Sometimes it starts with a 555 timer and an op-amp. Other times with a small microcontroller. But the timing’s not so great and needs a dedicated timing crystal circuit. And maybe some more memory, and maybe the ATtiny should be swapped out for some 74LS-series chips. And now of course it needs video output too. Before you know it, you’re staring at a 40-chip computer that hearkens back to a simpler, yet somehow more complex, time of computing. At least that’s where [Marcel] is with his breadboard computer based on 1970s-era chips.
For what it does, this homebrew computer is relatively simple and straightforward. It gets 8 bits of processing power from 34 TTL chips. Another 6 round out the other features needed for the computer to operate. It is capable of rendering 64 colors in software and has more than enough memory for a computer of this sort. So far the only recurring problem [Marcel] has had has been with breadboard fatigue, as some of the chips keep popping out of the sockets.
This is a great project for anyone interested in homebrew or 8-bit computing, partially because of some of the self-imposed limitations that [Marcel] imposed on himself, like “only chips from the 70s”. It’s an impressive build on its own and looks to get much better since future plans call for a dedicated PCB to solve the issue with the worn-out breadboards. If you’re already invested in a project like this, don’t forget that the rabbit hole can go a little deeper: you can build a computer out of discrete transistors as well.
39 thoughts on “Homemade Computer From 1970s Chips”
Hehehehe, another TTLer !
“If you’re already invested in a project like this, don’t forget that the rabbit hole can go a little deeper: you can build a computer out of discrete transistors as well.”
Or a difference engine.
Or relays or vacuum tubes or gears and levers
Thanks, that is great! A lot of excellent related videos to watch as well.
That ding at the end made me laugh out loud, lol :)
I really respect all forms of complicated mechanical (and electromechanical) devices, so hats off to this guy for managing to pull this off.
I’ve found that if you put a wire wrap socket into the breadboard this solves the problem of popping out chips. They come in all the sizes and once in, they stay in.
Never had chip pop out of the breadboard.
The troublemakers are the two EEPROM’s for programming: just had too many software updates and no in-situ programming.
It must be me… but the projects on hackaday.io are absolutely unreadable. All of them, so not a fault of the author. If you look at the hackaday.io vs., e.g., instructables, the hackaday.io has information all over the place, no simple way to see the project at a glance, poor structure, etc. Gosh, I like hackaday.com but can’t make heads over tails on hackaday.io.
Finally! Someone who calls out Hackaday’s mess for the mess it is. And it’s the entire thing; not just the projects.
It’s hard work being organized, Hackaday. You obviously don’t like working hard.
Another HaD.io project (mechanical laser show) has a horrible HaD.io page but fortunately there you can find a link to a page on thingiverse.com where everything is clear, the STL files can be found and clear images of the project found in a simple and well organized way. It almost seems like HaD.io is working hard to make things hard… Maybe some sort of pedantic “we are the cool kids, we don’t want no outsiders”?
few pople (including me) theorized it was developed for phone consumption only
So, your criteria for “successful way to present a project” is the ability to find .STL files. Then you compare .io with Thingiverse, a site specifically designed to share STL files, and only share STL files. Your conclusion is that had.io is horrible.
Like, yeah. I get it. That’s technically a valid argument. It’s also a stupid argument.
.io is designed to host projects and project logs in some sort of ‘diary’ format. This is unique. You might counter that the entire purpose of .io is to be like instructables or hackster, where the entire interface is driven around putting up build instructions, BOM, and code. .io also has this functionality, but the main driver is built around project logs.
There’s a lot of criticism of .io, but it basically boils down to 1) stupid arguments, like the one you presented and 2) some sort of nebulous usability issue that isn’t specific enough to turn into actionable design choices. I don’t know if it’s the black background that turns people off. We should do some A/B testing.
Getting salty? Yes it is a HORRIBLE interface/scheme/whatever. It is EXTREMELY DIFFICULT to find ANYTHING on io. Schematics are not there, STL files are not there, mechanical drawings, list of materials, etc. The typical IO page has NOTHING or these are VERY DIFFICULT TO FIND. It is like a soft story with no technical info well presented, and, may I say it… badly organized info. A typical instructable, or thingiverse project is easy to follow, but also easy to reproduce by others. What is the point of publishing a hack if you do not want others to try it and improve on it? Just vanity? Sorry, your entire argument is completely nonsense.
Unbelievable response. Absolutely unbelievable.
“…We should do some A/B testing.”
No, Mr Benchoff, if this is the best you have to offer, what you should do is consider the old saying, “Lead, follow, or get out of the way.” And then, after having seriously considered those choices, get out of the way.
Glancing through, I see the space for some of this stuff but it’s not there on some projects. Really, if a project creator doesn’t want to put it on their page, nothing the software can do about it.
Are you listening, reading, and taking notes
I have to agree with you. I have no idea why they do it that way.
Instructables are for users that want to recreate a project whereas hackaday.io allow project owners to log their project. A lot of hackers(?) have trouble / dislike documenting their projects; a diary format allows people to just dump their thoughts and progress without having to worry about format. I would argue that if hackaday.io would not have a diary format there would be a lot less projects. So I agree but don’t know if a different format would improve things..
Documentation has always been one of those task geeks don’t like. Part of the problem is that a lot of the tools one uses aren’t really geared with an eye towards documentation. Just getting the job done, and moving on.
Yeah, I have to agree. It’s nice to have the capability but it’s confusing to build and the results are not satisfying. Consider:
The wordpress page was much easier to build and more satisfying (for me) to read
The structure is likewise confusing for authors, and don’t get me started on the gallery. For this project, for the longest time I meticulously maintained the “details”, but after one wrong copy-paste a big chunk was irrecoverably lost. After that I just gave up and only use the log section for deltas.
Why not the 74181? That was around in the 70s.
Good question, but in these kind of projects any choice is arbitrary. Having said that, the ‘181 is relatively difficult to obtain and still quite complex. I wanted only simple chips, not necessarily just 1970’s. Jameco had only 6 left when I checked. Now I have the same functionality from 8 identical multiplexers and 2 adders. Much cuter.
If you need another computer project, how about a version of the “paperclip computer” that actually functions automatically? The design in the book is more of a computer simulator because the “memory” is just a bunch of manually operated switches, as are several other functions. I was giving a thought to building one, at first glance it looked like the stone knives and bear skins mnemonic memory circuit sort of thing. “Building magnetic core memory from coiled paperclips? No, wait, they’re just switches… and this is manually operated too. Bah! It’s not actually a computer. Skip this!”
Shouldn’t be too difficult to replace the manual functions with relays or solid state electronics so one can crank the program drum and have it automatically load into memory and do all the shifting and processing like an actual computer.
The scratch build BCD switch design is quite cool though. The whole thing could be built from off the shelf components but the idea of the book is to craft everything but the light bulbs from scratch. Why buy a load of toggle switches when you can twist up some paperclips and screw them to a board? Back when it was published a BCD switch wasn’t something one could easily find and likely would’ve been frightfully expensive.
I was hoping to find on OSHpark more boards that used 70’s/early 80’s CPU chips to make a computer.
i.e. CPU, memory, glue logic, serial/terminal port(s), expansion ports/bus. A lot of those that do, have very little documentation.
I mean, IIRC, I have a couple of N800 chips, Motorola 146805 chips, Lots of EPROMS and memory chips (such as MH*116)and others that I would like to build into something, but I no experience in actually building one.
Oh, and I should mention “monitor” firmware?
This is excellent. Clever idea to work backwards from the video and it worked out well. I’m amazed at the speed.
It has speed because back in the 1970’s Microsoft was building its bloatware.
If you don’t have a bunch of hardware abstraction layers or need to drive a GUI a 6MHz processor can actually be quite speedy at many tasks.
As demonstrated by a lot of GUIs from that era, the processors were up to the task.
I’m waiting for the person that makes 1970s chips. It seems doable using lasers and spinning mirrors.
There is a homebrew transistor subculture. You can do diffusion with household chemicals, and oxide growing in a pottery furnace. http://makezine.com/2013/08/07/how-to-make-a-homebrew-nmos-transistor/ But you probably would still start from industrial doped silicon wafers. A true hero draws her own ignots of course.
There’s also a company that makes modern equivalents to those 1970s chips that can run at a GHz clocks.
I want to see someone build PDP 11 from them.
Once again I find myself reading HaD because some ideas are cool, unfortunately HaD.io is horrible.
These things are like a right of passage for some people. It is an interesting build, very neat.
Nice project :)
Please be kind and respectful to help make the comments section excellent. (Comment Policy)