(also, no video output at all)

IBM 486 board troubleshooting

Led to another thread.
Dunno if it is useful

Could you share the BIOS ROM dump from PC-1? Then I could find the code that outputs 0xAA.

kmeaw wrote on 2024-10-12, 11:08:

Could you share the BIOS ROM dump from PC-1? Then I could find the code that outputs 0xAA.

Attached! Thank you! 😀

myne wrote on 2024-10-12, 11:00:

IBM 486 board troubleshooting

Led to another thread.
Dunno if it is useful

The POST status in the post you're referencing seems to be the same (00 AA) but I've got an original IBM power supply hooked up to the board, so I don't think it's related to the 3-pin Rapid Resume connector.

I don't think the POST card is expected to be on port 0x80.

The code that interacts with this port is located at F000:752C. It reads 4 bytes from F000:751B. For every byte read, it writes it to ports 0x81..0x8E, then reads them back (and freezes the machine if it doesn't get back the same value).
For some reason, every but the first iteration covers a slightly different port range - 0x80..0x8E, so the first byte at F000:751B (which has the value of 0xFF) doesn't reach your POST card but the next three (0x55 0xAA 0x00) do.
Once the loop is complete, it does a similar write-read-compare thing but with different values and port range - 3 bytes from F000:751F (0xCC 0x33 0x00) would go to ports 0x8F (which would get a byte) and 0x4F (which gets a leftover byte from the previous loop).

Unfortunatelly, I don't know much about non-Award BIOSes, so it doesn't make much sense to me. The code has some reasonable structure - it is definitely not some random data and the 1-byte checksum is correct (all bytes of every half add up to 0).

You can try flashing this one to see if you would get anything on your POST card.

Let us know when you find out what CPU is in PC1. Very curious to find out.

Looks like this motherboard maybe?

https://theretroweb.com/motherboard/manual/34788.pdf

Maybe check the status of J15? Can't quite see it in your photo.

myne wrote on 2024-10-01, 14:32:

Chuck them in the dishwasher without soap.
Dry thoroughly.

Is he talking about you in this video?

https://youtu.be/5j1QfCqnvzI?si=12vE08gmrdfCPMlQ&t=636

Heh
If it's stupid but it works, it's not stupid

douglar wrote on 2024-10-17, 12:55:

Let us know when you find out what CPU is in PC1. Very curious to find out. […]
Show full quote

Let us know when you find out what CPU is in PC1. Very curious to find out.

Looks like this motherboard maybe?

https://theretroweb.com/motherboard/manual/34788.pdf

Maybe check the status of J15? Can't quite see it in your photo.

I think you're right about the motherboard. The cooler is glued to the CPU and I haven't mustered the courage to try to remove it. Underneath it says:

8426921SNB
MALAY
ES 447
INTEL 89 92
i50

kmeaw wrote on 2024-10-17, 12:34:

You can try flashing this one to see if you would get anything on your POST card.

The plot thickens. I was able to read the existing data from the AM27C010 EEPROM, write it back and verify it. When I tried to write your ROM, I got an error from minipro at exactly 50%. My initial thought is that one of the address connections is broken somehow. I've ordered a couple of new AM272C010's, and will report back once they show up and I get a chance to burn them.

saustrup wrote on 2024-10-18, 10:23:

kmeaw wrote on 2024-10-17, 12:34:

You can try flashing this one to see if you would get anything on your POST card.

The plot thickens. I was able to read the existing data from the AM27C010 EEPROM, write it back and verify it. When I tried to write your ROM, I got an error from minipro at exactly 50%. My initial thought is that one of the address connections is broken somehow. I've ordered a couple of new AM272C010's, and will report back once they show up and I get a chance to burn them.

I temporarily borrowed the M28F101 EEPROM from the other dead Aptiva, wrote your test ROM and tested it. Result (what I caught, since it updated quite fast):

- "04", "05", "06", "08", "0F", "10", "12", "13"
- Three low pitch beeps with ~3s between them.
- ~10s pause
- "00"
- Three beeps again, 10s pause, repeat.

Tried with and without RAM. Same result. Tried J15 in both positions. Same result.

IBM boards can be unreasonably picky about RAM. I don't even know what their ideal is sometimes, get apparently very similar modules side by side and one works and one doesn't.

On POST code 0x12 this ROM resets CRT video registers.
On POST code 0x13 it initializes the chipset registers and performs RAM tests.
The next POST fragment (0x14) is expected to set up the timer (8254).
I don't expect this test ROM to succeed as it has been pulled from a different 486 board.

Have you tried writing your backup (aptiva1.img) on that EEPROM?

kmeaw wrote on 2024-10-18, 15:17:

On POST code 0x12 this ROM resets CRT video registers. On POST code 0x13 it initializes the chipset registers and performs RAM […]
Show full quote

On POST code 0x12 this ROM resets CRT video registers.
On POST code 0x13 it initializes the chipset registers and performs RAM tests.
The next POST fragment (0x14) is expected to set up the timer (8254).
I don't expect this test ROM to succeed as it has been pulled from a different 486 board.

Have you tried writing your backup (aptiva1.img) on that EEPROM?

Yes, same result as the original ROM. But since I had issues writing to the original EEPROM, I probably had - unknowingly - issues reading it too. I suspect my backup is bad. From the looks of it, the data looks ok though.

I can't seem to find any reference to my model on https://aptivasupport.com/uas/bios/ - assuming that the cases and case covers for these three motherboards go together, this is a 2144-121 (SL-I). I know at least the COUFLASH.EXE file unpacks to a floppy and contains multiple BIOS ROMs, so I'm guessing at least one of the flash packages must have a ROM for my model - but no idea which one.

https://www.driverguide.com/driver/detail.php … driverid=603204

Maybe?

If not ask this guy for a dump
IBM Aptiva 2144 won't boot from hard disk

Your original ROM has a string "96G0926" which looks like a label IBM puts on ROM chips. Looking for that string on the web leads to https://github.com/MobyGamer/TOPBENCH/blob/ma … er/DATABASE.INI which has a matching entry of IBM Aptiva 350 Model 2168-26P. Doing another web search give me a page with system specs - http://ps-2.kev009.com/pcpartnerinfo/ctstips/e572.htm
aptivasupport.com table lists BG2US0M as the BIOS update for that machine - getting an update and unpacking it results in $imageg2.usf but I'm not sure if it is the right one - it has a BIOS date of 12/16/98

I have a 2144 like pc1. It is a dumpster find. Mine has a Cyrix Cx486 Dx266. I don't know if this is the original CPU. Probably no because i didn't change any jumper and the bios report Dx2 100. I don't own any EPROM programmer to dump and upload the bios. Does anyone know if i can put cache memory on this motherboard?

saustrup wrote on 2024-10-01, 16:04:

myne wrote on 2024-10-01, 14:32:

Chuck them in the dishwasher without soap.
Dry thoroughly.

[...] I'm worried about the salt?

Just FYI, the salty water in the small tank in the bottom is not used (in any significant quantity) for washing - only to wash balls or cubes of this stuff: https://www.duboischemicals.com/water-treatme … resin-beads.jpg which are in a hidden tank next to the one you fill with salt, it only flows that way once every a few runs ("a few" = depends on how you set the water hardness option), at all other times "that stuff" is attracting calcium off the tap water coming in 😀

In fact the manual will tell you to be extremely careful not to forget the cap or getting other stuff in there!

For the same reasons, 3 in 1 soap including salt and rinseaid (that one is mixed into the last water change, well after the regular soap tank opens) is largely useless...