Back when the IBM PC was new, laying out an ISA board was a daunting task. You probably didn’t have a very fast ‘scope, if you had one at all. Board layout was almost certainly done on a drafting table with big pieces of tape. It was hard for small companies, much less hobbyists, to make a new card. You could buy a prototype board and wirewrap or otherwise put together something, but that was also not for the faint of heart. But with modern tools, something like that is a very doable project and [profdc9] has, in fact, done it. The card uses an ATMega328P and provides two SD cards for use as mass storage on an old computer.
The design tries to use parts that won’t be hard to get in the future. At least for a while, yet. There’s capacity for expansion, too, as there is an interface for a Wiznet 5500 Ethernet adapter.
Can you imagine if you could transport this card back to the days when the ISA bus was what you had? Just having a computer fast enough to manipulate the bus would have been sorcery in those days.
We don’t know if you need an ISA mass storage card, but if you do, [profdc9] has you covered. Then again, you do have options. Or, if you’d rather take a deep dive into the technology, we can help there, too.
For a prototype get a broken card,Dremel away the components from the slot connections and solder wires from there to your card in mezzanine,that is how I did a custom card for an EPROM programmer in the 90s, I didn’t know better than that
Heh I did exactly the same at roughly the same time. Still have it and the birds nest of wires somewhere in a box at my parents house.
IBM used to do a neat trick with some of their cards, they broke out the ISA bus onto a row of pads above the edge connector which made debugging the bus easier but also allowed you to tack on a proto card.
You didn’t need to sacrifice the card either, the ISA bus is pretty simple and as long as you don’t decode to the same address space, IRQ or DMA, you can stack cards in one slot (you obviously had to be conscious of fanout).
But these days, I can’t imagine not getting a proto card made to spec by one of the cheap PCB houses
This is a cool hack. I especially like the convenience of 2 SD cards instead of just one.
I have one of these: https://texelec.com/product/picomem/ . It works very well – which is good, because it uses an RP2040 and has more power than a truckload of 8088 PCs.
Note, I did not specify the size of the truck.
ISA was quite easy, you just need a handful of TTL IC’s to do the buffering and some address decoding. But on later many commercial designs those were probably all integrated in some ASIC.
But where do you find a PC that still has ISA slots?
I also did a quick search for PCI express prototype card and found lots of PCB’s with both a prototyping area and an FPGA to handle the bus traffic. But with the common re-programmability of FPGA’s you can also re-purpose an existing “mass produced” board and modify it to do your own bidding, such as for example:
https://hackaday.com/2024/12/07/cheap-fpga-pcie-development/
LPC bus adapter. Presto, ISA bus.
Presto ? I know LPC bus is serializing ISA, but never seen an adapter doing it. Link ?
https://www.vogons.org/viewtopic.php?t=93291
i would find a pc like that in my basement :) My friend found it literally in a barn 20 years ago, and now it’s in my basement.
That is not true. In the eighties I made my own card for Apple[]+ and PC and I was a school kid at this time. That was not so special!
ISA is boring easy!
Everything is relative.
By “hard” it was probably meant that there was no online PCB etching service available at the time.
Just like going to local public library was “hard” in retrospect.
Or buying and reading printed books and magazines to educate yourself.
Everything nowadays is being seen from an always-online perspective, with everything being on your fingertips with zero effort.
From that point of view, soldering through-hole components on a veroboard was a total chore, too.
In retrospect, I mean. Back then it was perceived differently, we weren’t such “softies” yet and had fun to experiment and to learn through failure, too.
A burned finger or hand from tip of a soldering iron was part of the experience.
Back then, you had to etch your copper board at home in bathroom using iron-3 chlorid (?) etc.
PCB design was either done by drawing everything by hand with a pen,
by using photo copies of PCBs shown in magazines or by using software.
Eagle Cad for DOS PCs, for example, was popular in Germany of late 80s I think.
There also was P-Cad (?) and various shareware/public domain programs for PC or Atari ST/Amiga/C64 etc.
ISA/PC-Bus, ECB and Apple II prototyping cards surely had been available too at some point.
Printed books also sometimes contained experimental PCBs that had to be populated by the user.
They were inside a plastic sleeve that was glued to the inner side of the cover.
The 5,25″ or 3,5″ floppy was stored same ways, often. On the other end of the book.
This was so-called “bookware”, if memory serves.
Likewise, online services and online shopping had existed by the mid-1980s, already.
CompuServe, Genie, Minitel (Videotex family) etc.
There had been early services that were similar to nowaday’s online experience.
It simply just wasn’t the www/internet yet, but commercial/proprietary online services.
That sounds limiting, but many mail order companies/travel agencies or banks had offered their service on more than one online service.
Of course, this required knowledge. Users had to actually research and inform themselves about current technology at the time.
The users had to seek out for information, read news and computer magazines.
Or ask their computer vendor/bank/travel agency via phone, at least.
There was no oracle of Delphi to ask yet. Or ChatGPT, for the matter.
Human interaction did still matter.
idk, this article is way over selling everything it says. and it says very little. manually wiring a prototype board? thats how I make pcbs. its about as much work as making it in a schematics program then routing and board layout. then DEBUGGING the board then making the final solution. is about as much work as let me place the components on an actual pcb and wire it up.
I did a number of these in the 80’s using EE Designer III. We even did SMD ISA boards in like 1986.