Ok. Yeah, I know EISA is backwards compatible. Just didn't make any sense why it would only be EISA.

Here's the gist of it (layout)
PS/2 (bottom kboard, top mouse)
USB
VGA
Serial
CF

_______ __ (PCI)

________ ____ (EISA)
________ ____ (EISA)
________ ____ (EISA)

Nu layout ; )

I am only worried about how long the old hardware lasts and that would be main thing in wanting a 486 motherboard made of newer components and working with a newer power supply.
I still want to have my 486 33Mhz dos PC 20 years from now. It is definitely my favourite retro pc.

Baoran wrote on 2020-03-05, 19:23:

I am only worried about how long the old hardware lasts and that would be main thing in wanting a 486 motherboard made of newer components and working with a newer power supply.
I still want to have my 486 33Mhz dos PC 20 years from now. It is definitely my favourite retro pc.

Agreed. I want my Compaq Housecall Reciever to be working when I have kids, in 15-20 years.

Tbh this is a schizofrenic idea...

Either:
1) do a barebones So3 ATX board with socket and slots to run real hardware, with FPGA basically only doing core chipset functionality.
2) integrate everything into a bigger FPGA and do it mITX-style, maybe with a single ISA slot.

Combining the two makes it unnecessarily expensive and won't satisfy everybody. Moreover, Voodoo2 and 486? Seriously, no Voodoo2 software runs adequately on any 486-like CPU, so this is a waste of money and FPGA space. Save that for a newer system or - better - stick it standalone on a PCI card. In fact, same applies to all that video stuff: make it a standalone card. The only thing that would be sensible to integrate onto the motherboard is I/O as that would let you populate the ATX backplate. I'd suggest sticking a CF-card slot in the backplate too, linked directly to onboard IDE.

FPGA based motherboards are in my todo list, hard part is the BIOS that I have actually slowly began working on.

- - -

dionb wrote on 2020-03-05, 19:31:

I'd suggest sticking a CF-card slot in the backplate too, linked directly to onboard IDE.

That's what my diagram shows...

Am I understanding right, that you're hoping to contract someone else to do the design if you can stir up enough interest? Because (I don't mean this to be offensive), you don't sound like you're very confident with the hardware design aspect.

I'm going to throw my 2c in, which agrees with other sentiments here:

I would potentially buy a recreated 486 motherboard that had a mATX footprint. ~$200 is reasonable. It's a complicated project in small quantities, I'm willing to pay a bit more for that. North of $200 and I can't justify it except as a "support" thing, and even then it's difficult to reconcile. I understand this is a considerable undertaking, but that's what market analysis is for. We may love the idea, but have to agree that it's not financially viable. Such is life.

In terms of hardware:

Go with a standard motherboard design. ISA, VLB or PCI, perhaps on-board local bus IDE and I/O -- and that's it. I do think a second variant with on-board local bus SVGA and SB16-compatible audio would be dandy, but I don't think the market will support it. Too many people would want to customize the video and audio, and/or wouldn't be willing to pay for the additional FPGA resources and development. I do support those peripherals becoming a side project on standard add-in cards that could be used with this motherboard, though.

Drop the USB unless you want to build in a USB-to-PS2 bridge for KB and mouse (and then only if it's trivial to do, and stupid simple *and reliable* to assign USB peripherals to virtual PS/2 devices.) Actual pass-through USB on a 486 is pointless. DOS, Win 3.x, and Win95 don't support it. Late OEM 95 will use it... but only barely. Even 98's support is fairly minimal. Anything beyond 98 won't run well on a 486. So, not useful at all.

Personally, I would rather not have on-board storage, since plenty of existing solutions exist, but it wouldn't be a huge turn-off either. IMO, CF is going the way of the dinosaur. SD is probably a better option for longevity. The BIOS would have to have some way of limiting large SD cards to a 486-size chunk. One idea might be to support exposing partitions on the SD card as the whole disk to the OS. Mapping partitions to IDE channels would be useful. Either way, a method of completely disconnecting the flash storage from the IDE bus is crucial, so those of us with spinning rust can keep on spinning rust.

The BIOS is going to be the biggest hurdle, I think. (Not that the board design is a cake walk, but...) If you don't have someone ready, willing, and technically capable of providing it, this is dead in the water.

SirNickity wrote on 2020-03-05, 23:22:

Am I understanding right, that you're hoping to contract someone else to do the design if you can stir up enough interest? Because (I don't mean this to be offensive), you don't sound like you're very confident with the hardware design aspect.

Yeah; I'm leaping before I look. I have pinouts, diagrams, and ideas, but I can't design FPGAs or boards. I need help and I'm reaching out for it.

I'm confident this can be done, but I'm going to need a team. So I guess this means: #teamNu486 !

SirNickity wrote on 2020-03-05, 23:22:

Personally, I would rather not have on-board storage, since plenty of existing solutions exist, but it wouldn't be a huge turn-off either. IMO, CF is going the way of the dinosaur. SD is probably a better option for longevity. The BIOS would have to have some way of limiting large SD cards to a 486-size chunk. One idea might be to support exposing partitions on the SD card as the whole disk to the OS. Mapping partitions to IDE channels would be useful. Either way, a method of completely disconnecting the flash storage from the IDE bus is crucial, so those of us with spinning rust can keep on spinning rust.

The BIOS is going to be the biggest hurdle, I think. (Not that the board design is a cake walk, but...) If you don't have someone ready, willing, and technically capable of providing it, this is dead in the water.

SD is a bit too large (storage wise) and CF natively uses IDE. Well, PATA anyway. Besides, this way I could use XTIDE Universal BIOS.

SirNickity wrote on 2020-03-05, 23:22:

I would potentially buy a recreated 486 motherboard that had a mATX footprint. ~$200 is reasonable. It's a complicated project in small quantities, I'm willing to pay a bit more for that. North of $200 and I can't justify it except as a "support" thing, and even then it's difficult to reconcile. I understand this is a considerable undertaking, but that's what market analysis is for. We may love the idea, but have to agree that it's not financially viable. Such is life.

Nah. I hope to make, at the very least, about nine hundred.

If anyone can help, please let me know 😀

Nitroraptor53 wrote on 2020-03-06, 02:03:

Nah. I hope to make, at the very least, about nine hundred.

Are you serious about this?

keenmaster486 wrote on 2020-03-06, 14:39:

Nitroraptor53 wrote on 2020-03-06, 02:03:

Nah. I hope to make, at the very least, about nine hundred.

Are you serious about this?

I meant 900 boards, over the course of a year or two, most of them as DIY kits or just boards.

Nitroraptor53 wrote on 2020-03-06, 15:05:

I meant 900 boards, over the course of a year or two, most of them as DIY kits or just boards.

Lol! I thought you meant you wanted to make $900 per board.

keenmaster486 wrote on 2020-03-06, 16:16:

Nitroraptor53 wrote on 2020-03-06, 15:05:

I meant 900 boards, over the course of a year or two, most of them as DIY kits or just boards.

Lol! I thought you meant you wanted to make $900 per board.

Oh, of course not! Anywhere from $130-$370 could be the price.

Given that the cheapest NuXT is $280, I don't see how you can make a new cheap 486 motherboard. If you can't get an existing 486 chipset, creating one from scratch in FPGA would probably be prohibitive and I don't think ao486 performance is what people are looking for when they build a 486:
https://www.youtube.com/watch?v=G32lGUVeodQ
Personally, I would buy only such board if it had performance comparable to the late 486 PCI boards. On the hardware side it would need at least several PCI and ISA slots, so micro-ATX is a no-go. 486 Baby-AT boards had often as much as 7 slots, including long VLB ones.