I don't think it's freezing at IDE detection, that part is done at the point where it complains about the time and date not being set. Assuming you also can't press F1 to go into the BIOS settings and set the date/time, it may need the bios re-flashed due to age.

NeoG_ wrote on 2025-12-08, 00:23:

I don't think it's freezing at IDE detection, that part is done at the point where it complains about the time and date not being set. Assuming you also can't press F1 to go into the BIOS settings and set the date/time, it may need the bios re-flashed due to age.

Thank you for helping. You’re probably right that the freeze doesn’t actually happen during IDE detection.
I can actually access the BIOS setup by pressing F1, and the settings page loads normally. The date/time error appears only because I recorded the video right after clearing the CMOS.
I’m keeping the BIOS reflash in mind, but I’ll only try it once I’ve exhausted everything else, since I don’t have an EPROM programmer and the BIOS for this system would probably be extremely hard to find.

In the end, it was... a capacitor. I spotted it with a thermal camera because it wasn't swollen at all, but it was hot.
It's strange because the behavior was always the same, freezing at the same moment, whether hot or cold. And shortly after I was able to install Windows 95. I've never seen a capacitor cause this kind of problem.
Maybe it was something else I fixed by manipulating the motherboard, because it doesn't look like a capacitor problem.
In any case, it's running perfectly! Now I'm going to open up the original power supply and see what I can do.
Maybe I should dump the BIOS and send a photo of the motherboard to Theretroweb because there doesn't seem to be anything about it on the Internet.

JustJulião wrote on 2025-12-11, 09:29:

In the end, it was... a capacitor.

I was thinking about it, although not so sure since your list of solutions covers practically the most important things to boot properly.

That date and keyboard issue, sounds something phisically, and if it's not the battery...

Hello Juliao,

That is not an old Olivetti as Olivetti stopped producing custom machines and went full generic OEM a few years after the release of that one. And it shouldn't be undocumented, it's a common slim box.
What is the exact board model, e.g. BAxxxx? You should have stickers on the board indicating so. The board looks like Trigem Milano/Torino apart from the positioning of RAM slots, which is in your case much more accessible, allowing RAM insertion without removing the top case bridge.

zb10948 wrote on 2025-12-11, 12:03:

Hello Juliao,

That is not an old Olivetti as Olivetti stopped producing custom machines and went full generic OEM a few years after the release of that one. And it shouldn't be undocumented, it's a common slim box.
What is the exact board model, e.g. BAxxxx? You should have stickers on the board indicating so. The board looks like Trigem Milano/Torino apart from the positioning of RAM slots, which is in your case much more accessible, allowing RAM insertion without removing the top case bridge.

Yes I meant "old" generally speaking, not for an Olivetti.
Even if most topics are about old PCs here anyway.
I didn't dig too deep, but I tried to find the motherboard by BAxxxx. It's the BA2324.
I didn't succeed.
The commercial model name is Augustus 133 SL.
Although the chassis resembles common Olivetti models (which I also have), this one has a different facade from most.
I also tried to find the motherboard based on its characteristics on Theretroweb:
Socket 7, 430tx (by the looks of it), integrated ATi Mach64.
No luck.

EDIT: chipset is a i430HX.