Analog Computers: Digital vs. Theoretical Possibilities

[ISamson]

There are two types of 'reading'. Digital and analog. There are digital computers, which function with 'on' and 'off' data. But can it possible to construct an analog computer? Can it be theoretically possible?

 

[fresh_42]

There are two types of 'reading'. Digital and analog. There are digital computers, which function with 'on' and 'off' data. But can it possible to construct an analog computer? Can it be theoretically possible?

Not only theoretically: https://en.wikipedia.org/wiki/Analog_computer
In the end it depends on what you allow to be called a computer. There are even more examples on other Wikipedia language pages, so I suppose those examples can also be found in English on some webpages.

 

[Vanadium 50]

https://en.wikipedia.org/wiki/Analog_computer

You've been here long enough that you should know you have to put in some effort.

 

[lekh2003]

Do you think this is relevant: https://en.wikipedia.org/wiki/Antikythera_mechanism

It is probably the world's first scientific calculator discovered in a shipwreck in 1901. Apparently it was built by the Ancient Greeks. Analog computers have been around a long time.

 

[ISamson]

Do you think this is relevant: https://en.wikipedia.org/wiki/Antikythera_mechanism

Greeks had computers. Cool!
Well, a mechanism to be precise...

 

[lekh2003]

Greeks had computers. Cool!
Well, a mechanism to be precise...

But still a computer, again, anything can be a computer. I like to think of a computer as a man-made machine with automated computational abilities. So this would include something like the Antikythera machine, but would exclude an abacus.

 

[ISamson]

Why are analog computres utilised nowadays, how are they useful or more convenient in some ways than digital, normal, computers?
If such analog computers are used, then how are they better than digital?
I don't understand this.

 

[lekh2003]

Why are analog computres utilised nowadays, how are they useful or more convenient in some ways than digital, normal, computers?
If such analog computers are used, then how are they better than digital?
I don't understand this.

Well, for starters they won't require electricity. I don't think the Ancient Greeks had anything like electricity so how in the world would they make a digital computer. In today's world, I can't think of any advantage besides not needing electricity. Analog computers are simply too big and bulky and do too little. Simple multiplication machines are the sizes of rooms (you could make it smaller and more efficient, but why?).

 

[ISamson]

But still a computer, again, anything can be a computer. I like to think of a computer as a man-made machine with automated computational abilities.

A computer is a device that can be instructed to carry out arbitrary sequences of arithmetic or logical operations automatically. The ability of computers to follow generalized sets of operations, called programs, enables them to perform an extremely wide range of tasks.

https://en.wikipedia.org/wiki/Computer

I think it is more accurate to think of it as a device capable of operating logic in computationally and mathenatically logical ways.

 

[lekh2003]

I think it is more accurate to think of it as a device capable of operating logic in computationally and mathenatically logical ways.

Well that's the definition of automated computational abilities.

 

[ISamson]

I can't think of any advantage besides not needing electricity.

How does a computer not need electricity? How on earth?

 

[lekh2003]

How does a computer not need electricity? How on earth?

It's an analog computer, it can use mechanical energy and convert that into motion to run its parts. There are a plethora of methods. Take a look here: https://en.wikipedia.org/wiki/Analog_computer#Mechanical_analog_computers

You can find this stuff online, my friend.

 

[fresh_42]

A slide rule is certainly a computer, because you can do computations with it. And it is without electricity as well as analog.

 

[ISamson]

In recent years, analog computers have proven to be much more efficient at simulating biological systems than digital computers. But existing analog computers have to be programmed by hand, a complex process that would be prohibitively time consuming for large-scale simulations.

The work could help pave the way to highly efficient, highly accurate analog simulations of entire organs, if not organisms.

http://news.mit.edu/2016/analog-computing-organs-organisms-0620

 

[lekh2003]

http://news.mit.edu/2016/analog-computing-organs-organisms-0620

Yeah, but then again, they just said how difficult it is. You have to mechanically fix everything, not as easy as coding. To create a fully functioning organism, the analog computer would take decades to build and would occupy a city.

 

[ISamson]

You can find this stuff online, my friend.

If you are talking about mechanical, not electrical, or silicon computing, then you are totally right!

 

[Rive]

Why are analog computres utilised nowadays, how are they useful or more convenient in some ways than digital, normal, computers?
If such analog computers are used, then how are they better than digital?
I don't understand this.

Nowadays this art can be considered as a dead end. There are still some examples when analog functions are implemented with the knowledge and parts originated from analog computers, but that's all. Microcontrollers and AD-DA converters are so cheap, that most cases it's just faster, cheaper and more flexible to do everything digital.

In theory, analog computers were way more sturdy in some environments, and such things like a random high energy particle was just a source of some additional error instead of a deathblow (like it is for a common digital circuit).

 

[Borek]

Analog computer can be easily an electrical one. Nothing unusual to use an electrical circuit to simulate a process and read the results from a voltmeter.

My Mom in front of an analog electrical grid simulator, early sixties:

 

[Svein]

Analog computers...

Of course you can add and subtract voltages with op-amps. You can also take logs and do exponentials. There are multiplying circuits (http://www.analog.com/media/en/training-seminars/tutorials/MT-079.pdf) as well.

So what else do you want to do?

 

[ISamson]

Analog computer can be easily an electrical one. Nothing unusual to use an electrical circuit to simulate a process and read the results from a voltmeter.

Voltimeter? That easy?
Do you have a link?

 

[ISamson]

Of course you can add and subtract voltages with op-amps. You can also take logs and do exponentials. There are multiplying circuits (http://www.analog.com/media/en/training-seminars/tutorials/MT-079.pdf) as well.

Thank you. Extremely good link!
But a bit complex...

 

[Rive]

Voltimeter? That easy?

It's just unusual but not really different than what you get from a common calculator...

 

[Borek]

It's just unusual but not really different than what you get from a common calculator...

Completely different. Do you understand the difference between analog and digital?

 

[Rive]

Completely different. Do you understand the difference between analog and digital?

Yes. In this case: around 1% accuracy vs. 8 digit accuracy.

(C'mon, both are simple, one dimensional readings...)

 

[phinds]

Greeks had computers. Cool!
Well, a mechanism to be precise...

Keep in mind that the word "computer" was first applied to PEOPLE, not machines.

 

[donpacino]

Even in the modern day the "flight control computers" on some aircraft are all analog (granted the products in design now are not).

 

[donpacino]

http://www.popularmechanics.com/mil...r-were-a-mid-century-engineering-masterpiece/
So the B-29 had one of the first advanced electronic computing systems that were not the size of a room during ww2. Read the above article!

TLDR: The B-29 was pressurized so to utilize less crew and allow the crew to properly aim weapons, the machine gun turrets were controlled via an analog computer and remote control. They even had logic to prevent the aircraft hitting itself.

 

[donpacino]

Yes. In this case: around 1% accuracy vs. 8 digit accuracy.

(C'mon, both are simple, one dimensional readings...)

There is no such thing as a purely digital voltmeter. A digital voltmeter is an analog voltmeter with a digital interface to the outside world along with possibly (depending on the model) some process and function handling in the digital domain as well.

His question was not regarding accuracy. It was do you understand the differences beneath the hood.

 

[CWatters]

Here is a water based analogue computer from 1949 running a simulation of the UK economy...

 

[Rive]

There is no such thing as a purely digital voltmeter.

Since the topic is about (analog) computing, the analog/digital measurement is completely off. What matters is, the type of output, what's a single number in this case: the same what a common (digital) calculator would provide. To mix in the type of output is kind of like setting up categories on computer performance according to display type (CRT, LCD).
That was behind my slightly sarcastic answer about accuracy.