Intel Atom C2000 bug - help locate LPC_CLKOUT0 on PCB

[giosif]

Hi,

I've recently acquired a McAfee FW-325 unit in non-working condition and, upon inspection, I think it is suffering from the well known Intel Atom C2000 bug.
Not wanting to give up on this unit, I am trying to apply the workaround for this bug - i.e. add a pull-up resistor between LPC_CLKOUT0 and the 3.3V supply rail.

Problem is I don't know where on the board LPC_CLKOUT0 goes.

So far, I've tried probing the trace pads near the CPU (mostly those which are unpopulated or with resistors on them), as well as any pin headers or pin header slots (i.e. unpopulated) on the board, in the hopes I can detect a 25 MHz signal (a weak one should still be present, based on Dave's video on the Synology NAS unit which had a similar problem).
But had no luck so far, so I'm starting to think maybe I could locate the signal from LPC_CLKOUT0 starting from the BIOS IC instead (to my limited understanding, the signal from LPC_CLKOUT0 ultimately controls how the CPU accesses the BIOS code from the EEPROM IC).
However, I am treading unknown waters here and I wanted to ask for any advice or indications if my approach makes sense and/or if I could locate LPC_CLKOUT0 somehow else.

Below are pictures with the entire board, the CPU and two ICs, one of which I suspect hold the BIOS code.

Thank you!

[amyk]

I'm not too familiar with the Atom C2000 but that looks like an SPI BIOS which would not be connected directly to the LPC; you'll need to figure out what other ICs are on the board and which ones are connected to the LPC.

The "Customer Reference Board Guide" here has a block diagram which might be close to your system:

https://www.intel.pl/content/dam/www/public/us/en/documents/guides/atom-c2000-crb-guide.pdf

...and it shows an SPI BIOS with miscellaneous other stuff on the LPC.

[giosif]

Ok, thank you!
I will do some reading and come back with any findings/updates.

[YetAnotherTechie]

I remember that the fix required 2 resistors, not just one?
Also, the fix works if its done early, as the board it's used the metal migrates inside the transistors and they stop working.
At that point you would need an external 25Mhz generator applied to the pins.

[Monkeh]

I remember that the fix required 2 resistors, not just one?
Also, the fix works if its done early, as the board it's used the metal migrates inside the transistors and they stop working.
At that point you would need an external 25Mhz generator applied to the pins.

No, just one, and no, it works after the fact - that's the point.

[YetAnotherTechie]

I remember that the fix required 2 resistors, not just one?
Also, the fix works if its done early, as the board it's used the metal migrates inside the transistors and they stop working.
At that point you would need an external 25Mhz generator applied to the pins.

No, just one, and no, it works after the fact - that's the point.

There are 2 lines, and it will depend on motherboard implementation if it uses one, the other or both. My supermicro board, fixed by supermicro under warranty had at least 2 resistors added.
Intel Spec Sheet:
"AVR54. System May Experience Inability to Boot or May Cease Operation
Problem: The SoC LPC_CLKOUT0 and/or LPC_CLKOUT1 signals (Low Pin Count bus clock outputs) may stop functioning."

[Monkeh]

I remember that the fix required 2 resistors, not just one?
Also, the fix works if its done early, as the board it's used the metal migrates inside the transistors and they stop working.
At that point you would need an external 25Mhz generator applied to the pins.

No, just one, and no, it works after the fact - that's the point.

There are 2 lines, and it will depend on motherboard implementation if it uses one, the other or both. My supermicro board, fixed by supermicro under warranty had at least 2 resistors added.
Intel Spec Sheet:
"AVR54. System May Experience Inability to Boot or May Cease Operation
Problem: The SoC LPC_CLKOUT0 and/or LPC_CLKOUT1 signals (Low Pin Count bus clock outputs) may stop functioning."

Very true, but he did ask about only one of them. A simple board like this is unlikely to use both.

E: Well, actually, I suppose it might depending on how that SI/O is implemented..

[giosif]

Thank you both!
Yes, it is possible both pins need fixing but, at the moment, I would be happy to find just the one.