I have some 486 waiting to be repaired, one was bought 30 years ago, it only worked for a couple of days, and went to service a couple of times, without them having to solve the problem (freeza shortly after starting up), I kept it, and several years ago, I tried to turn it back on.

Honestly, I don't know if the motherboard returned from the assistance is the same one I had sent, that is, if they changed my motherboard with another (?).

Doing the boot tests, I tried the CPU in another PC, and it was broken, putting the one working, I saw the first BIOS screen on video, and then it crashed, every subsequent attempt showed nothing more.

I update the current configuration, instead of the Cyrix DX2-80 (5V) mounted on an intermediate socket, I tried other CPUs, with an AMD DX2-66 I saw the BIOS screen on the video, then an Intel DX-33 and finally an SX-25.

The RAMs are 30 PIN, four of 1 MB each (NEC 2+1 chip), but in this card it is also possible to use a 72 PIN RAM, or both types.

On the paper manual, there are various combinations, in the 30 PIN ones, there are some strange 512 KB and 2 MB modules, but it could be a translation error (at least I think), and all the jumpers are set correctly.

The motherboard is still almost like new, there is only a slight dust, observing it I notice a trace in the battery area, oxidized green, I will have to investigate to understand where it goes, I see that it passes under the BIOS chip, maybe that could be the problem (?).

Unfortunately it will have to be scratched, to make the COPPER underneath appear, fortunately the warranty has expired, since 01/1996!

If I find an interruption, the problem of not starting could be solved, for the original problem, I don't think it was caused by the Cyrix CPU, which maybe started to fail, because with the AMD it happened again ️

After performing some checks, I can rule out that the problem comes from the UMC chipset, by the way the motherboard is a Power Tech MB457, AT Industrial format, usually these boards are sold with U5 CPUs, here maybe in Europe you couldn't sell such CPUs, but there was still the socket (1 or 2?), in which to insert the 486 at 5V, unfortunately it does not have linear controllers, to use 486 CPU at lower voltage, if you wanted to use you needed an interposer with regulators on board.

As anticipated, originally the CPU was a 5V Cyrix Cx486 DX2-80, but it was mounted on an intermediate socket, which in the back has some pins connected to each other, I see three very thin wires of different colors, which bridge two socket pins, from what I know, they serve the Cyrix CPU for the cache, but when the PC worked, the intermediate socket had been removed.

On the fact that the CPU does not work, there would be the same doubt as before, that is, if the motherboard sent for service, has been changed with another, in that case the CPU could also have been changed (?).

The card was bought new, and when it worked it was very fast, compared to other 486 DX2-66, although this card does not have the L2 cache, it is also true that at that time, the fake cache was very widespread, so having it or not was the same, that is, if they were not real cache chips, the performance was comparable.

Maybe, however, the FSB at 40 MHz instead of 33, could make the cards work faster, as VGA ISA I had a 1 MB Trident 8900, all in all not bad for DOS use, with Windows 3.1 only at high resolution you could see a bit of slowness, but I think it was quite normal at that time.