Hello,
I’m in the planning stage of building a mini PUMA arm using the sort of parts and servos that Lynxmotion sells. Does anyone have any advice on how you would put together a mechanism like the PUMA? For those not familiar with a PUMA-type robot see webpages.uncc.edu/~jmvannoy/puma/.
I’m not necessarily trying to make it look just like a real PUMA but I want to have the same basic mechanism. I have looked at the “What can I build” examples such as the aluminum tubing arm, and I’m thinking the low-profile axis assembly is going to be key to making this work.
Let me clarify, I’m not trying to duplicate the same layout of the motors as those cutaway diagrams show; I just want to duplicate the same joint positions and axes. Actuating with servos is the goal here (not the complex shaft-driven approach shown above).
If you’re willing to scale it down (the old PUMA relics at my college are as tall as I am, when fully extended), it might be possible.
You’ll also have to keep within the confines of a servo’s 180 degrees max rotation.
For joints that need more rotation than that, combining two servos on the same axis would give you 360 degrees.
The other limiting factor would be payload.
With quality servos (say 645’s) a tabletop-sized arm with that many joints wouldn’t have any real carrying capabilities.
If you go with 5645’s or better, you’ll get more mobility, but don’t expect to be lifting anything over a half pound or so.
I’d suggest using digital servos.
While you won’t be able to recreate the strength of that industrial arm, you may very well be able to come close to its precision with digital servos and good IK programming.
Oh, by the way…
It sounds like you want to replicate that long joint by using the hollow aluminum tubing.
It’s certaintly possible, but only if you support the tubing with bearings.
Hanging a load far off of a servo’s output shaft is asking for trouble.
Yep, a scaled-down version is exactly what I’m going for. And the 180-deg. joint limit is OK (for now).
I’m not too concerned about payload right now; I’m more concerned about precision and repeatability in the joint positions. So I believe the digital servos, as you suggested, would be appropriate.
Also, you’re exactly right about supporting each joint with bearings. I don’t want any of the joints be composed simply of the servo’s output shaft. That’s where I think the LPA assembly would come into play. The part I’m having the most trouble with here is how to setup the shoulder & elbow joints so that the weight of the link is not on the servo shaft, and yet a servo can actuate that joint.
I think I can get 3 dof in the wrist if I setup the forearm to rotate, and then separately setup the gripper to rotate. All this plus a “wrist bend” joint similar to the one in the sample tubing arm.
the 4 little holes hold are for the servo, to hook it up, the bigger hole in the center is for the servo horn to rest, the part that goes over the output shaft, it rests ion there, i was going to use a ball bearing, but i didn’t have the room of there, but you might be able too because you will probably have more space than me
, just realized that my idea may not work for you, because the other side of the arm would not be supported unless you make the servo go in the middle of the arm and then make the casing around that
, just realized that my idea may not work for you, because the other side of the arm would not be supported unless you make the servo go in the middle of the arm and then make the casing around that