Ho hum. Experimenting with larger pits to manipulate minecarts, the most interesting but so far also remarkably useless was the "pentomino" - ramped pits arranged like this:
▼.....║
¢¢▼...╣║=
.▼.....║
Pits with track ramps as seen from above, two with hatches over them, ramp orientation to the right.
With both hatches closed, the cart goes in a simple loop, but if the left hatch is open, things get strange. The cart accelerates past derail speed quite quickly, so i added derail track ends with high-friction track stops to keep the cart in the loop. While this succeeds, the cart goes on some convoluted spaghetti course with a ridiculously long period. It gets even crazier when both hatches are open.
The cart passes through nearly every bit of track visible here, in a crazy pattern of ever-changing loops. The actual repeating period seems to include twenty passages through the pits, possibly more.
For now, it's just a conversation piece. I've as yet no idea what use to make of it, although it'd make a nice screensaver.
Other than that, i couldn't come up with a more compact powerless memory cell. I made a working model, but it's nothing new. I think there might be an alternative design with a single straight pit possible, but it promises to be a bit fiddly to build and less versatile.
In other news, the duchess is still busy populating the fort and adopting cats. The first mayoral elections were held and i promptly overrode the result in favour of the old expedition leader. Still two years until the first children grow up.
EDIT: yeah, the memory cell works:
Above the core memory - a single straight double-ramp pit. The cart is either in the northwestern or the southeastern loop and will keep cycling indefinitely, in a stable orbit fully restricted to that loop. The difficulty lies in extracting the cart from the loop when clearing the memory resp. feeding the cart into the loop when "writing" to memory. To achieve that, at the bottom the complete memory cell: the two zinc (teal) hatches are operated by a single signal, clearing memory when the signal goes off, activating the memory when the signal goes on: no matter which way round the cart cycles, it'll pass across the hatch cover over the memory pit to the west, coming to rest by bumping against the wall and sits on top of the northern hatch cover over the western pit. The southern hatch cover is regulated by the data input. When the clear/write hatches open again, the cart falls into the pit, leaves it either to the south or north, depending on the input received, and files either into the northwestern loop (through the southern branch) or the southeastern loop (through the northern branch) and remains in that state regardless of changing data input, until the memory is cleared again.
The clear/load signals must be potentially stable over a quite long time, so work best when run from a lever and not so well when activated by pressure plate. The cell distinguishes between "signal positive", "signal negative" and "memory not loaded". It takes a signal and "remembers" its state indefinitely, ignoring later signal switches.
Damn, i think i'll have to make another computing thread soon, i guess.
Author
Topic: What's going on in your fort? (Read 6097144 times)