>> Think about how a transistor works: You have one circuit that controls
>> another. So how hard would it be to rig up a valve where pressure from
>> one pipe moves the valve allowing (or blocking) water from flowing
>> through a seperate circuit? In principle it ought to be pretty trivial.
>> (Of course, making a valve that actually works well in practise probably
>> requires far more equipment than I personally have...)
> 
> Right OK, just like all of the analogies. The gates would be controlled 
> by current, then. I suppose you could control them by voltage (pressure) 
> as well, but the valve would need to be easily actuated by pressure. 
> Hmm, the analogy holds well.

The analogy is pretty perfect. Electrical current = water current. 
(Litres per second rather than Columbs per second.) Electrical potential 
(i.e., voltage) = water pressure. Resistance = ...well, resistence.

Wikipedia even has a name for it:

http://en.wikipedia.org/wiki/Hydraulic_analogy

An electric circuit is just like a water circuit, but pushing water 
around instead of electrons. All the same things like flow-rate, 
density, pressure, etc all make sense.

I envisiged pressure-actuated valves. (I.e., pressure on the input 
channel slides a valve, revealing an outlet. A spring slides the valve 
back again when the input pressure drops. The slilding of the valve 
opens or closes a channel for another water circuit.) If you wanted 
current-actuated, you'd need a turbine or something to measure flow-rate 
rather than pressure.

>> The problem is going to be that once you have more than a few of these
>> linked together, effects like gravity and insertia become significant.
>> These don't affect electronics, for some reason...
> 
> Well, water molecules are several orders of magnitude larger and more 
> massive than electrons... so it makes sense. Though, I suppose it could 
> be said a coil imparts a certain inertia to the motion of electrons.
> 
> I realize I'm stretching the analogy with that, but you'll see the point.

Even so, you would have thought starting and stopping the motion of 
several hundred trillion electrons would take more energy than just 
keeping them flowing.

>> Maybe that's what I did wrong... I was expecting an open TTL input to
>> float low. Anyway, I don't think I shall go down the FPGA route. (!)
> 
> Yep. That got me when I fist started playing with TTL gates.

Hmm, interesting.

> Here's and FPGA with 3000 cells for around US$18 ... Expensive in terms 
> of a single IC, but not too terribly bad.
> 
> However they can also run as high as US$6,485.79 (!!) Oh, and good luck 
> soldering that one to a board at home (Though, I've heard you can 
> actually attach a BGA to a circuit board using a toaster oven, but very 
> hard to verify that no pins are bridged or poorly connected.

I was looking at the price of a "kit", which contains not only the FPGA 
chip itself, but also the [expensive] gizmo for programming it. The chip 
also comes ready-mounted on a board with a USB interface and there's 
driver software so you can control the thing from your PC. But yes, the 
cost is far, far more than the piffling price of one little chip. (And 
then there's the version of the simulator software which can actually 
*program* the physical chip, not just simulate what it would do...)

>> Another thing I thought about was a lego-style kit where you have lumps
>> of plastic in the shape of logic gates, with nice connectors for the
>> inputs and outputs, and LEDs in each input and output to indicate which
>> logic state it's at. The trick, of course, is power routing. ;-)
> 
> Meh, that should be too hard to do. Either you have a substrate that you 
> drop the blocks on, or the power routes through additional pins.

Yeah, the substrate is probably the way to go. It draws attention from 
the ubiquitous power lines and focuses it on the routine of the signal 
lines.

> I've 
> seen a lot of electronics kits for kids that use lego-like blocks to 
> connect components together, rather than the old spring terminals that 
> my kit had when I was a kid.

Hell, I've still *got* the spring-terminal kit. ;-)

Yes, today you could make some cool stuff. Trouble is, kids today don't 
have the patience for it, it seems. Why spend hours wiring up a circuit 
so that a few LEDs twinkle when you can sit in front of a PS3?

Post a reply to this message