In the back of my mind I have ideas for various models that I just can't work out. I know what I want to do but haven't worked out how. To me, these models are the really interesting ones. They're especially satisfying when I can finally work out a solution. The latest model to move from this mythical collection into the realm of the solved was the XOR goat. I still have plenty of mechanisms I can't work out though.

Randomness: I like the idea of a paper decision maker. Ask a question, pull the handle and either the Yes character pops up of the No character. For it to work though the chances of each character popping up needs to be randomly 50/50. Making something repeatable in a mechanism is easy, making it random; not so much.

Power: There are a few different models that would benefit from a power source. The wave machine shown here would be neat if it could be made to wave at you across the room, without the need to crank any handles. There are a few possible approaches to powering a model, sand or salt pouring into a waterwheel type machine. Elastic band power – the problem being stopping the model from collapsing under the force of the elastic. Or how about a clock type mechanism with a counter weight and an escapement.
Powered models are the ones I feel that I could solve with a bit of effort and some research time.

Gears. So I actually have a few models that already use gears and they work fine. A single gear to another gear is okay, especially if you are gearing down. But three or more, or gearing up, or even just trying to actually transmit a bit of power and it all goes pear shaped. I think, based on a few experiments, that the main problem lies in the axle rather than the gear itself. A square axle going through a round hole is, by necessity, going to have issues with tolerances. This lets the gear wheel move, and not in a good, round and round, way. If the gear wheel is not accurately fixed on its axis then as soon as you try and do anything interesting with it, the teeth jump. Small holes and a wooden skewer would probably solve the problem but I'm a purist, I don't like idea of using wooden skewers in a paper model so gears stay on my list – for the moment.

There are more of these tricky mechanisms; things that I have yet to solve. For example, I really like the idea of making a butterfly that sits on the end of a thin, delicate branch and opens and closes its wings, the mechanism fiendishly hidden within the branch.

I'm thinking of setting up a section on the websites where I can list these problem mechanisms then tick them off when they are solved.

Any ideas/solutions to any of these would be gratefully received. Back to work on the catapult!