested) is that it has some stability issues. Making it as precise and stable with a tool on it is much more difficult than something that is supported on both ends.

But, guaranteed, somebody will make one...

I did some poking around, and the small robot arms (serial robots) that kuka makes:
[www.kuka.com]
have a positioning accuracy of 0.02-0.05 mm. That's not a good as a cnc machining center with a comparable sized workspace, and a positioning accuracy of .0002 mil (0.005 mm) but you could probably improve it by adding rotary or linear encoders and crossbracing, (i.e. making it more look like a RepRap.) I think small parallel robots (aka stewart platforms,delta robots) have much better error than small serial robots, but I can't find any numbers just now.

As Simon says, get Darwin out the door, then experiment.

I glanced at Kuka's page. Looks expensive, when they invite you to a one on one experience rather than just posting a price list.

Agreed. I don't know of the robots in the continuum between Kuka and these robot arm kits:
[www.hobbytron.com]
but there must be a bunch of them out there. Extruder heads are light enough that repurposed pick and place robots might do the job; although I think I'd start by trying to RepRap a robot arm.

I'd think Forrest has that as a goal for his telepresence project but he's been busy. I certainly think that lots of what we have would be suitable for an arm as part of an enhancement for manipulation and self feeding etc.