A Rare ATX XT Motherboard by Hyundai Electronics

Posted by BEST_NK@reddit | vintagecomputing | View on Reddit | 19 comments

I recently came across an XT motherboard made by Hyundai Electronics, now known as SK hynix. It appears to be called the HYUNDAI SUPER-88. What makes it particularly interesting is that it was built in the ATX form factor. I’m not sure why it was designed this way, but it seems likely that it was built for some specialized purpose. Otherwise, the board itself is a fairly standard 10 MHz 8088 turbo XT.

I installed this unusual motherboard in a SilverStone FLP01 case. Since the motherboard uses an AT power connector, I used ATnoX (a design by hkzlab) to make it compatible with the case. The FLP01 offers virtually no space for cable management, so I had to put in quite a bit of effort to keep the interior tidy.

[-]

TxM_2404@reddit

This is not ATX. It was designed for a PC with a custom form factor and it just happens to have the I/O match up with ATX.

[-]

gcc-O2@reddit

Yeah. Manufacturers were experimenting with putting the onboard i/o in that general area as far back as the Columbia Data Products, the first PC clone at all (IIRC). The reason it didn't become more standard is: OEM machines detoured to LPX (using a riser card to hold the expansion slots) so the onboard ports blocked where the expansion slots are in an AT/ATX case, and Baby AT chopped a full AT motherboard in half but in the wrong direction so as not to leave room for any onboard slots, while ATX re-chopped it in half but the other way so that there's an onboard i/o area.

Still it's a neat coincidence that this board happens to be close-enough to ATX to fit into that case, just because Hyundai also stumbled upon chopping Full AT in half in that direction rather than the other.

[-]

dst1980@reddit

From what I recall, the Compaq Portable was the first 100% compatible clone, including reverse engineered BIOS.

[-]

starcube@reddit

It doesn't even match up at all. The connectors are not in the same positions. They're just in the same quadrant as ATX.

[-]

fadedspark@reddit

Yeah. It happened to fit the footprint but the mounting points are way far away, and part of the ATX standard is ATX Power supply support, and this is still AT.

[-]

Educational_Bee_6245@reddit

Very interesting. The chips have date codes from 1991 but ATX specification was release in 1996.

[-]

mallardtheduck@reddit

Revision 1.0 of the ATX specification was published in July 1995 (according to a note in the 1.1 spec from February 1996, which seems to be the oldest surviving version available online). There were probably draft specs circulating around hardware vendors for a year or two before that, so maybe some hardware based on draft specs could have been made as early as late 1993. Still a bit late for an XT board with 1991 date codes though...

[-]

gcc-O2@reddit

And the final innovation was pop-in i/o shields, making any motherboard still fit in any case like Baby AT (misaligned screw holes aside). The first generation of ATX motherboards/cases like 430HX and 430FX ones didn't use the modern pop-in i/o shield yet.

[-]

Albos_Mum@reddit

There were also hybrid ATX cases that'd allow you to install an AT or BabyAT motherboard, some of which also came with ATX power and front panel connectors in case you had one of those cases and an ATX PSU.

I've got a Super 7 hybrid BabyAT board with the ATX connectors as well as the AT ones.

[-]

neighborofbrak@reddit

Date codes on chips just means that the build is no earlier than the date printed.

Absolutely crazy to see an XT-class board in an ATX board profile.

[-]

Educational_Bee_6245@reddit

True, but how likely is it to build such a board 1996 for ATX?

Considering that the board is missing other ATX features (like the power connector) I think it's more likely someone had the same idea (turning a Baby AT board by 90 degrees) for their proprietary system a few years earlier? That it fits in an ATX case is just a happy coincidence.

[-]

AppropriateCap8891@reddit

The XT actually had a fairly long lifespan, I was still making brand new XT machines as late as 1993.

I think a lot of the problem we keep seeing in these threads is that most people were not around at the time, so really have no idea the state of the computer industry in this era.

And the fact is, the "death of the XT" was pretty much at the same time as the death of the 286 and 386. And that was in 1995 when Windows 95 came out. Before then, most software was still DOS driven, and the actual hardware used really did not matter. If you were using WordStar, Lotus 1-2-3 or dBase III, it did not matter if you were using an 8086 or an 80486. And over 90% of turnkey solutions for everything from pawn shops, dry cleaners, hotels and video stores were all DOS based.

It was only after Win95 that we saw a sudden dumping of those older systems, and the demand that all software now support Windows.

And some of the late model XT systems were actually quite impressive. The late model NEC V33 CPU in particular was damned amazing. NEC was given a license to produce XT clone chips, and they took the chips far beyond what Intel ever imagined. The late ones would actually support up to 16 MB of RAM (as opposed to 1 MB of the original), and outperform 80286 processors.

But this board is simply another proprietary XT board that would have been made for a specific manufacturer. Not unlike their Super-16 series of boards.

https://theretroweb.com/motherboards/s/hyundai-super-16t

The fact that this would never work in a standard XT-AT case screams it is proprietary. And the oddity of having two floppy connectors also has me shaking my head a bit.

[-]

gcc-O2@reddit

(turning a Baby AT board by 90 degrees)

Other way around - if you look at the IBM Full AT motherboard https://minuszerodegrees.net/misc/5170_motherboard_type_1.jpg all that extra space for potential I/O ports is staring you in the face, if you just chop it in half at the end of the expansion slots (like ATX), rather than chopping off everything past the keyboard port (like Baby AT)

[-]