Maybe the TAG chip is causing the issue. Remove it, spray some contact cleaner on the socket and put the chip back on.

IntelInsidest wrote on 2025-09-22, 22:54:

Also check if all the legs of the PCISet chips are solid and soldered to pads, and cache chips too as you are at it.
Any cut traces on the back of the board.

I wouldn't suspect fake chips on an Asus board and not when the traces are in place for the cache.

I'd echo the suggestions above to check for continuity for all pins for the cache, as well as loose pins. You could also check voltages and make sure the chips are getting the appropriate voltage.

Perhaps also check any adjacent components (capacitors, resistors) for any signs of failure.

Thanks for the suggestions!

I went through all the pins of the chips and everything seemed solid. The tag chip and its socket got a squirt of contact cleaner and quick brushie-brushie.

I took out my scope and open the data sheet for W25P010AF-8 and sure enough everything seems to be there. I/O and supply voltages have a solid 3.3V, host clock is present and even data I/O pins show some serial data. Same applies for both of the chips, but one difference I noticed is that the one on the top in the image gets hot, while the lower one stays at room temperature.

What I'm now curious about is that why does it look like there data I/O on the chips while POST reports there's no external cache?

I even tried all BIOS roms I could find (0105, 0118 and 0205) just to see if that'd make a difference, but no luck.

What an annoying fault. I like the system very much and thought it'd be my ultimate Pentium/3Dfx DOS/W95 machine and could excuse me to get rid of a whole bunch of miscellaneous parts 😁

IntelInsidest wrote on 2025-09-27, 08:21:

Thanks for the suggestions! […]
Show full quote

Thanks for the suggestions!

I went through all the pins of the chips and everything seemed solid. The tag chip and its socket got a squirt of contact cleaner and quick brushie-brushie.

I took out my scope and open the data sheet for W25P010AF-8 and sure enough everything seems to be there. I/O and supply voltages have a solid 3.3V, host clock is present and even data I/O pins show some serial data. Same applies for both of the chips, but one difference I noticed is that the one on the top in the image gets hot, while the lower one stays at room temperature.

What I'm now curious about is that why does it look like there data I/O on the chips while POST reports there's no external cache?

I even tried all BIOS roms I could find (0105, 0118 and 0205) just to see if that'd make a difference, but no luck.

What an annoying fault. I like the system very much and thought it'd be my ultimate Pentium/3Dfx DOS/W95 machine and could excuse me to get rid of a whole bunch of miscellaneous parts 😁

Usually they get hot under use, not at idle.
Could be a bad chip (they can die).

Like Shponglefan suggested:
any capacitor / resistance around getting hot as well? If you have some isopropyl alcohol you can splash the area and check what evaporates first.

Maybe something is pulling to ground and not allowing L2 sizing (my guess yet again).

Nexxen wrote on 2025-09-27, 08:29:

Usually they get hot under use, not at idle. Could be a bad chip (they can die). […]
Show full quote

Usually they get hot under use, not at idle.
Could be a bad chip (they can die).

Like Shponglefan suggested:
any capacitor / resistance around getting hot as well? If you have some isopropyl alcohol you can splash the area and check what evaporates first.

Maybe something is pulling to ground and not allowing L2 sizing (my guess yet again).

No caps get extremely hot. I also measured all of them in that general area in circuit with power off and couldn't find any shorts or open circuits.

I also measured from ground (VSS) to power supply (VDD) on both of the chips and there's no short there either. Resistance starts to ramp up from sub-100 to ~300 ohms which to me looks like a healthy coupling capacitor charging up.