I done even know how is this supposed to work. You need special IRQ arbitration (IOAPIC) for multi CPU systems.
https://en.wikipedia.org/wiki/Advanced_Progra … rupt_Controller https://wiki.osdev.org/IOAPIC
The only way this board could operate is as normal single CPU board with slave processor. There will be zero software support for this arrangement.

some old expired ebay auction for https://picclick.com/ECS-Elitegroup-AC486-486 … 2833237298.html single CPU variant https://theretroweb.com/motherboards/s/ecs-ac486-pqfp , shows wrong photo, but crucially this
"BIOS STRING: 40-0401-001131-00101111-121291-ACC2046-F"

EDIT: looking at jumper document we get this
"80486SX (primary) N/A Open Pins 2 & 3 closed
80487SX (upgrade) Pins 1 & 2 closed Closed N/A
80486DX (primary) N/A Open Pins 1 & 2, 3 & 4 closed
80486DX (upgrade) Pins 2 & 3 closed Closed N/A
80486DX2 (primary) N/A Open Pins 1 & 2, 3 & 4 closed
80486DX2 (upgrade)
Primary CPU is CPU1 & upgrade CPU is CPU2."
Really confusing, cant tell if its some attempt at non standard SMP system or just really stupid two socket arrangement. Why would anyone need two sockets to upgrade from one 486DX2 to another 486DX2 without pulling out the first one?

rasz_pl wrote on 2024-01-21, 18:30:

I done even know how is this supposed to work. You need special IRQ arbitration (IOAPIC) for multi CPU systems. https://en.wikip […]
Show full quote

I done even know how is this supposed to work. You need special IRQ arbitration (IOAPIC) for multi CPU systems.
https://en.wikipedia.org/wiki/Advanced_Progra … rupt_Controller https://wiki.osdev.org/IOAPIC
The only way this board could operate is as normal single CPU board with slave processor. There will be zero software support for this arrangement.

some old expired ebay auction for https://picclick.com/ECS-Elitegroup-AC486-486 … 2833237298.html single CPU variant https://theretroweb.com/motherboards/s/ecs-ac486-pqfp , shows wrong photo, but crucially this
"BIOS STRING: 40-0401-001131-00101111-121291-ACC2046-F"

EDIT: looking at jumper document we get this
"80486SX (primary) N/A Open Pins 2 & 3 closed
80487SX (upgrade) Pins 1 & 2 closed Closed N/A
80486DX (primary) N/A Open Pins 1 & 2, 3 & 4 closed
80486DX (upgrade) Pins 2 & 3 closed Closed N/A
80486DX2 (primary) N/A Open Pins 1 & 2, 3 & 4 closed
80486DX2 (upgrade)
Primary CPU is CPU1 & upgrade CPU is CPU2."
Really confusing, cant tell if its some attempt at non standard SMP system or just really stupid two socket arrangement. Why would anyone need two sockets to upgrade from one 486DX2 to another 486DX2 without pulling out the first one?

Thank you for turning this to a constructive direction.
I've just read the same manual and totally can't get what's the point of having two PGA sockets when using only one CPU at once. Maybe only for the case when user get theirself a computer with pre-installed SX processor and later going to upgrade it to DX/DX2 but doesn't know how to pull a processor or wants to switch between two sometimes... (this probably could be a reasonable option when buying or testing only 2 modules of expensive high-capacity RAM)

If so, probably I really can try any similar BIOS, since another manual says the following:

1. Revision C.0 motherboards may have the CPU installed in the upgrade socket (U45) instead of the CPU
socket (U47), and will have to be removed in order to upgrade.

, so this 2xPGA version seems to be essentially a version of pre-soldered CPU + PGA upgrade socket motherboards but with double PGA instead.
But I don't think the given POST string will fit since it's for another chipset, ACC 2046.

ebay seller claimed thats what ecs-ac486 was reporting. At this point its the only clue as to ecs-ac486 bios string on the whole internet. That string leads to https://theretroweb.com/motherboards/s/unknown-acc2046 seemingly the only motherboard ever with ACC2046 chipset. Most likely ACC2186 is software compatible with older sibling and ECS was too lazy to modify old bios. With any luck some other ACC bios will work for you.

rasz_pl wrote on 2024-01-21, 19:08:

ebay seller claimed thats what ecs-ac486 was reporting. At this point its the only clue as to ecs-ac486 bios string on the whole internet. That string leads to https://theretroweb.com/motherboards/s/unknown-acc2046 seemingly the only motherboard ever with ACC2046 chipset. Most likely ACC2186 is software compatible with older sibling and ECS was too lazy to modify old bios. With any luck some other ACC bios will work for you.

Sounds reasonable. Will try to start it with few ACC images after cleaning it up and replacing the resistor array and the BIOS socket.

rpocc, be sure to stock up on some capacitors.
There will surely be a tantalum on that board that throws a tantrum. 😀

this looks like a good candidate https://theretroweb.com/motherboards/s/juko-at046dx3#bios
ACC Micro 2168 with "30-0004-428002-00101111-070791-ACC2046-0" bios string

rasz_pl wrote on 2024-01-21, 19:31:

this looks like a good candidate https://theretroweb.com/motherboards/s/juko-at046dx3#bios
ACC Micro 2168 with "30-0004-428002-00101111-070791-ACC2046-0" bios string

No idea if it matters, just an observation, that board carries a 386 cpu.

rasz_pl wrote on 2024-01-21, 19:31:

this looks like a good candidate https://theretroweb.com/motherboards/s/juko-at046dx3#bios
ACC Micro 2168 with "30-0004-428002-00101111-070791-ACC2046-0" bios string

I've tried this BIOS, and it starts, loads the VGA BIOS, starts to count RAM, finishes with that correctly, shows the "Wait..." message and hangs up. Post card shows 0x00 as the last code. IIRR, it's CPU halt. So, i'm not sure if the BIOS has partial incompatibility, or the MB is just broken.

The keyboard only blinks (appears to be suspiciously fast) at the very start, then does not react to any keypresses, so you can't interrupt the RAM test or do anything. I see no signs of broken traces and even tried another 8042 chip with exactly the same results.

Moreover, this board is very selective about memory modules. The first ones which at least confidently complete the RAM test are 60ns 4MB modules.

Actually I can't see any other candidate that I could try. Few other pages does not contain BIOS images, and packard-bell images are 128K and it's not some simply glued-together machine and VGA images.

weedeewee wrote on 2024-01-21, 19:46:

rasz_pl wrote on 2024-01-21, 19:31:

this looks like a good candidate https://theretroweb.com/motherboards/s/juko-at046dx3#bios
ACC Micro 2168 with "30-0004-428002-00101111-070791-ACC2046-0" bios string

No idea if it matters, just an observation, that board carries a 386 cpu.

Even the file name is acc386.bin, however the shown CPU looks more like 486sx to me.

Normally 00 means BIOS finished initialization successfully and will now try loading bootloader (int 19h).

Btw that last letter of bios string made me remember something https://alexandrugroza.ro/microelectronics/es … bios/index.html
Ebay "BIOS STRING: 40-0401-001131-00101111-121291-ACC2046-F" ends in -F, but juko one with -0 so it might expect different keyboard controller, thus non working keyboard. You might need to either swap keyboard controller or try to patch bios code

Almost certain this was made for the 486sx/487 upgrade chip marketing by Intel.

When using the 487 upgrade chip, the original is supposed to be disabled (pin C11 on the primary CPU). The 487 is mostly pin compatible with 486DX/DX2 CPUs as well. So I think the board designer gave the user two full sockets and a lot of options, but they don't actually work at the same time. One CPU is always disabled...

rpocc wrote on 2024-01-23, 14:16:

I've tried this BIOS, and it starts, loads the VGA BIOS, starts to count RAM, finishes with that correctly, shows the "Wait..." message and hangs up.
The keyboard only blinks (appears to be suspiciously fast) at the very start, then does not react to any keypresses

well, what do I know
"This poor motherboard needs my help! TGM #4" - Tony359 https://www.youtube.com/watch?v=yVg2pvyYbn0&t=805
Similar case, here was getting stuck on Wait... because of Cache enabled in BIOS by default.
Looks like you will need to figure out compatible keyboard controller situation for this bios first.

Or maybe chkcpu can be bothered to have a look at it and possibly patch it 😉

Reporting progress so far:

1. I have spent several hours analyzing assembly code with IDA Pro, which is a separate torture because of so-called ROM calls (A method of calling functions without using any RAM: xxxx:xxxx mov sp, return_addr_ptr // jmp function_addr), which prevents the whole engine from tracing the correct call graph. This image doesn't contain any I/O operations with ports 60h and 64h related to detection of the keyboard controller version, so anyway, I can't see any practical way nor reason to patch the BIOS.
2. I've found a chip JetKey v3.0, which is exactly the same as shown on the picture of JUKO Motherboard and once was able to pass the memory test and enter BIOS setup. By the way, the cache is disabled in default settings. However, when I restarted machine after that, I've got several hangs at POST code 13h, sometimes with beeps, sometimes not, later got hangs up at memory tests, and later got the same hang-up at Please wait... state with the same last POST code 00h. So, apparently the root of the problem is either a hardware malfunction causing RAM errors (including bad caps and bad line drivers).
3. I've tried to start the board again, with the old keyboard controller and once, after being long enough unpowered, it has passed the memory tests and loaded up to the message "CMOS Battery low, etc". After next restart, it keeps hanging at 13h and buzzing.

I'm not sure but maybe the problem also may have something to do with the absent battery, which is unusual but who knows how this ACC works.
Anyway, I think the BIOS is OK for this motherboard and simply it has hardware issues, which are hard to detect. This board isnt' so good in terms of computing performance but electrically it's absolute magic: it has no even a single electrolytic caps: all SMD ceramic and tantalum.

Interesting thing: when I was reading the assembly code I've found a great document called Ralf Brown's Interrupt List, which has a file PORTS A containing a list of AMI post codes which I can confirm to be really correct for older boards, unlike most of the lists on the internet, which seem to be from a totally different generation. So I recommend anyone to download RBIL and use as a reference for debugging old hardware.

At the end of the day I'm afraid I'm stuck with a malfunction which I can't solve using my current experience and probably it doesn't deserve so much time to spend.

rpocc wrote on 2024-01-27, 13:31:

1. I have spent several hours analyzing assembly code with IDA Pro

lovely stuff 😀

rpocc wrote on 2024-01-27, 13:31:

This image doesn't contain any I/O operations with ports 60h and 64h related to detection of the keyboard controller version, so anyway, I can't see any practical way nor reason to patch the BIOS.

so it has hardcoded particular keyboard controller, but since you

> I've found a chip JetKey v3.0

then patching is indeed not necessary

rpocc wrote on 2024-01-27, 13:31:

once was able to pass the memory test and enter BIOS setup. By the way, the cache is disabled in default settings. However, when I restarted machine after that, I've got several hangs at POST code 13h, sometimes with beeps, sometimes not, later got hangs up at memory tests, and later got the same hang-up at Please wait...

> I've tried to start the board again, with the old keyboard controller and once, after being long enough unpowered, it has passed the memory tests and loaded up to the message "CMOS Battery low, etc". After next restart, it keeps hanging at 13h and buzzing.

That sounds random and erratic, the worst kind of defect. Might be weak/broken trace/via somewhere randomly flipping between connection/break. Good once over with a loupe/microscope, cold spray (canned Air), percussive maintenance (poking at it with a wooden/esd safe plastic stick, slightly bending) are all good troubleshooting steps.

>including bad caps
the only bad caps on 486 can be burning and shorting tantalums, electrolytics wont give you trouble unless its very late 486 board with leaked SMD ones.

>no even a single electrolytic caps: all SMD ceramic and tantalum.

smd ceramic cap failure mode is short circuit, you would notice not powering up or smoking 😀, can disregard those.

>Anyway, I think the BIOS is OK for this motherboard and simply it has hardware issues, which are hard to detect.

that is also what I think

>Interesting thing: when I was reading the assembly code I've found a great document called Ralf Brown's Interrupt List, which has a file PORTS A containing a list of AMI post codes which I can confirm to be really correct for older boards, unlike most of the lists on the internet, which seem to be from a totally different generation.

I use this list https://mrbios.com/techsupport/award/postcodes.htm

rasz_pl wrote on 2024-01-27, 18:43:

I use this list https://mrbios.com/techsupport/award/postcodes.htm

Just try this link:

https://github.com/cirosantilli/ralf-brown-in … nter61d/PORTS.A

At line 7837 you’ll find the list which is not provided in the document which you have referenced.
I’ve traced down the execution of my BIOS up to warm restart after final UMB test and can guarantee that at least for this version of AMI BIOS the list is precise.

So, that can be a neat addition to the onowledge base.

I’ll give the board another try with some scope line tests later. So far I feel totally tired in my struggle to get in back on its feet.

I used bioscentral before the site went to a parked domain crud a few weeks ago.
So now it means always going to the archive for the info.
https://web.archive.org/web/20230925005155/ht … des/amibios.htm

It has several versions of the post codes for ami award etc...