It’s a seemingly simple task: bounce a ping-pong ball on a wooden paddle. So simple that almost anyone can pick up a ball and a paddle and make a reasonable job of it. Now, close your eyes and try to do it just by the sound the ball makes when it hits the paddle. That’s a little tougher, but this stepper-driven platform juggler manages it with aplomb.
That’s not to say that the path to the finished product in the video below was a smooth one for [tkuhn]. He went through multiple iterations over the last two years, including a version that surrounded the juggling platform with a fence of phototransistors to track where the ball was at any time. That drove four stepper motors through a cross-linkage that popped the platform up at just the right moment to keep the ball moving, and at just the right angle to nudge it back toward the center of the platform. The current version of the platform does away with the optical sensors in favor of four small microphones. The mics pick up the sharp, well-defined sound of the ball hitting the platform, process the signal through an analog circuit, and use that signal to trigger a flip-flop if the signal exceeds a setpoint. An Arduino then measures the time delay between arriving signals, calculates the ball’s position on the platform, and drives the steppers through a PID loop to issue the corrective bounce.
The video below is entrancing, but we found ourselves wishing for a side view of the action too. It’s an impressive build nonetheless, one that reminds us of the many maze-runner and Stewart platform robots we’ve seen.
10 thoughts on “Juggling Machine Listens To The Bounce To Keep Ball In The Air”
this has a cool factor of 100%. can’t wait for the paddle version.
He actually did do a paddle version. It was prior to this one. Check out his blog.
Based on the the description and behavior, the control loops are bang-bang type. Imagine a full analog negative feedback loop on each axis. It would be less interesting to watch with the ball quickly finding the middle and just bouncing in place, but it would probably run until the wiggly wires wear loose. :-) ***EXCELLENT PROJECT*** (up to 11 on the cool-o-meter)
This is one of those insanely cool projects that just makes me think…. Why?
Is there a word in english for a question that contains its own answer?
Pretty cool (even if not actually juggling).
If the board is smooth enough, and
if the ball is smooth enough, and
if the precision of the mechanism is tight enough,
then it seems like you could eventually get it to converge to various kinds of steady state: bouncing straight up and down in one spot, bouncing back and forth between two spots, … various patterns with N spots.
How about multiple balls at once?
Now this is a fun project… though I wonder how the people in his direct vicinity react to it.
Watched the video for a few minutes… that was OK… but this project must have took many many hours of tonk-tonk-tonk-tonk-tonk-tonk-tonktonk….. tonk-tonk-tonk-tonk-tonk I guess that if I was living next door to someone with this kind of project I’m not sure if I would still think this as a fun project.
But since I don’t… cool project, interesting approach!
I’m surprised that the steppers can move the plate to an angle fast enough to keep the ball controlled, considering that it can’t start moving until slightly after the ball actually contacts the plate.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)