In the late 80s I enjoyed playing Mach3 on an IBM PC/XT (at work). At home I had an IBM PC/AT with 1 MB RAM and 20 MB HDD. Оn the picture - its motherboard (since the coprocessor got very hot, I put a copper five-kopeck on top)
Edit
View attachment 16527
Interesting to see this type of board again. The C&T 2-chip set, I wonder if that's their NEAT (new, enhanced AT) chipset. Interesting to see the AMD 286 made under license from intel. The chip under the 2 kopecks looks like a regular 287. 3 texas asic's too, perhaps those are C&T designs too and made under license by texas, I remember the early C&T chipsets were 5 asics. Nice to see the 16-bit ISA slots, I used to design cards for those
2 bios chips, a PAL.. a decent chunk of micron drams.
Nice clean layout with only 1 jumper. I remember my genuine IBM AT board had something like 4 or 5 jumper wires on it, and of course all discrete logic, no asics. On the ibm board the xtal was socketed and you could replace it with a faster one to overclock the cpu. The power supply made a ton of noise but the big red switch was a plus point. I can just make out "copyright Award" on the bios chip labels, so it's got an Award bios. Likely to be a 6-layer board I think. The ability to reliably mass-produce a board like that with total hardware compatibility to the ibm original was really ground-breaking at the time. Taiwan made major investments in digital of course. Another historical curiosity on this board is the old 5-pin DIN keyboard connector
The history of C&T and their role in kick starting the taiwan PC clone industry is quite interesting, there's a good article on wikpedia here
https://en.wikipedia.org/wiki/Chips_and_Technologies. ALI were the other big player - ie, Acer. Those chipsets are what made the whole PC clone industry economically viable.
Just for interest I dug out this old prototype 8-bit ISA-bus card I designed back then, perhaps around 1986-7. This was a high-precision (for a PC card) dual Burr-Brown DAC80 card with an 8-bit PIO port, one of the first boards I did for the PC. The DAC80 was a high quality 12-bit DAC
https://www.alldatasheet.com/datasheet-pdf/pdf/56642/BURR-BROWN/DAC80.html, which was industry standard at the time, they were pricey chips in ceramic packages with gold-plated pins.
The design features on-board voltage regulators to supply the DACs with stable levels, an attempt to isolate the DACs from fluctuations on the noisy old PC power supply lines. The large package at the left is a hybrid regulated DC-DC converter from IPS
https://tvsat.com.pl/PDF/4/4CCX_MG.pdf. Because of the application for which this card was designed, which was controlling a high quality analogue video pattern generator (both static and dynamic patterns could be produced), it was important that the DAC outputs exhibited both low noise and high stability over time, since any glitches or noise in the control voltages from the DACs would have been visible to the user on the video monitor screen. Furthermore the DAC80's themselves required symmetric +/- 12 V supply lines, and the DC-DC convertor was used to generate those from the PC's 5V power rail. The convertor in the photo is actually a 'N' (unregulated) variant, that was simply because there was a supply problem getting the regulated 'R' variant at the time that particular prototype board was made, however the regulated 'R' version was used in production. In fact the unregulated convertor performed adquately anyway, probably because there was enough downstream regulation on the board.
From memory, the specification for the required precision of the input control voltage to the video generator was 1 part in 1000, so I basically threw away 2 bits of precision from what the DAC80's were capable of at the limit. At the time the DAC80's were the best chips that were readily available. Component cost was not a big issue because this was not a volume production card (as you can probably guess!), and the external equipment being controlled was very expensive, so the cost of the PC card components was a small fraction of the whole.
The trimmer pots along the top of the card are to trim/calibrate the regulators and DACs, which could be done from above with the card mounted vertically in the PC-AT slot while monitoring the output on a scope or voltmeter. I wrote some software to take the card through a calibration procedure so that the company that made and assembled the PCB's for me could calibrate them. I designed the pcb layout myself by hand, using letraset transfers onto film? , then the pcb masks were made from the film masters... no CAD! It's a 2-layer, PTH board on FR4 with solder mask on both sides. Well, we've all got to start somewhere!
Nowdays you just send your gerber files off to shenzhen and the boards come back by airmail.
Of course you would never get away with a layout like that in modern high frequency PC electronics, but back then bus speeds were much slower and you could get away with this kind of thing. But nowadays you would drive a card like this from a USB chip or something like a Pi anyway. The board itself actually worked very well, the DAC outputs on the two phono sockets were rock solid stable and accurately tracked the input data. The driver and test/calibration software was all written in assembler - MASM.
One interesting little historical note.. I still remember the name of the company that made that parallel port IDC connector socket: Foxconn; that was long before they made iphones for Apple!
Seeing your photo of that old clone mobo took me back, hahaha. Most of the other cards I did are lost in the mists of time, but I kept a couple of the early ones. And I did actually get a CAD later on! ?
As a comparision of my old 1980's design with something that is the current state of the art, have a look at this incredible piece of equipment used to produce control voltages to 25 bits of precision(that's one part in 33 million!!) for quantum computer applications:
https://www.quantum-machines.co/products/qdac (I have no connection to them). The basic principles are the same, of course, but their system is 1000 times more advanced than my design. It's incredible to see how this technology has evolved over time.
I guess this hasn't got much to do with FreeBSD ? which wasn't around at the time I made that card, but it does tell you a bit about me, much more interesting than hearing about my hobbies! I've been a software engineer as well of course, I've worked on windows, os/2, then mostly unix and linux, embedded, voice, realtime, image, satellites, image processing, etc etc. I've done some work on storage systems. The usual nonsense. Used FreeBSD on and off for maybe around 15 years, its a nice system. I used to particularly enjoy running frenzy
http://frenzy.org.ua/en/ entirely from RAM.
