Autonetics D-17

>Autonetics D-17 guidance computer from a Minuteman-I missile.

How do these circuits even work? It's huge.

Can Sup Forums explain it

Other urls found in this thread:

allaboutcircuits.com/textbook/semiconductors/chpt-9/computational-circuits/
twitter.com/NSFWRedditVideo

well, not only does the missile need to home in on a specific set of GPS coordinates, but it also forms a network with the other missiles launched and can recognize targets and set target priorities, avoid antimissile countermeasures designed to stop it, and makes sure that the high priority targets are destroyed first and that each missile is covering a different target
if a missile gets destroyed or is unsuccessful, one of the remaining missiles will divert its course to destroy the open target

It's big because it's a mix of redundant systems and being built in 1962. It was probably pretty tough to build a guidance system before the advent of microprocessors.

air bearings aligning a receiver to whatever tower system was providing it with locational data, gyroscopes read the missiles locational data and compare it, the circuits automatically send out out the orders for changing course

the circuits are basically an extremely basic system thats just taking input from a gyroscope, comparing it to what data its receiving, and adjusting the rocket to get as close to identical streams of data as possible

and its big because its old, pcb lithography at the time was not too fancy

I don't even understand what I'm looking at. When I think about electronics I'm reminded of this... thing? All these densely packed data paths.

The missile's system looks like it was drawn by a child. It's all irregular and stuff. And what are all those components doing? How do they make a computer? When I look at that, I think of voltages rather than bits.

Nah in 1962 they couldn't do advanced PCB like that. They were probably still drawing the PCBs by hand. lol

I mean... any voltage above a certain point is 1 and under it is 0 so that is exactly what it is. Just think of the innards of a Microprocessor, how many transistors and resistors and pathways there are in there and then rebuild it with off the shelf parts.

Single sided hand-traced PCBs with large through-hole components implementing a very complex system. So it is to be expected.
If you have any idea how difficult it is to manually build analog circuits to implement complex logic systems then you will understand why it is so big.
allaboutcircuits.com/textbook/semiconductors/chpt-9/computational-circuits/ for a bit of explanation on how it can be done. I'm not sure if this circuit uses much logic (which can be converted to from analog using schmitt triggers, for example), but probably a bit at least.

Yeah what I mean is I'm used to thinking about electronics in terms of high-level components like processors, chips and other black-boxes, and these paths as the connections between them. Like a network.

When I look at a pile of PCBs like that with a pile of what looks like resistors on top, I have no idea what is going on, as in, I don't understand how does that translate into "guidance" functionality.

Even a bunch of discrete component radios intimidate me, I look at it and it seems like pure magic

shiiieeet i didnt know missiles could do that

Imagine the simple hello world program. A simple high level statement. This is what your describing with processors.

Now imagine windows written in assembly. This is your motherboard built with nothing but off the shelf resistors, transistors and capacitors. That is basically what is pictured.

I mean if you spent several years studying the assembly code you could figure out what each function did in the grand scheme just as you could derive the function of each board.

Yeah m8 US military shits out some of the most impressive tech I've ever seen even in old times. I spent all day reading about military satellites.

Soviet superscience is awesome too.

So those ARE resistors, transistors and shit? Those everyday components that go into electrical circuits people make in universities to learn?

Holy shit man. How in the hell could someone make a "guidance system" out of this? It boggles the mind. It's awesome

Yea, it's all they had.

Years of work. How were the 15million lines of the Linux kernel written? One at a time. Each process of guidance can be broken down into discrete processes and then done by everyday components.

It's still done by "everyday components" they've just been miniaturized beyond belief into processors.

I mean even 's description sounds too algorithmic for this

How do you make these components:

1. read output of on-board sensors and communications
2. compare them
3. Somehow apply feedback to the system to get them to converge

The Nazis had mechanical counters on their V-1 bombs that had a little propeller on the front, at a predetermined number of revolutions the counter would cut the engine fuel and jam the control surfaces down. All the Nazis had to do was point them in the right direction, figure out the number of revolutions it would take between the launch and target and they were set.

missile targeting is relativity simple if you know where the target it

>How in the hell could someone make a "guidance system" out of this?
This is a small part of the work that had to happen before you can simply type
>print "Hello World"
on your stupid fucking interpreter. If you think your cool python program you wrote is special, consider the work that had to go in to get to that point, all without the aid of any computer tools.
Be grateful, or git gud and understand it

Also please do your own research before making a new thread on Sup Forums

Yeah sure but it's so easy to forget the billion transistors and god knows what else goes into these chips when I have an instruction set that can manipulate all these electronics in a rather abstract way

So what you're saying is this is kind of a hardcoded program? It's like they took out a high level concept of the problem, figured out the "assembly code" it maps to, figured out the components that code would drive and then just hooked them up directly?

I created this thread because I was awestruck and wanted to make sense of this after being unable to even after researching all day

Not because I wanted to make fun of it

>So what you're saying is this is kind of a hardcoded program? It's like they took out a high level concept of the problem, figured out the "assembly code" it maps to, figured out the components that code would drive and then just hooked them up directly?

yea, kinda. They didn't have the concept of assembly code in the day though. They more likely said stuff like "When this input is higher than this other input we need out output a high signal to drive this other circuit which does X other function." and that was their analogous
CMP A1, B1
JMP 04

That is awesome

To think wizards like that once walked this earth and today everybody makes webapps

>It's like they took out a high level concept of the problem, figured out the "assembly code" it maps to, figured out the components that code would drive and then just hooked them up directly?
Kind of the right concept, but you're thinking of it backwards.
Assembly code (and processors) didn't exist for the longest time. Up until that point, all circuits - digital and analogue - had a fixed (usually one) number of functions, and yes it was 'hardwired' in this way. Look up processor architecture if you want to see what makes assembly code (technically the machine code - zeroes and ones that assembly maps to) execution possible.

>I created this thread because I was awestruck and wanted to make sense of this
Research these topic:
Operational Amplifiers
Analog computational circuits ( allaboutcircuits.com/textbook/semiconductors/chpt-9/computational-circuits/ )
Transistors in Digital Logic
Digital Design
Processor Architecture

We truly stand on the shoulders or giants. I love studying the old way of doing things because it reminds me how far the ground really is.

you realize those things are meant to kill people right?

The best things were originally built to do that. There's a fair amount of innovation in war time.

This is from like the 1960s, right? Assembly code existed back then, but software assemblers did not. Assembly was something you wrote on paper while planning a program, then converted into the binary opcodes which you would load into the computer.

>you're thinking of it backwards

Yeah it's the only way I can understand the thought process employed by the people who designed that system

I was watching some MIT class videos the other day, they explained a processor bottom up, making it look easy, I mean hey there's lots of electrical components that you can selectively charge and discharge based on what you want to do. But then I see components used in a stand-alone way and I just don't understand what they're doing

I like reading about military technology, especially from the cold war era. It's one of the most motivating things I ever found in my life. These guys somehow invented the most incredible shit and all they had were these fundamental circuit components

Is it surprising that the most amazing technology comes out of the military? The internet came out of the military.

'62. You may be right, I am neither an encyclopedia nor do I consult one when I post on Sup Forums.

It's TTL, much more resistant and less prior to errors then microchips.

that looks really old, definitely before GPS. Their is no way it does half of that shit.

It probably just controls the control surfaces with some kind of feedback control to keep it straight and cancel out wind/turbulence/whatever. I bet it has circuitry for keeping track of time and detecting combinations of failures.

Once you have it going in a straight line, The only other variable is how far you want it to go. So all you would have to do is point it at Moscow and let it rip.

What

That is nothing compared to THIS computer though.

until one wire shorts out on something, or the tiniest EMP goes off next to it

ITT ANALOG COMPUTERS MY DICK

Take one of these.
Now put 1.4 billion of them together (with accompanying resistors and whatnot).
That's an i7 quadcore.
Yes, in the beginning, they had to be put together by hand. Older processors, not i7's.

oops

Where do you study the old way of doing things? I could use some enlightenment

you know how now transistors are in the silicon? back then they were individually soldered on to the board. otherwise same principal

>made in 1960
>dsub ports

>thinkpad T450s in 2015
>still has dsub ports

Never know when you'll have to reprogram a minuteman missile.

Even though Turing complete digital computers existed in 1960s you probably couldn't fit one in a minuteman and run a program to guide it.

The internet is your best resource. Look for sites detailing analog circuitry then look at the footnotes for a cool book if you're that interested.

What's wrong with those ports?

well, judging by how many fucking discrete components you would need for an actual digital processor, this is probably analog.

watched an excellent history channel documentary called race for the superbomb other day, highly recommend if you find it. Much better than what they usually dish out

How technical is it?

I hardly watch any documentaries, the few I have watched glossed over all the fun engineering stuff. They were showing me the great stuff, but they didn't care about how great engineers made it. In the end, they amounted to nothing but "this amazing thing exists/has happened."

The most fun part for me is when knowledgeable people start talking about how they achieved their objectives and why they did it the way they did and not in some other manner. It's like reading a good hacker's source code and then talking to him about it.