The Mines at Ria

[home]
[about]
[morris]
[pockets]
[email]

QuAD was neat as a project to start with, but when I finished the ball sorter I realised that accuracy of position would defeat me in the long run, if I wanted to do any more complex programming. A beam-rider (technically a supported monobeam system) will cure those problems. This will allow me to sort stuff more accurately, without having to use the light sensor for line-following.

The working name Ria is taken from a monorail which Lartigue built at Ria and which I've modelled a couple of times in the past. On the right is a picture of the railway at Ria, from the 1886 Lartigue booklet. I have a simulation on my laptop which works a bit like a sliding blocks puzzle and which has given me a few ideas for neat tricks to play with the programming, so Ria it is.

I documented the project as I went along, mistakes and all. Learning from your own mistakes is good, but learning from someone else's is better. I'm taking my inspiration from the Victorian era monorails by Fell and Lartigue, but there's a serious problem with building one exactly like either of those. You can see that the Lartigue system uses equal weight on both sides of the beam, hanging down for stability. The RCX with batteries if too heavy to do that in any meaningful way, and I haven't had it long enough to start butchering it.

Note to parents: You may feel it was unwise to publicise a model that needs a six-foot (longer is better) piece of wood to run along. In my defence, at no time did I say to nail the wood to the furniture: if you build the model correctly, you should be able to just rest it on two things of roughly the same height. If you or your offspring have interests in programming, the Ria software works with Joe Nagata's train, which doesn't even need the six feet of pine at head height.

In truth, I did have enough parts to make the guide wheels and in the end it was the only way to make the thing work accurately. You can see the result on the light sensor software page. If I'd followed the original design mechanics this wouldn't have taken so long.

The Fell system uses complex guide-wheels below the level of the rails. I thought I didn't have enough pieces to duplicate these, but was heartened by a supply of fairly straight pine which neatly fitted between a pair of black beams when the latter were set up to support the RCX . This should have kept the centre of gravity low enough for the thing to work with the beams as simple guides. Now read on...

My design constraints were:

the RCX had to be easy to access, unlike on QuAD;
driven wheels to be as small as I could get for accuracy in positioning;
larger idler wheels at the other end because I only had two of the small ones;
motors to fit with the special motor clips, unlike on QuAD;

That's my proof of concept. Click on it for a bigger picture. Here's the order in which I started to do the project:

I then aimed to

add a more complex hoist, one of a number of drawings from real systems, which I've never been able to try out;
write (port from my little sim) some more complex shunting/sorting mechanisms

Now, project plans never survive contact with reality, so the stuff at the end never got done, but I'm looking at the development of a more complex system to solve a different type of problem. This is a cross-fertilisation with a model railway project.

[home] [pockets] [mindstorms] [tour]