3 ms·

Others have answered why not later, but why not earlier is also a question. Earlier computers were a lot more work. CPUs weren't single chips, they were collect

Others have answered why not later, but why not earlier is also a question. Earlier computers were a lot more work. CPUs weren't single chips, they were collections of dozens or hundreds of ICs. Before that they were thousands of individual transistors, Ben Eater has a Youtube series designing and building such a machine. Before that came vacuum tubes, in types which are no longer produced and aren't at all common to find working. Before that came electromechanical systems with machined cams, cogs, wheels, lots of relays, etc. So the earlier in computing capability you go the more expensive and difficult it gets. Early 80s parts are a bit of a sweet spot combination of low price, enough challenge to be interesting but not so much as to be frustratingly difficult, and an interesting turning point in computing history as the first single-IC CPUs became popular.

flir 2y ago
Ownership? It was the personal computing revolution, after all.
- hx8 2y ago
  I think that's a big draw to 80s computers. It was the first time most people knew someone that had a personal computer.
dcminter 2y ago
I'm fascinated by the early PDP-11 range - and if I saw one readily available, reasonably priced, and close at hand (shipping is not trivial for those machines) then I don't know if I'm actually competent enough to handle a machine where normal configuration duties include changing the wirewrapped backplane! That's part of the fascination too though.
crote 2y ago
Yeah, exactly this. I've been toying with the idea of making a computer out of individual transistors, but building something capable enough of having an interactive CLI would probably take years, thousands of dollars, and some nontrivial engineering. Hooking up a 68k to some peripherals? Quite doable. Building something out of 7400 ICs? Tricky, but if your expectations aren't too high definitely possible. Transistors, vacuum tubes, mechanical stuff? You better be really dedicated to it!
- tzs 2y ago
  > Hooking up a 68k to some peripherals? Quite doable. I did that for my project in the microprocessor lab class I took in college in the early '80s. It is indeed quite doable [1][2]. It was a bit scary, because the 68k was quite new and Motorola only gave the school a small number of samples. I was told that if I fried it or broke it by not being careful enough when removing it from a socket I would not be given another one. I could not afford to buy one either. And I was taking this class in the last term of the school year as a senior, so failing it would mean not graduating. > Building something out of 7400 ICs? Tricky, but if your expectations aren't too high definitely possible. I considered doing that. EdX has a version of MIT's 6.191, "Computation Structure", which goes through how logic gates work at the MOSFET level on chip, then goes through combinatorial and sequential logic, followed by how to design a 32-bit RISC processor, which you them build and test in a logic level simulator. (You don't do all of the processor. You are given pre-defined modules for a RAM bank, a 32 x 32bit register file, and ROM). That's the first two parts of the course. The third part adds pipelining and caching to processor. I took the first two parts at EdX and afterwards seriously considered actually building my processor out of 74xx series chips. My parts list came to 350 chips. And that's not counting whatever parts I'd need for the RAM, register file, and ROM. That's way to big for my breadboard! :-) My ALU design includes a shift unit that can do left or right shifts or rotates from 0 to 31 bits in one cycle and that uses around 90 chips. I could drop that, and change things so that shift and rotate instructions trap, and then emulate them in software. That cuts the chip count down to around 260. Still too big for me. Even changing from 32 bits to 16 bits, or even 8 bits, would be too big, and so the idea to build it was discarded. [1] https://imgur.com/Ts9wcfW https://imgur.com/Ts9wcfW [2] https://imgur.com/3D4rvdC https://imgur.com/3D4rvdC

flir 2y ago

Ownership? It was the personal computing revolution, after all.

hx8 2y ago
I think that's a big draw to 80s computers. It was the first time most people knew someone that had a personal computer.

dcminter 2y ago

I'm fascinated by the early PDP-11 range - and if I saw one readily available, reasonably priced, and close at hand (shipping is not trivial for those machines) then I don't know if I'm actually competent enough to handle a machine where normal configuration duties include changing the wirewrapped backplane! That's part of the fascination too though.

crote 2y ago

Yeah, exactly this. I've been toying with the idea of making a computer out of individual transistors, but building something capable enough of having an interactive CLI would probably take years, thousands of dollars, and some nontrivial engineering. Hooking up a 68k to some peripherals? Quite doable. Building something out of 7400 ICs? Tricky, but if your expectations aren't too high definitely possible. Transistors, vacuum tubes, mechanical stuff? You better be really dedicated to it!

tzs 2y ago
> Hooking up a 68k to some peripherals? Quite doable. I did that for my project in the microprocessor lab class I took in college in the early '80s. It is indeed quite doable [1][2]. It was a bit scary, because the 68k was quite new and Motorola only gave the school a small number of samples. I was told that if I fried it or broke it by not being careful enough when removing it from a socket I would not be given another one. I could not afford to buy one either. And I was taking this class in the last term of the school year as a senior, so failing it would mean not graduating. > Building something out of 7400 ICs? Tricky, but if your expectations aren't too high definitely possible. I considered doing that. EdX has a version of MIT's 6.191, "Computation Structure", which goes through how logic gates work at the MOSFET level on chip, then goes through combinatorial and sequential logic, followed by how to design a 32-bit RISC processor, which you them build and test in a logic level simulator. (You don't do all of the processor. You are given pre-defined modules for a RAM bank, a 32 x 32bit register file, and ROM). That's the first two parts of the course. The third part adds pipelining and caching to processor. I took the first two parts at EdX and afterwards seriously considered actually building my processor out of 74xx series chips. My parts list came to 350 chips. And that's not counting whatever parts I'd need for the RAM, register file, and ROM. That's way to big for my breadboard! :-) My ALU design includes a shift unit that can do left or right shifts or rotates from 0 to 31 bits in one cycle and that uses around 90 chips. I could drop that, and change things so that shift and rotate instructions trap, and then emulate them in software. That cuts the chip count down to around 260. Still too big for me. Even changing from 32 bits to 16 bits, or even 8 bits, would be too big, and so the idea to build it was discarded. [1] https://imgur.com/Ts9wcfW https://imgur.com/Ts9wcfW [2] https://imgur.com/3D4rvdC https://imgur.com/3D4rvdC