• Please review our updated Terms and Rules here

Homebrew Analog Computer?

Aaron Teeling said:
Does anyone know of anyone who has built or is building a analog computer?
Click to expand...

Hi,

Yes. We built those in college. They are basically a collection of op-amps and passive devices used for solving differential equations. The military used to use them for all sorts of applications like ballistic solutions, mixing equations, etc

They work well but are not terribly exciting, IMO. I am sure you could build some based if you got a book on analog computing from the library and ordered some precision op-amps.

Best of luck!

Andrew Lynch

PS, ordinary passive components generally aren't going to work since the precise values are necessary for accurate solutions. Generally, you'll need the high precision 1% tolerance or better components.
 
Aaron Teeling said:
Does anyone know of anyone who has built or is building a analog computer?
Click to expand...

Hi
There are a number of different types of analog computers.
As was mentioned, many think of the op-amp type as the only
type. There were many of what I'd call the slide ruler type.
These were quite common from places like radio shack.
These were just three pots, a battery and a meter.
I have a couple of op-amp type analog computers but
I've not had time to play with them much. One is the
EC-1 by HeathKit.
Many problems in physics can be done with analog computers.
The US had some of the best fire control computers that
used mechanical analog computers, during WW2.
Early pong games were done with analog computing.
Dwight
 
Analog computers are not like a programmable digital computer... imagine it as a tool to solve a single problem.

A simple example:

Example: a+b=c

Input A = 5v dc
Input B = 2v dc
Answer: 7v dc (approximate read by meter)

The neat thing about analog computers is that the answer is produced very quickly as there is no program loop.

You could also build an analog computer mechanically rather than electronically.

There is a good book about this (scanned to pdf) which I will share when I find it later today.
 
How many have constructed the more advanced mechanical analog computers? e.g. ball-and-disc integrators. Pneumatic computation was fairly common in industry years ago. For example, if you want to measure the flow rate of a liquid or gas, you can use a differential-pressure method, which uses Bernoulli's equation. To get the flow rate, you need to extract the square root of the pressure differential. There are mechanical ways to do that with pneumatics.
 
The most common thing to do with analog computers is to integrate. This allows modeling of physical systems to see how they will respond in real life.
While multiplying by a constant is easy with an analog computer, multiplying to variable values is a little more difficult. Diodes are relatively handy as multipliers. The current is exponential to the voltage. This means that one can add the current of two diodes with input voltages, across them, and then take that current to match a voltage across another diode, making a linear voltage again.
Most simple systems can be emulated without xy multipliers but to make a great attractor, you need a XY multipliers.
Older analog computers did often had mechanical multipliers. Such non-linear operations were often handled by contoured surfaces, where one value was a rotation and another was a movement along the length of the shaft.
Dwight
 
Well the book I have (in print) is called "basics of analog computers" which was once available online in google books which is how I found it but has since been removed.

Amazon has it, but it's not cheap.

It is probably the best book for beginners on the subject as it gives a mechanical example for each circuit.
 
Most physical systems are often described as differential equations. It is good to know how to translate differential equations to integral equations. Although, one can make differentiators with analog computers, they are more susceptible to noise and small DC offsets.
Don't ask me to do any, as it has been too many years since I did this my self.
Dwight
 
Actually, Dwight, we use differentiation in lots of applications in analog and digital electonics in the guise of RC networks:

rc-rc34.gif


But, as you said, the problem is doing it with precision.
 
One of the annoying problems of analog computers is the two different types of offset. One can have input voltage offset and input current offset. They don't often happen at the same point. It is interesting to see how the control computer used in the V2 ( A4 ) rockets dealt with offsets using simple troide tube amplifiers.
See: https://www.cdvandt.org/Hoelzer V4.pdf sorry in German!
The schematics are still worth looking at.
I manually translated it once. Many would think poorly of Helmut Hoelzer because of his contribution to such a nasty weapon. He was just an engineer working on a problem. If he'd not been such a determined engineer, the V2 might never have been successful. In fact ballistic rockets in general would likely have not been created for years into the future. We'd not have satellites for communications or been able to send men to the moon. We'd have gotten there eventually but it would have been delayed quite a while. The military success of the V2 was what made the USA and USSR interested in the technology. Like most things, it can be used to benefit or destroy.
It is interesting that his control system made the rocket possible but it seems few understood the genius in his design.
He'd worked on the design in secret, even from his superiors. If he'd got caught at the early stages, thousands of lives would have been saved at the time, by stopping him from wasting time and material.
Things are like that sometimes.
Dwight
 
Chuck(G) said:
Given the intended purpose of the V2, I'd hesitate at labeling it a "success", Dwight.
Click to expand...

I'm trying to think of another word. Its use was clearly bad. It didn't really do what it was intended to do and that was to demoralize the British. It didn't do that, if anything it did just the opposite. It likely cost more to build than an equivalent air raid.
I only meant that the analog computer made it possible to control what others thought couldn't be done.
Dwight
 
I've got to concur here, Dwight... under the circumstances, the V2 programme seems like one of the least bad things that could have happened, even though its basic purpose was a bad one. It is easy to imagine the war being longer and worse if no resources had been spent on the V2, or if the technological ingenuity that went into it had been applied to more effective weapon concepts. Most of the tech in it ended up benefiting the postwar powers much more than its original builders, clearly a good thing. I find it easier to appreciate the V2 as an impressive achievement - several advanced technologies, for the time, integrated to make an entirely new kind of thing - precisely because it wasn't militarily or terroristically particularly effective.

(I say this even though I'm about 400 yards from the nearest place hit by one, and you can infer the damage it did even now from the places where older buildings abruptly stop. Scary stuff from our perspective, it must have made a hell of a big hole, but it cost a lot to make that hole and didn't alter the course of the war much if at all)

Anyway, analogue computers. There was a really good demonstration of them at the National Museum of Computing at Bletchley - I really hope they make it through the current fuss, I've donated - and I would jump at the chance to refurbish an old one, but I really don't know what I'd do with it. I dimly remember building single integrators and maybe differentiators? from op amps at A level (high school?) but somehow never needed one since. I'll keep my eye out for an excuse! Thinking out loud here, it seems kind of silly to build an analogue computation and read the result on a digital oscilloscope :) but it might be more fun to use one to control something physical.
 
Recall, also that industry has used both pneumatic and hydraulic methods for computation. They have the advantage of ruggedness and operating over a wide temperature range. The disadvantage, of course, is that you use plumbing rather than wire.
 
Air tubing is quite fun. Some of the "reproducing" player pianos (that have dynamics coded on the punched music rolls, rather than just the notes) have, ok not computation, but at least a bit of logic implemented entirely with tubing and wood and rubberised cloth. I'm sure I've seen a 4-bit DAC in one - 16 levels from 4 channels - made just from bellows, that was pretty cool. To do anything more complex with that approach would probably be physically enormous, perhaps fine in a factory, but hard to fit in at home...
 
You might want to look into industrial control instrumentation for pneumatics. The standard there was generally 3-15 psi (i.e. all positive pressure). Lots of 1/4" plated copper tubing. Some very ingenious solutions, such as the square-root extractor that I mentioned (really works on a cosine over a limited range) can be performed.
 
The analog fire control computer on an WWII Iowa class battleship makes your fine Swiss watch look like something in the checkout lane at the supermarket. If you have the opportunity to tour one of those magnificent ships, you'll want to make a point of checking it out.

Also, when I was a young A4D-5 Skyhawk plane captain, the LABS (low altitude bombing system) was part of the pre-flight inspection, at least in my squadron. Mostly mechanical with some relays and what all, the pilot could do the 'over the shoulder' loft maneuver and the LABS would handle the release of the shape. Big boom to follow shortly thereafter.
 
To kind of go along with Chucks mentions, they even had pneumatic flip flops and gates based on the Bernoulli effects. One might think that it had some moving parts, similar to a automatic transmission but these logic devices had no moving parts, they just had a small air stream.
Come to mention it, the automatic transmission is one of the most complicated fluid computers one could find. Most people don't even think about the complexity of making the car change gears without stalling the engine or slamming peoples heads against the head rest.
Dwight
 
Yup, industrial pneumatics are usually static affairs. But fluidics was very hot back in the 60s--and has been quietly applied to quite a number of applications, including control of large ship bow-thrusters. And, of course, numerous applications in aerospace.

Then there's MONIAC ca. 1949 or the older Russian Water integrator.

Or the digital FLODAC ca. 1964

I'm a bit surprised that fluidics hasn't enjoyed a popular renaissance, what with the advent of affordable 3D printers. Earlier systems required precision machining.
 
You must log in or register to reply here.
Back
Top