What motor drive for an arm?

[nervis1]

I need to drive one joint in one plane of motion, like an elbow...exactly like an elbow with 50 or so degrees of movement, I need feedback so the thing will reverse at the end of the throw, and need to be able to adjust the throw. The movement needs to be slow, like 20 ipm, maybe a bit more.

I can take care of the mechanics but I have no idea how to drive this thing. I have a ball screw and motor but it's a bit clumsy for this app. Is there a rotary motor that can drive an arm like this directly at the joint with the feedback built in? A google search did not turn up much but I don't think I'm looking for the right thing, I really don't know what this type of motor / mechanism is called.

Any guidance or links you can provide will be greatly appreciated.

[murphy625]

http://www.motionvillage.com/products/motors/ddr/

Is that what you want?

Murphy

[nervis1]

Well that might work, if the unit is small enough to fit at the joint without hitting anything. I'll give em a call. Thanks for the link.

[Evodyne]

Nervis,

Hey! You didn't really talk much about the load. Also, you stated you want it to reverse at the end of the throw. What it made me think about is the windshield wiper setup on a car. A gearmotor (runs slow, tons of torque) has a small bellcrank and is located at the "shoulder". The "forearm" is attached at the "elbow" and has a longer bellcrank. The two bellcranks are tied togther via a tierod.

Activate the motor and the forearm will "wave" once per motor revolution. Options would be a pulse-width modulated drive to speed-slow the motor (wave faster or slower), a limit switch to reset the circuit once per revolution (one wave only each time activated), and an adjustable length bellcrank and tierod (to determine the arc the forearm swings through).

Pretty simple.

[nervis1]

Good idea. I think it may need to be more sophisticated than that. saftey is one issue, more lever arms is more potential pinch points, all of that will need guarding...more weight...more money. I'd like to try and keep the motor directly coupled to the joint if I can.

It needs to be kept light, simple, cost isn't an issue...yet.

I'll need to make this thing adjustable, via a controller, speed and throw in degrees. I'll need it to move slow... like across 90 degrees in one or two min.

I talked to an applications engineer today that said I'll need a geared servo. Lot's of tourqe. The load will be placed across the "forearm" of the elbow and the upper arm fixed. Load may range from 10-30 lbs. the forearm part is about 12" long,

I still have not figured out how to actually attach this motor to the arm, what kind of coupling to use.

Anyone who has done this type of thing...I'm all ears.

Thanks for the windshield idea, I just don't think thats what I'm after with this device though.

[rafe]

Considered servo motors yet?

Regards,
rafe

[Al_The_Man]

If you do not want to use a linear actuator to move the arm through an arc, which is effective but takes up space on the side of the arm, another way to get around the torque/speed requirements would be to use a servo driving a planetary gearbox, either in-line type or rightangle, Bayside are a popular make and show up reglularly on eBay.
The arm would be attached directly to the gearbox output shaft.
I have just completed a similar project, I used professional equipment obtained from ebay, Galil single axis stand-alone servo controller and a Maple Sytems OIT for a small operator interface to communicate with the Galil, via the RS232 port.
DC servo and amplifier.
Al.

[unterhaus]

Penn State had a robot that was all hydraulic cylinders that they sold for scrap. I wanted it, but realized that it would be incredibly stupid for me to buy.

My GE robot was chain drive from the base so the motors were remote from the joint, worked pretty well from what I understand, I scrapped it out, not worth poking a hole in the ceiling with. The chain drive was driven by a servomotor through a harmonic drive gearbox. The thing is that any one move will move the whole thing. I've been thinking about putting a gearbox and a motor at each joint. Harmonic drive gearboxes are pretty neat. They have big gear reduction in fairly small packages with no backlash.

If you have a reasonable amount of room in the arm, then the motor can be hidden inside, and the joint itself can incorporate the gearbox. One issue with moving the motors is that you reduce the weight capacity of the mechanism by the weight of the motor. It's a big concern in many cases.

Those Kollmorgen direct drive systems look pretty good. I forget the Japanese company that actually makes the motors for them. They show up on ebay every once in a while. Either way, it's $$$

[unterhaus]

The arm would be attached directly to the gearbox output shaft.
I have just completed a similar project, I used professional equipment obtained from ebay, Galil single axis stand-alone servo controller and a Maple Sytems OIT for a small operator interface to communicate with the Galil, via the RS232 port.
DC servo and amplifier.
Al.

Al, did you use gearboxes? I have been looking for a suitable substitute for the iRobot dexterous manipulator

It has 4 cameras onboard, which seems like overkill to me, but if you are picking up bombs I suppose you can't have too many views. I'm thinking their arm looks a lot like the arms that you see people manipulating nuclear waste with. They are not being forthcoming about selling us an arm, so I think I'll just mount one myself.

[Al_The_Man]

I have used Bayside and others, planetary GB, but as you say, if the arm is carrying the weight of the motor & GB, they can weight more than a few pounds.
I never had a weight restriction problem so I have not searched out light GB versions.
Al.