CPU out of logic gates

[DrenDran]

DrenDran submitted a new Showcase Item:

CPU out of logic gates

Inspired by a few YouTube videos showing projects of the same nature, I decided to create my own breadboard computer. Basically, it stores the program to be run as well as that data it's using on a RAM chip. Using somewhere between 8 and 16 instructions I'll be able to make it do basic things like add numbers, subtract numbers, store data in ram, add data to ram, compare values, jump to other parts of the code, etc.

It's far from done, but I've gotten the program counter, command registers, clock, and memory unit wired so far.

Attached is a photo but I've also created a YouTube video:

View attachment 33186

Read more about this showcase item here...

 

[Bluejets]

In the late 70's one could build your own robot from discrete chips etc.

http://cyberneticzoo.com/cybernetic...programming-robot-david-l-heiserman-american/

Problem was at the time parts were difficult to source and cost was at a premium.

 

[KeepItSimpleStupid]

Back in the 80's , three of us decided to breadboard a very simple computer using SSI (5 Volt 74xx TTL) integrated circuits for a college course with a lab. This was about the time the 4004 and 8008 processors came out. One of us took the memory, I did the program counter/branching and another did the data structures.

I learned the "hard way" why the program counter has to point to the NEXT instruction.

The instruction set was 16 bits x 16 words and what we would call microcoded. It could sort a list of 4 bit numbers on ascending or descending order depending on the microcode.


Memory is cloudy. I think we had a bit that did "branch", branch on less (A<B). Branch on greater (A>B), swap, load A and load B and probably store A and store B. I just can't remember. That's almost 37 years ago.

I think the clock may have been a 555 timer. It wasn't fast and it wasn't intended to be.

The PDP-11 mini-computer schematics were an inspiration. If I remember right, instructions for that machine were microcoded into a 128 bit wide microcode.

I should still have my lab notes.

==

In the same time frame, two of us wrote an OS and machine simulator for a Computer Science class. A large feat for two people. Everyone else had 4 or 5 in the group.

==

At home, I wrote code for the 1802 processor. Everything was compiled manually. There is no subroutine instructions in the 1802. It passes the program counter to execute a branch. Any register can become the program counter. The first order of business was to write code that would emulate a subroutine call and return.

 

[Irv]

...
At home, I wrote code for the 1802 processor. Everything was compiled manually. There is no subroutine instructions in the 1802. It passes the program counter to execute a branch. Any register can become the program counter. The first order of business was to write code that would emulate a subroutine call and return.

I still have 2 or 3 of those, one a Netronics ELFII, and the others home-built, wirewrapped, with extended memory and an EPROM programmer. I suspect that, like most thing electronic, they'll no longer work after sitting around for 30+ years

The KENBAK-1, arguably the first "personal" computer, since you could buy one for only a month or two's salary, was built with logic gates, no microprocessor. Recently a Kenbak-1 sold for about $30,000.

 

[ratstar]

Thats a great system you have there, would be cool watching it run a simple game on an led matrix like maybe tetris or pong.
Im not into the standard way to make a cpu, I prefer something cacheless maybe using circular access memory (has to be accessed in order, so its good for program instructions. - but note u dont get a goto statement ) and on a really fast oscillation. Im yet to get it going yet, hopefully my brain stays together till I manage to finish the thing, sad for us all when it happens to us tech guys...