how do I build my own 8 bit cpu using only transistors?
It's easy?
>>958824
Start with something simple, a 4 instruction cpu that can load a register and add two numbers or something. Forget microcode, use combinational logic based on the instruction for now.
Start at a high level, define your alu, controller, memory etc. Then figure out their operation, create from logic gates then simply swap logic with the fet equivalent
>>958834
It's there any webpage with some easy to follow instructions?
>>958837
Yes if you look for them a lot of colleges/universities will have coursework similar to this idea that you should be able to find but don't expect to be just given the answer.
There are plenty of people who build these and document them online but usually the design is complex and I wouldn't recommend for a starter.
Download something like quartus for fpga programming and you can design the system in verilog or gates and test it out first
>>958846
Keep in mind that basic mcu design is usually done in third year or so of uni/college and requires a fair bit of background knowledge.
>>958824
I hope you like troubleshooting.
This guy has a couple articles about it. http://www.bigmessowires.com/bmow1/
you're gonna get real comfy with wire wrap.
>>958896
I want you to close your eyes and imagine for a second, that you just finished this project. As you are about to turn it on to see if it actually works, someone walks in, grabs a fist full of wires, and I dont mean a girly fist, I mean a mans hand full of wires and rips them out.
>>958907
All the wires sticking up are test leads. The wire wrap wires are not really grabbable. They'd have to grab wires with a pair of pliers.
>>958916
fine then, imagine some edgelord with those wolverine claws you see at flea markets comes by and rips out a ton of wires and then tips his fedora at you.
>>958931
>>958824
I'd use relays if you're just starting out. Easier to use, harder to break, easier to debug (click, clack), and can do some things that will make the job easier (multiple pole relays, latching relays, etc.).
>>958837
Look up boolean logic to get an idea of how it works, then CPU architecture to understand the big picture, then relay logic to understand how to implement it.
>>958875
Doesn't mean it's hard. College is more of a money-for-paperwork exchange than a real intellectual challenge.
>>958824
Shouldnt be extremely hard, people have done it in minecraft
>>958824
I m drooln'
>>958824
Google around, you can find a schematic for a 6502 microprocessor.
>>958824
Introduction to Computing Systems by Schoken
>>958824
Use logic chips, it's pretty hardcore if you only have BJTs or FETs.
>>958943
college can be hard, it all depends on what school you attend and how driven you are
>>959517
2nd on this, there's a course based on the book that the authors teach called NAND to Tetris. It's actually running on Coursera right now, I'm taking it and have been pleased so far.
>>959591
that's cheating
>>958824
Read this book.
I'm not even kidding when I say ignore everything everyone else has said and just buy & read this book. It'll explain everything you need to know to even begin to understand the task. You don't even know what you don't know until you've read this book.
>>959650
Not OP but how much?
>>958824
Fuck off Ahmed I'm not teaching you anything you fraudulent little shit. Go read a fucking book.
Step 1: build it in a hardware description language and get it to work in simulation
Step 2: translate the HDL into transistors in real life.
>>958896
http://megaprocessor.com/index.html
>>959780
>>958824
Did an ALU for 4 bit computer addition, subtraction and most logic operations on simulation, and got it working about halfway through because of bad design of my logic gates. Picture is the pcb layout for half the thing, without any memory, about 500 components.
It's possible to do it, but it will takes you load of time. I had almost no knowledge about electronic and ALU when I started it.
First simulate your logic circuit and get it running as you expect. I'd say it is necessary, as you will sometime forget which part of your circuit do what. Mostly, it will save you from "Are you wasting your time to make a circuit work when the logic design is wrong in the first place?".
If after you're still up to build your circuit, make it modular so that you can divide the work.
Second look this up, this guy did a fine job but don't use the design for the logic gates, because you can't line more than about 10 of them with that design:
http://www.waitingforfriday.com/index.php/4-Bit_Computer
Finally pick a support to build your circuit: breadboards are good to start, but erf find one that fits 200 transistors on it or just to have the whole thing wired is a mess. I did the first try on breadboard, test a couple logic gates, then when it's seemed fine, I did a PCB, but if you don't feel confident printing your own PCB, just try on stripboard.
>>959771
there's a torrent somewhere.
