Great Ball Contraption

A couple months before the August 2005 Brickfest in Virginia, a couple friends and I got together weekly for a couple hours to make an entry for the Great Ball Contraption event.  We didn't actually plan to go, since it was on the other side of the country.  But the goal to make an entry gave us a common focus.  We came up with maybe 100 ideas for cool things to do that involved moving Lego basketball and soccer balls to and fro.  Substituting marbles for the soccer balls we didn't have, we soon found out that 90% of our ideas were fantasy when compared with our ability to control the marbles' behavior and the limitations of the Lego pieces.  Make that 95%.  Still, we made a number of proof of concept models and eventually got one fully assembled module that successfully transported about 80% of the marbles through from input to output.  This mostly-working module was first operational about 1 day before Brickfest. 



We experienced two major difficulties: creating a stream of marbles from a batch, and raising the marbles. 
No matter what physical shape we made the box or ramp or whatever you want to call it, there was always SOME configuration where multiple marbles could (and would) get jammed together, stopping the stream of marbles leaving the box.  One way out of that is to assume that marbles arrive at the input bin one at a time.  Not very flexible, but possible.  The other way out is to stir the marbles such that they can't stay in a jammed configuration.  An agitator.  We thought we could get away with passive marble manipulation.  We didn't use an agitator; consequently we became agitated.
The other difficulty was raising the marbles.  Gravity helps them roll down just fine, thank you very much.  Greg made a very cool mechanism that moved kind of like a magician rolling a quarter over his fingers.  The rotating elements seemed to almost go through each other.  The marbles got passed from one rotator to the next.  There were 4 rotators, limited by the number of one particular part.  Unfortunately, it was a very tempermental mechanism.  Another way to make things go up is by chain.  Lego makes chain, but the connections are weak and there is little way to connect pieces to the chain.  There are many other methods of raising things.  The method we used on the 'working' module was a ferris wheel.  The mechanism to feed a single marble into the cage as it came by mostly worked, but also was a little tempermental.  When the cage reached the top, the marble rolled out onto a ramp.  That description makes it sound like a nice, smooth, reliable process.  Hardly!  It was more like the marble fell out of the cage at an angle, dropped a couple inches, and bounced onto (and sometimes off of) the ramp.



A year came and went.  'Round about middle of August 2006, I found a good deal on a couple thousand pieces at a garage sale ($5 !!!), and while sorting through them got the bug.  The previous year we made some models that threatened to work, but never actually did.  This time I was able to work out a pump idea, that raised the balls a few inches. 


Ball Pump Movie (7 MB)

A week later, I figured out a kicker; it can raise the balls a foot or more.  You'll notice the tower is mostly Duplo blocks: they are bigger and make quick work of a tower.

Then I needed to feed the kicker with one ball at a time, so I made a couple gate mechanisms that operate from rotation, so they require no special contraptions. 


At the time I write this, it is the end of August 2006, and I have just finished the gate mechanisms.  I have not yet connected the gates to the kicker.
I want to make an input bin with an agitator, so I can connect all these pieces into one module.