Having taken a sabbatical from programming Big Chaser, I decided to do a ball bot.
There are a pair of pudding bowls or mixing bowls attached directly onto a threaded bar. Also attached to this bar is a servo (modified for continuous rotation). When the servo is powered, it moves the bar, attempting to turn it and the pudding bowls.
Hanging from and perpendicular to this servo is an unmodified servo, attached to which is the battery and the control. When this second servo moves, it swings the battery etc (the "weight") to either side, causing the ball to steer.
When the main servo is powered, it causes the battery to attempt to "rotate" around the bar, which it cannot do, so the ball rolls forward.
When the microcontroller detects a current over a certain amount, it assumes the ball has hit an obstacle and attempts a three-point-turn.
Yes, clearly the device is not yet under microprocessor control. In fact, you can hear me twanging the radio control joysticks in the background, but what's written here is a goal.
I wanted the "ball" to be as near spherical as possible so I cut the rims off the bowls. Of course this was what had given them their strength. After several attempts, I came up with the idea of making interlocking teeth on the edges of the bowls and that's what I'm trying next.
This is a bunch of photos and videos of the gubbins being assembled and working. I don't know if I'll do much more with it, now. I've sort of demonstrated the principle.
18-Oct-2009
I'm getting all fired up about the mark II. Look what my wife decided to throw out:
That's gonna be far more managable than the two pudding bowls!!
Rolls about under microprocessor control. Avoids obstacles.
Actuators / output devices: Acoms R/C Servos
Control method: Autonomous.
CPU: Microchip PIC 16F628
Power source: 4x AA 1.5V cells
Programming language: Microchip RISC
Sensors / input devices: None. Current sense only.
This is a terribly terribly cool and simplistic aproach to basic locomotion. So cool, in fact that it gave me an idea…
What would happen if you mounted a hard drive motor/plater on the a servo and mounter that arrangement on the setup you allready have? if you “whipped” it around in the same way - the basic idea would be that the center servo is attached to the center post as you have done, with it’s axis around the center of the sphere. The second servo rides on top of the center servo oposed 90deg with the motor/plater mounted on top of that. The top servo/motor/plater would provide stability as well as affect a right or left movement.
Thanks. There were a couple of “round” robots appearing and that’s what gave me the idea.
Yes. In fact, there’s a guided misile which works on a similar principle. The “whipping around” is the problem. If you had the platter spinning and suddenly braked it, it might produce enough movement to rotate the 'bot. I don’t know if the platter could be spun up quickly enough to affect a movement, though.
FF rocks! as your robot does! What i like is the fact that it is different from many other robots and as casca said, it also has a neat and simple "motor" inside. Your son also seems to like it!
good to see you have enough balls to get a project like this to work
Will the microprocessor addition be next? It’s a an interesting solution vs the pushing by gyroscope which I guess would need a gimble setup. Senors should be challenging too: When you said that one of the feedback messages would be “when it stops” - are you going to use an encoder or some other means to determine when its stopped. I imagine (depending on the material) IR would work rather well, but not sonar, capacitance changes could be possible on a metal ball maybe…
Keep up the great work (aren’t you supposed to be at the relatives?)
To be honest, the locomotion needs some help. First of all, it really needs a bigger motor. I have a perfect one. It’s from an old casette player and it has a lovely gearbox built-in. Trouble is it’s 15V. So I need a new battery, too. Extra weight probably means it needs a heavier steering servo.
Perhaps more importantly, the cross-section of the pudding bowls has a pointy bit which means it always wants to fall one way or the other. If you suddently change diection, it kind of spins and it’s a bit unpredictable. There is no arrangement of the bowls which has a flat spot. I’m at a point where the bowls are dispensable, so next I will cut downwards towards the base of the bowl and try to manufacture a flat edge by melting it back into shape.
This polycarbonate is a pain to saw. It’s nice and flexible if you work it slowly, but if you go too fast with a saw, or hit it too hard, it basically shatters.
You are spectacularly astute. Most people would have been foolish and insensitive enough to say “holiday”. We’re away on Thursday morning.
oh cool you’re going on holiday boa? ( well seriously, maybe staying away from the pc for a while helps you relax a bit)
anyways i too think a geared motor would be better than a servo in that case, it may make the whole thing faster but i am worried about precision and maneuverability (i might have spelled that badly)
Better Yes, that woud have been better. My main problem was the seam bewteen the two bowls. I didn’t even know large hampster balls existed. If I can find a more spherical shell of some sort I will ressurect this project.
There are a pair of pudding bowls or mixing bowls attached directly onto a threaded bar. Also attached to this bar is a servo (modified for continuous rotation). When the servo is powered, it moves the bar, attempting to turn it and the pudding bowls.
Hanging from and perpendicular to this servo is an unmodified servo, attached to which is the battery and the control. When this second servo moves, it swings the battery etc (the "weight") to either side, causing the ball to steer.
When the main servo is powered, it causes the battery to attempt to "rotate" around the bar, which it cannot do, so the ball rolls forward.
When the microcontroller detects a current over a certain amount, it assumes the ball has hit an obstacle and attempts a three-point-turn.
