Me and two friends built this robot arm for our mechatronics class at CMU in spring of 2004. It was capable of shuffling cards, cutting the deck twice, and dealing cards to any number of players.
(Interestingly, the three of us formed a team of electrical engineer, mechanical engineer, and computer scientist - a perfect spectrum!)
At the time I had a poor quality digital camera that could only take 15 seconds of film, so I was never able to film an entire successful run. I made sure I included the bad prototype runs in the video, but towards the end of the class we got it to work perfectly 75% of the time. The other 25% was due to some occasional mysterious encoder error we were never able to figure out.
Enjoy the video!
Anyway, the first challenge was to design a way to pick up individual cards. After much thought we decided a suction cup would be best. This is our initial design of an arm with a suction cup. Obviously 3D CAD wasn't so common back in the day . . .
To test the concept we bult a mockup - a suction cup attached to a stick!
Then we hooked up the stick to two servos to see if we could apply enough suction force to lift the card up (just like in the above sketch). The servos were remote controlled. We found that a small weight (a steel bolt) attached to the stick end helped increase force.
It worked decently well, so we continued on . . .
So I used 2D CAD to throw some measurements together . . .
And got it to work like this animation. The edged blades were used to cut and pick stuff up, while the suction cup just picked up individual cards.
Here is a close-up of the gripper. This is after we added the servo and driving linkages. Its hard to see, but between the center linear shaft and the rotating edged clamps are two high-friction (no slip) rubber wheels.
And a mathematical analysis . . .
And more math . . . a basic statics analysis.
Here you can see a mockup built from scrap wood.
If you haven't noticed already, we attached our batteries on the back end of the robot arm. Why? Well, we needed a counter weight, and batteries are heavy . . . obvious now? =P
The arm had two power sources: one 7.2 NiCAD battery (regulated to 5V) for the electronics, while two other unregulated 7.2V NiCAD's were put in parallel to drive the motor and two servos. I wonder what these mislabeled buttons do?
The whole system never drew more than about 3 amps at any time.
If you want to learn more about the theory behind building and programming robot arms, please check out my robot arm tutorial page.
Has this site helped you with your robot? Give us credit -
link back, and help others in the forums!
Society of Robots copyright 2005-2013