BIOS upgrade went wrong on 430FX board

[Zare]

The board is Olivetti BA2270 - https://theretroweb.com/motherboards/s/trigem-milano
Original BIOS version was 1.13. I got 1.30 from the site and flashed it. Reason was to try to increase BIOS HDD limit, I was only able to run less than 8 GB drives in it.

The computer now freezes just as it displays the BIOS banner.

I checked here - https://blog.theretroweb.com/2024/01/20/amibios-beep-and-post-codes-list/#Version_22 , and I guess it could be AMI WinBIOS codes.
I hang at 38 which is "Initialize BUS types", which means 39 "display bus error messages" is never reached.

A problem is that BIOS is soldered onboard so looks I'm like in a bit of a situation.

Any tips?

 

[GiGaBiTe]

I hope you dumped the original BIOS before you flashed that one.

You should never trust BIOS images from TRR, or from anywhere but the manufacturer. Nobody verifies them, so they could be junk. There are also different board revisions that could require different BIOS versions.

Since the board doesn't boot, your only choice is to desolder the BIOS chip and ideally reflash it with the original BIOS image. As for reinstalling it on the board, I'd recommend to install a PLCC socket instead, because it reduces the stress on the board and on the EEPROM. You can just pop the chip out of the socket instead of having to hot air it off.

If you lost the original BIOS, and neither of the BIOS images on TRR work, the only thing left to you is to try cross flashing a similar boards BIOS with the 430FX chipset. The only problem with this approach is I don't see a VGA ROM, which means that it's multiplexed in the same EEPROM with the System BIOS. You may be able to get away with another board's BIOS, but the integrated video is going to be dead without the VGA BIOS.

 

[Zare]

I supposed so. I did not dump it. I of course thought about it and then decided not to do it, what could go wrong, I could only screw the BIOS. And fixing problems is fun.
Well it looks like I screwed the BIOS.

About TRR, the board I have is designated OK, and the files seem to match. The system board is called Trigem Milano (sourced from Olivetti pocket service guides), but there are several revisions, and seems TRR conflates them all into a single entry with two BIOS provided.

In that archive there are files - milano.rom or something like that. The BIOS ID string presented on TRR, and shown after flashing, is the same string I had originally on the working BIOS.

I don't think tracking down the BIOS will be that hard because it's an Olivetti system (PCS5130) thus there may be owners around here or vogons willing to help me. And it's going to take time and space to remove and resocket the eprom.

In any case I can try to disassemble the rom and see what's going on before emission of post code 39. There is limited amount of jumpers on board and it doesn't seem like anything relevant can be set. I of course tried powering it on with CMOS cleared, battery disconnected, bus devices, extra memory, cache card removed...same situation.

 

[Starcat]

I would try disabling the onboard video controller and see what happens if you boot from serial. Can you access the BIOS via serial?

Also, if you haven't removed the PCI/ISA bus riser, I would try that as well.

Have you checked power supply voltages?

 

[Zare]

The voltages seem to be ok judging by the leds on the post card. Also the problem happened immediately after flash so I don't think a PS failure coincidence in same moment is doable.

The onboard Trident is not active. I have a discrete adapter in PCI and its video BIOS is executed because the main BIOS banner in graphical mode appears, with Award logo and everything.

I never accessed a desktop PC bios 'console' that way. Should I just connect other end of serial with 9600bps standard settings and fire up the machine with all graphics cards removed?

 

[Starcat]

I never accessed a desktop PC bios 'console' that way. Should I just connect other end of serial with 9600bps standard settings and fire up the machine with all graphics cards removed?

I've never seen a desktop PC with that functionality either, just servers. But I noticed this in the user manual:

Jumpers J7 and J10 - Boot from Serial Port Disable
J7: Position 1-2 Disables the serial ports
J10: Position 1-2
J7: Position 2-3 Enables the serial ports (Default).
J10: Position 2-3

That's what I would try, 9600-8-N-1 with all graphic cards removed.

 

[Starcat]

Judging from this Olivetti BIOS manual for the M4-82 / PCS52E, it seems to imply that Olivetti PCs could actually boot somehow from the serial port. Perhaps that's what they're actually eluding to and not console functionality.
https://www.ardent-tool.com/Olivetti/Docs/service_guide/systems2/caph2.pdf

Bootstrap From
the Serial Port
Enabled
Disabled
Enables or disables the operating system being loaded from
the serial port.
The Supervisor Password is required in order to modify
these options.

 

[1ST1]

Original BIOS version was 1.13. I got 1.30 from the site and flashed it. Reason was to try to increase BIOS HDD limit, I was only able to run less than 8 GB drives in it.

I think that will not help. Those 4x0FX boards supported only 8 GB, that was the era... (nobody could imagine bigger drives than that at these times...)

I sugest you to use a flash eprom device and reprogram the chip in there. I hope that you also have a backup of the 1.30 Bios. Maybe 1.31 needs another mainboard revision.

If you want to use bigger harddisk, switch to SCSI or use ontrack disk manager. Maybe you also can use a network card with boot rom socket and run XT-IDE BIOS from there. But then you need to disable IDE controller in the mainboard bios or set harddisk type to none.