h a l f b a k e r yThe embarrassing drunkard uncle of invention.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
To me, little model steam engines are entertaining but their power is wasted on their tiny flywheels. So, an idea.
Imagine a piston pushing against a spring. Its position would be proportional to the input steam pressure. If there were slide valves exposed when the pressure reached a certain level,
we basically have a transistor.
So, using logic gates from steam pistons we could in theory construct a steam-powered computer. Not steam-engine-powered, but genuinely steam-powered. Naturally the best application of this technology would be a desktop calculator.
The beauty of the arrangement is that steam pressure would directly correspond to electrical voltage, and assuming the cylinders were glass, one could watch it calculating and know that the flow of steam is virtually identical to the flow of charge in the electronic version.
Alternatively, the working fluid could be cold water, and the pressure provided by a stand pipe with hydrostatic pressure. That way, you needn't wait half an hour for the boiler to heat up when you want to do a calculation.
Mechanical Calculator
http://en.wikipedia...chanical_calculator
Or you could just drive one of these with the engine. [MechE, Feb 24 2014]
Steam punk genius
http://www.girlgeniusonline.com/comic.php
This girl could have your calculator up and running in a month. [popbottle, Feb 24 2014]
description of a calculator - with simulator
http://files.righto...ator_simulator.html
a very old, relatively cheap calculator [Loris, Feb 25 2014]
ANITA info
http://www.anita-ca...tml/anita_mk_8.html
[mitxela, Feb 25 2014]
[link]
|
| |
This will give the Borg an erection. |
|
| |
[+] (Though not for that reason). |
|
| |
How complicated a calculator are you going for?
I suspect that the design will be less desktop and
more in the heavy-industry range. |
|
| |
(I don't usually
indicate my vote, but mine gives
this idea a grand total of one. Max's typing hand
doesn't know what his mouse hand is doing.) |
|
| |
I was thinking about the size, and I think the pistons could be made at least as small as ordinary vacuum tubes. |
|
| |
Although you might be able to cram the workings into a desktop machine, it would make more sense to spread them out in a large display case for easier viewing. |
|
| |
Babbage would have loved this. |
|
| |
You need to use hydrogen or helium as your working fluid, for
minimal viscosity and maximum speed of operation. And it needs to
be dry; condensing steam would clog small pipes and cause back-
pressure; either that, or keep the whole thing at 105 C - which might
cause problems with lubrication. |
|
| |
Sadly, this is by no means an innovative idea. Jut type "pneumatic
computer" into google ... |
|
| |
"Wow, your steam calculator's not working, and you
look really dejected." |
|
| |
I'm not so sure if this has been done. There have been many fluidic computers but every one I've found has been analogue. The few 'fluid logic gates' I've seen are demonstrations only; they rely on gravity and the output pressure isn't amplified so they can't be chained. |
|
| |
I suppose using pistons as transistors is a logical step when you're thinking about the electricity-fluid analogy, but if anyone has actually built a working example I would love to see it. |
|
| |
Cool. You could use sublimation of dry ice to avoid the moisture issues and it could operate at room temperature. |
|
| |
Beautiful, really beautiful. I love stuff like this. |
|
| |
While you could design steam-powered transistor-like devices, or you could design standard gates (nand, or, and, and not) to build up in the usual way, it seems like you could probably build it more efficiently with clever contraptions to make more complex basic elements. |
|
| |
No, [8th] you don't need to use dry air as your working fluid. That's already been done, and if you run it at 105 C, again it's just more of the same. The point of a steam engine is that it takes advantage of the phase change between water and steam. Designing to work with the phase change and deal with issues such as condensation and lubrication is part of the challenge. |
|
| |
One possibility might be to have the cyclinders cooled so as soon as the flow of steam is cut off, it condenses and pulls the piston back down, giving push-pull action without relying on gravity or complimentary logic. You would need some tiny float valves that would let the condensed water escape the cylinder without allowing too much steam to escape. |
|
| |
Alternately with complimentary logic, you could make it more like an old steam engine so every time a logic gate changes state, it emits a puff of steam. That might actually look cooler, but doesn't really take advantages of the properties of steam. It also wouldn't be practical in an office where you don't want that much humidity. |
|
| |
Oh, and you don't need to wait half an hour for the boiler to warm up. My wife's clothes steamer gets going pretty quick. Maybe 15 seconds? Depending on your design, you might need a few of these runing in parallel. I think it just heats up the element then it somehow exposes the water to the element in small quantities. So when you're done steaming you're clothes the excess water you dump out is barely warm. If you spent the money to build your steam calculator, I'm sure you could find (or design) a boiler that ramps up even faster. |
|
| |
Just a couple of days ago I was considering a micro-
hydraulic system powered by Mentos and carbonated
water, or some such. |
|
| |
I suppose a couple of AlkaSeltzer might run a pocket
calculator long enough to do your taxes... |
|
| |
Love it - a series of Leibniz "pipes" (hollow Leibniz
Wheels with holes appearing in stepped locations around
the circumference) might provide an alternate method for
converting analogue steam pressure into a more digital
function - you lose the analogue electronic/fluidic
analogy in one sense, but I don't think it takes away
all that much. |
|
| |
Suppose we wanted to create a very basic calculator. I've linked to a basic one, which does the four basic operations. I would consider that mission accomplished. If we could do that, we could also do more clever things and get a rather slower low-accuracy scientific calculator for the same resource. |
|
| |
So what would that need? Apparently it had a 320x11bit prom. That would be managable. I've not found a reference to the transistor count of the processor - a TMS0800 series. The sinclair scientific used a TMS0805.
My guess (based on an article listing chip transistor counts vs time) is of the order of 5000 transistors. |
|
| |
Oh, I should point out that a steam boiler might be different in operation to a clothes steamer. My father has as a hobby set up a miniature railway in his back garden. The (IIRC 1/6th scale) steam engine is indeed a thing of beauty, but it takes well over half an hour to get up to operating temperature. On a good day. |
|
| |
I really like that simulator. |
|
| |
One of the calculators I had in mind was the ANITA (link). The logic was achieved with less than 200 tubes and some number of relays. I think the reason they were so small (relatively) is that they used decimal logic. In a way the dekatron counters are analogous to Liebniz wheels. |
|
| |
Regarding steam-up time, I actually visited the Science Museum today, and watched them start up the steam engine. It's interesting to note that the crowd was biggest just before it ran - I think the tension and mystery of the progressive leaks and the first twitch of the flywheel all add to the experience. |
|
| |
The quick-starting type of boiler is called a Flash boiler I think, as used on commercial production steam motor-cars. |
|
| |
Ah, yes I did mean to say that that was an upper bound calculation assuming no optimisations of logic at the valve level. |
|
| |
Decatrons are beautiful things, but I'm not convinced they could be easily replicated with steam - and particularly not at the same scale. |
|
| |
So far we've been assuming a decimal output. And both the ANITA and the early electronic calculators work in base ten internally, in some sense. It seems to me that it would be acceptable to work in binary exclusively. That would give quite a big saving in the logic required. If someone were to find that offensive they could implement decimal input and display devices as separate modules. |
|
| |