I'm Japanese, and the 80286 at 10MHz was huge for Japan's PC-98 scene. The V30 handled backward compatibility while the 286 ran much faster than what we had before. This project brought back memories—the 286 was the chip of my era, and it's great to see people still exploring its capabilities decades later.
I've seen a similar project a while ago and thought this was about the same thing at first: [1][2]
Both essentially built a DIY chip tester for a 286 and both built around a Harris 80C286.
If I understood it correctly, the goal behind this project seems simulating the rest of the PC, purely for the challenge and learning experience, documenting the process of building the chip tester (and getting mildly philosophical in the process).
The other project was more directly interested in the 286 itself, undocumented instructions, corner cases in segmentation behavior, instruction cycle timing, etc. and also trying to find out if there are any difference between the Harris and Intel variants.
I'm not a fan of breadboards, they tend to be unreliable even for trivial circuits. We need something more affordable and practical for home PCBs[^1]. Why is it that nobody has invented a tin 3D printer, or at least a 2D version of it, i.e. a tin plotter?
[^1]: I'm discouraged from home-etching by the chemicals and the dark-room part of the process.
For the 2D version, you might not need very much custom. Use a regular pen plotter and use a pen with conductive ink. These both exist today, though personally as a hobbyist PCB designer, I can get 2-layer and 4-layer boards cheap enough from JLCPCB or Oshpark or PCB Unlimited that I don't bother trying to make them myself.
I haven't tried it myself so you might be right, but I was thinking of the silver conductive pens from Chemtronics with a conductivity of 0.02-0.05 ohms/sq/mil.
For attachment, I'd evaluate their conductive epoxy or maybe glue down the underside of the component and then smother the lead with the silver conductive ink. But again, just hypothetical since I have a quickturn shop make cheap prototype PCBs for me and either hand solder or use a stencil, paste, and a hot air gun for my hobbyist projects.
Yes, conductive ink has too high resistance; at least the one that works on a basis of carbon; a "trace" can easily have kilo-Ohms, and the metal ink interface makes things worse.
I remember reading some "Sputnik" magazines from the 1970s where Russian scientists were searching for the holy grail of a good conductor resin. I didn't understand at the time why they found the (concept of the) thing so valuable; but now I've got an inkling...
Just use a fabrication service. You can't readily make plated through holes and vias at home. The services do a much better job than any hobbyist level tinkering can achieve.
I worked at a no longer extant networking equipment manufacturer as an intern in college in the late 1990's. My role was to work on software for an in-development 45Gb network switch, and a bunch of the software I wrote ran on prototype boards.
Since fabricating new boards took time and was expensive, a lot of work was done to make in situ modifications that involved an insane amount of wirewrapping. One member of the team did that all day, every day as their full time job, and I was always amazed by their ability to focus consistently at that level for so long.
For $2,500 the Carvera Air makes very nice 2 sided pcbs with solder mask. Though even in raw materials it is hard to match a finished board from China if you can wait a couple of months.
My last job had a desktop CNC called Volterra I think. It would squeeze out conductive ink, bake it, drill holes, and lay down and reflow solder paste.
It was expensive and performed terribly.
I think there really isn't any good way to improve on breadboards. Breadboards, in fact, are the improvement. They're called breadboards because we used to literally drill pieces of wood and do wire-wrap construction on the other side.
Breadboards are good enough for the kind of prototyping they're for. Spring loaded contacts are about the best you can get for removable parts. The signal integrity isn't that bad at modest speeds.
In today's world, the next step up from a breadboard is custom PCBs. You can have a set of five shipped from China for the same price as a set of breadboards. There's no real need or reason for anything in between so long as PCB manufacturing is so disgustingly cheap and fast.
Ha, little fall down the memory hole. I had a Harris 20mhz 286 back in the late 80's. I thought that thing was a beast at the time. Paired it with a Conner 100MB disk and I remember my brother laughing derisively, "Who the hell needs 100 megabytes?"
Looking at the amount of wires going into this, my instinct is that this cannot scale, in 5-10 years this won't be doable for a Pentium chip, at least not as an at home hobby project. But I actually think it could go the other way, and in 5-10 years you'll be able to do this at home for far more sophisticated kit, unlocking crazy amounts of reverse engineering possibilities that were once thought of as near impossible, or at least only possible for a nation state scale setup.
I'm pretty sure if you really want to, you could do something like this as a hobbyist with a Pentium right now.
Instead of futzing with wires on a breadboard you could simply designing a PCB up front, throw the design over the fence at JLC or PCBWay, insert coin, wait patiently at the mailbox, solder your scavenged Socket 7 onto the board.
The days of toner transfer and aquarium pumps are already long gone. Getting production quality, one-off, multi layer PCBs done as a hobbyist is dirt cheap these days, no government budgets required.
> you could simply designing a PCB up front, throw the design over the fence at JLC or PCBWay, insert coin, wait patiently at the mailbox
It blows my mind that I can use free-as-in-beer Free-as-in-Speech software to design a PCB, email it to a dude in China, and get a finished working professional-looking PCB back in my hand within a week, for the price of a couple of coffees. And if I want the components stuck on too, it'll cost a little extra, maybe three coffees it costs now.
If I want it really quickly then for the price of a decent takeaway curry I can have it flown over next day. What the actual hell?
Edit: the slowest part of "next day" is when it hits the UK, and if I could guarantee it just gets delivered to DHL's Edinburgh depot I could drive down there in two hours.
The reality is that's just how cheap commodity circuit construction is, and even small shops in the US sometimes approach that low cost, and we've been paying crazy markup on electronics for decades. That dirt cheap board is so profitable it is using air freight to get back to you. It is literally burning money just for convenience, yet that is the "cheap" option.
Electronics cost a lot to manufacture in the 70s, but is entirely automated now but for "reasons" we have only seen a small part of that savings.
When you buy the "Cheap" version on AliExpress, they are still making a healthy margin, yet Americans will happily buy the exact same product off Amazon for next day shipping for 10x the cost and think they are getting a "deal"
This extends to cars as well, with the F150 costing as little as $20k to build, even with "Expensive" very unionized and well compensated labor. The higher market trims only cost a little more to make but take in far higher profit margins. How much of China's supposedly "Subsidized" car price (as if the US doesn't do anything to subsidize cars) is just a lower profit margin?
Things should be way cheaper to western consumers. Where does all that extra money go? "Marketing and administration", basically bloated executive suites, bloated middle management, and the pockets of Meta, Google, AWS, and Apple. Oh gee, those exact companies seem absurdly wealthy and are basically responsible for all economic growth in the past few decades.
> even small shops in the US sometimes approach that low cost
It's ludicrously expensive to ship things to and from the US though, and since they're now paying some insane markup because no-one understands what tariffs are the prices have got even sillier.
> My first computer was a 286 with 1 MB of RAM and a 50 MB HDD
That was the configuration I had when I went to college. I had an amber monitor that had a really crisp image and 5.25" HD floppy drives. The AT case was fantastic, I wish I had one today.
I'm Japanese, and the 80286 at 10MHz was huge for Japan's PC-98 scene. The V30 handled backward compatibility while the 286 ran much faster than what we had before. This project brought back memories—the 286 was the chip of my era, and it's great to see people still exploring its capabilities decades later.
Japanese PC-98 games have an aesthetic that’s so unique. There’s this one Twitter account that posts images from visual novels from that era and they all look so pretty: https://x.com/PC98_bot Also on bluesky https://bsky.app/profile/pc98bot.gang-fight.com
I've seen a similar project a while ago and thought this was about the same thing at first: [1][2]
Both essentially built a DIY chip tester for a 286 and both built around a Harris 80C286.
If I understood it correctly, the goal behind this project seems simulating the rest of the PC, purely for the challenge and learning experience, documenting the process of building the chip tester (and getting mildly philosophical in the process).
The other project was more directly interested in the 286 itself, undocumented instructions, corner cases in segmentation behavior, instruction cycle timing, etc. and also trying to find out if there are any difference between the Harris and Intel variants.
[1] https://forum.vcfed.org/index.php?threads/286-cpu-experiment...
[2] https://github.com/dbalsom/arduinoX86
Cool! But the wires!
I'm not a fan of breadboards, they tend to be unreliable even for trivial circuits. We need something more affordable and practical for home PCBs[^1]. Why is it that nobody has invented a tin 3D printer, or at least a 2D version of it, i.e. a tin plotter?
[^1]: I'm discouraged from home-etching by the chemicals and the dark-room part of the process.
Bad Obsession Motorsport actually went this route, you can see their custom printer on their youtube[1] channel
[1]: https://www.youtube.com/watch?v=FzrZoVKT8gM
For the 2D version, you might not need very much custom. Use a regular pen plotter and use a pen with conductive ink. These both exist today, though personally as a hobbyist PCB designer, I can get 2-layer and 4-layer boards cheap enough from JLCPCB or Oshpark or PCB Unlimited that I don't bother trying to make them myself.
I always thought conductive ink had too high a resistance to use to make PCBs. That’s not the case?
How do you attach the components to it?
I haven't tried it myself so you might be right, but I was thinking of the silver conductive pens from Chemtronics with a conductivity of 0.02-0.05 ohms/sq/mil.
For attachment, I'd evaluate their conductive epoxy or maybe glue down the underside of the component and then smother the lead with the silver conductive ink. But again, just hypothetical since I have a quickturn shop make cheap prototype PCBs for me and either hand solder or use a stencil, paste, and a hot air gun for my hobbyist projects.
https://www.chemtronics.com/circuitworks-conductive-pen
https://www.chemtronics.com/circuitworks-conductive-epoxy-2
Yes, conductive ink has too high resistance; at least the one that works on a basis of carbon; a "trace" can easily have kilo-Ohms, and the metal ink interface makes things worse.
I remember reading some "Sputnik" magazines from the 1970s where Russian scientists were searching for the holy grail of a good conductor resin. I didn't understand at the time why they found the (concept of the) thing so valuable; but now I've got an inkling...
Just use a fabrication service. You can't readily make plated through holes and vias at home. The services do a much better job than any hobbyist level tinkering can achieve.
wirewrap circuit construction would be more reliable. However high speed (if needed ) could be a problem ? from crosstalk.
I worked at a no longer extant networking equipment manufacturer as an intern in college in the late 1990's. My role was to work on software for an in-development 45Gb network switch, and a bunch of the software I wrote ran on prototype boards.
Since fabricating new boards took time and was expensive, a lot of work was done to make in situ modifications that involved an insane amount of wirewrapping. One member of the team did that all day, every day as their full time job, and I was always amazed by their ability to focus consistently at that level for so long.
One maybe unrecognized problem, of breadboards is not so good for high speed digital circuits due to capacitance
For $2,500 the Carvera Air makes very nice 2 sided pcbs with solder mask. Though even in raw materials it is hard to match a finished board from China if you can wait a couple of months.
In my experience it's a couple of weeks, not months from China.
Depends how much you are happy to pay; but yes it has got faster.
My last job had a desktop CNC called Volterra I think. It would squeeze out conductive ink, bake it, drill holes, and lay down and reflow solder paste.
It was expensive and performed terribly.
I think there really isn't any good way to improve on breadboards. Breadboards, in fact, are the improvement. They're called breadboards because we used to literally drill pieces of wood and do wire-wrap construction on the other side.
Breadboards are good enough for the kind of prototyping they're for. Spring loaded contacts are about the best you can get for removable parts. The signal integrity isn't that bad at modest speeds.
In today's world, the next step up from a breadboard is custom PCBs. You can have a set of five shipped from China for the same price as a set of breadboards. There's no real need or reason for anything in between so long as PCB manufacturing is so disgustingly cheap and fast.
Ha, little fall down the memory hole. I had a Harris 20mhz 286 back in the late 80's. I thought that thing was a beast at the time. Paired it with a Conner 100MB disk and I remember my brother laughing derisively, "Who the hell needs 100 megabytes?"
Here's a project that also does this to ensure cycle-accuracy for their emulator: https://github.com/dbalsom/arduinoX86
So they're stepping a 286 very slowly and sorta... bit-banging the I/O pins on it?
Love it. No notes.
Looking at the amount of wires going into this, my instinct is that this cannot scale, in 5-10 years this won't be doable for a Pentium chip, at least not as an at home hobby project. But I actually think it could go the other way, and in 5-10 years you'll be able to do this at home for far more sophisticated kit, unlocking crazy amounts of reverse engineering possibilities that were once thought of as near impossible, or at least only possible for a nation state scale setup.
I'm pretty sure if you really want to, you could do something like this as a hobbyist with a Pentium right now.
Instead of futzing with wires on a breadboard you could simply designing a PCB up front, throw the design over the fence at JLC or PCBWay, insert coin, wait patiently at the mailbox, solder your scavenged Socket 7 onto the board.
The days of toner transfer and aquarium pumps are already long gone. Getting production quality, one-off, multi layer PCBs done as a hobbyist is dirt cheap these days, no government budgets required.
> you could simply designing a PCB up front, throw the design over the fence at JLC or PCBWay, insert coin, wait patiently at the mailbox
It blows my mind that I can use free-as-in-beer Free-as-in-Speech software to design a PCB, email it to a dude in China, and get a finished working professional-looking PCB back in my hand within a week, for the price of a couple of coffees. And if I want the components stuck on too, it'll cost a little extra, maybe three coffees it costs now.
If I want it really quickly then for the price of a decent takeaway curry I can have it flown over next day. What the actual hell?
Edit: the slowest part of "next day" is when it hits the UK, and if I could guarantee it just gets delivered to DHL's Edinburgh depot I could drive down there in two hours.
The reality is that's just how cheap commodity circuit construction is, and even small shops in the US sometimes approach that low cost, and we've been paying crazy markup on electronics for decades. That dirt cheap board is so profitable it is using air freight to get back to you. It is literally burning money just for convenience, yet that is the "cheap" option.
Electronics cost a lot to manufacture in the 70s, but is entirely automated now but for "reasons" we have only seen a small part of that savings.
When you buy the "Cheap" version on AliExpress, they are still making a healthy margin, yet Americans will happily buy the exact same product off Amazon for next day shipping for 10x the cost and think they are getting a "deal"
This extends to cars as well, with the F150 costing as little as $20k to build, even with "Expensive" very unionized and well compensated labor. The higher market trims only cost a little more to make but take in far higher profit margins. How much of China's supposedly "Subsidized" car price (as if the US doesn't do anything to subsidize cars) is just a lower profit margin?
Things should be way cheaper to western consumers. Where does all that extra money go? "Marketing and administration", basically bloated executive suites, bloated middle management, and the pockets of Meta, Google, AWS, and Apple. Oh gee, those exact companies seem absurdly wealthy and are basically responsible for all economic growth in the past few decades.
> even small shops in the US sometimes approach that low cost
It's ludicrously expensive to ship things to and from the US though, and since they're now paying some insane markup because no-one understands what tariffs are the prices have got even sillier.
I don't know about Pentium, but there's definitely some homebrew 486 projects out there, e.g.:
https://maniek86.xyz/projects/m8sbc_486.php
> My first computer was a 286 with 1 MB of RAM and a 50 MB HDD
That was the configuration I had when I went to college. I had an amber monitor that had a really crisp image and 5.25" HD floppy drives. The AT case was fantastic, I wish I had one today.
Reminds me of the Tamagotchi Matrix:
https://hackaday.com/2015/11/24/building-the-infinite-matrix...
Very impressive to use intel cpu without BIOS. Real hacker
[dead]