aeh5040

About 5 years ago I made the first version of this model, which used a combination of pneumatic and mechanical control. I was never completely satisfied with this - according my own particular ideas of purity it would be better for it to be either 100% pneumatic or 100% mechanical. Moreover, it was not completely reliable at shows, since the pneumatic elements tend to dry out and develop leaks over time. The new version is entirely mechanical, and just works from one motor. I have taken inspiration from the recent mechanical Container Transport GBC modules and put the control system out in front where it is visible to an audience. The Tower of Hanoi puzzle is to move a stack of "discs" (here 5 of them, not really disc shaped) from one peg to another, making use of a third peg, with the rules that only one disc can be moved at a time, and a larger disc can never go on top of a smaller one. It required 2^5 - 1 = 31 moves to complete. The solution is ususally described recursively, but here (as in the previous version of the model), it is implemented by a repeating sequence of moves between pegs A and B, B and C, C and A, etc, where in each case the smaller of the two discs on top is the one that moves. The main challenge with such models is to provide a precisley timed control mechanism for the various functions that is reliable and capable of supplying sufficient torque. After trying various options, I did this using a chain drive (similar to the Loop Coaster, but with controlled lowering as well as raising), and two Tomy Armatron mechanisms, one with feedback to advance by a precise amount, and the other using a slipping clutch in combination with end stops.

Yes, I think you are right. So quite a bit of stretching of the rubber for it to bend.

To me Train GBC means GBC that involves trains in some way. GBC is ultimately about variety - time to think outside the hopper! Anyway, no worries - glad you've got your inspiration back!

Maybe this could provide a starting point: https://www.eurobricks.com/forum/index.php?/forums/topic/49988-tomaszajacs-lego-mindstorms-caterpillar/ Although to be honest I'm not sure exactly how the tracks are constructed. The designer gives a remarkably large number of pictures while somehow not giving much clue about this crucial detail! Perhaps some of our reverse engineering experts can help...

1. Pass balls directly from one train to another. E.g. one goes over a bridge and opens a hatch dropping them into another waiting below. 2. A linear GBC module situated on the cars of the train itself, which passes the balls along from the back car to the front car in sequence while the train is in motion. 3. Something along the lines of Akiyuki's "Stuck Ball" module but for trains, where the train can only move if it contains enough balls, and the choice is made mechanically. 4. A module that requires sequencing, e.g. like Container Transport or a Pick and Place device, but the sequencing is done mechanically by a train. It goes continuously around a loop (not carrying any balls itself), and attached gear racks provide power to a sequence of gears situated along the track, powering the various actions. That should keep you busy for starters. I expect to see working prototypes by next week. 😉

Not sure an Sbrick could handle that. Might need a Buwizz... 😆

Wonderful! You might consider putting this on bricksafe (and possibly rebrickable).

That's an amazing peice of work! Perhaps you could consider making instructions, at least for the mechanism parts?

My most ambitious builds just take a very long time. They require lots of tinkering, looking for inspiration, searching for solutions, which I will do on and off for several years. I usually have 2-3 such models on the go, some perhaps in the very early planning stages. E.g. I am currently close (I think) to finishing a purely mechanical (as opposed to partly pneumatic) version of my Tower of Hanoi solver. This has been nearly 5 years in the making, and the earlier version probably took a similar period. Then there are much quicker builds where I have the idea and pretty much build in straight off, or as soon as I can get the parts. There might be 0-3 of these on the go at any time.

Amazing work - thanks for sharing! Sufficient precision and torque are of course the eternal headaches with things like this. I hope you are not discouraged by the difficulties, and continue to work on this awesome project. For a slightly different direction, have you considered trying a binary (rather than decimal) system? Obviously it would not be so convenient as a calculator, but it might help overcoming some of the challenges.

The pin with ball looks interesting. Any ideas what to use it for?

That's a very nice and neat mechanism!

My focus is to make mechanisms or structures that have never been seen before (in Lego or otherwise).

Yep. In particular, the 6t splat would not have many teeth in contact...

My guess is that sadly the screw will not mesh with splat gears as a worm gear. The screw spacing seems to be exactly one stud (the piece is six studs long). But that is not the right spacing for splat gears - see Akiyuki's mangle rack video. It's possible two screws at right angles might be able to turn each other though. The screw angle looks close to 45 degrees.

Indeed! In Police Training Academy. Moreover it looks as though they fit together properly to extend the helix...

Just to address the elephant in the room: PF would probably have been the ideal choice here, but since that is sadly discontinued you are stuck with the slightly less convenient PU, or paying a premium for increasingly rare PF parts.

This looks like an absolutely brilliant job on the Container Transport V3, and of course brilliant work by Akiyuki and Geoff leading up to it. Thanks especially for making free instructions. All the finest AFOL traditions of innovation and collaboration on display here!

This is a fascinating topic, but perhaps it might do better in the "Digital Lego" EB forum?

For the 12-sided cam, one starting point might be to use these at 4 different orientations.

That's amazing! Looking forward to further embellishments of this...

Very nice! For a further challenge, I wonder if it is possible to extend it to bending in three dimensions somehow?

Wow, this is an amazing idea! Is it perhaps an entirely new approach to a limited-slip diff, Lego or otherwise? That would be surprising, but certainly I have never heard of one like this. Does the centrifugal clutch still work if it is not horizontal, or does it automatically lock? (Not necessarily a bad thing, if your vehicle is at some crazy angle!)

While this is certainly true, I don't think it is the fundamental reason why the old double bevels run rough. I think there is a problem with the mould of the tan 20t double bevel specifically. Even with optimal bracing and bushes either side of it to prevent the above problem, it sticks badly, especially when turning a 16t, and most especially a red clutch 16t. Black ones are not as bad, and older LBG and LG hardly have the problem at all. I've no idea why. The new gears do not have this issue.