I can't speak to the damage on that particular Advantech board or the attempted repairs. However I did recently get a PCA-6166 board working that was having issues. I found it did not like having the CMOS clear jumper (J1) unpopulated, which looks to be the case in your photo. I would jumper pins 1 and 2, if that hasn't been done already. The manual states you can populate the DIMM slots in any configuration, so not sure why it only likes DIMM 1 populated. The PCA-6166 I have DOES require that the first DIMM slot is populated first before any others.

It should boot without a CMOS battery, it just won't hang on to BIOS settings for very long after power-off.

My board would boot but would get stuck at 'verifying DMI pool data'. The only way I was able to get it to complete the boot process was to go through the Award bootblock recovery procedure. This involves shorting two pins on the BIOS chip (in my case using pins 2 and 3 worked). You'll have to peel away the BIOS version sticker on top and look at a datasheet for your chip to see which are two high address (Ax) pins you can short to trick the board into thinking the BIOS is corrupt. Short the two pins as you're applying power. Remove the short as soon as the floppy engages. It will boot a floppy if you're lucky and from there you can re-flash the BIOS (uniflash.exe should work). My board has some sort of issue where this condition will return from time to time and I'll have to use this procedure each time to get it to boot normally again. You'll have to source a backplane to do this, and swap your vga cable from the on-board VGA to an ISA VGA card. If you can't get a backplane, you'll have to either find a way to get the boot disk to automate flashing the bios or memorize the keystrokes required on the keyboard, because the on-board video will likely output nothing.

Contact Advantech support. I was surprised that they responded to me reaching out. Though noting that the board was long out of support, they still sent me working BIOS files to flash. They may have some insight on the behavior and beep codes you are getting.

sizzlinbeef wrote on 2025-04-03, 20:20:

I found it did not like having the CMOS clear jumper (J1) unpopulated, which looks to be the case in your photo. I would jumper pins 1 and 2, if that hasn't been done already.

I don't remember trying this, so I'll try it

sizzlinbeef wrote on 2025-04-03, 20:20:

The only way I was able to get it to complete the boot process was to go through the Award bootblock recovery procedure. This involves shorting two pins on the BIOS chip (in my case using pins 2 and 3 worked). You'll have to peel away the BIOS version sticker on top and look at a datasheet for your chip to see which are two high address (Ax) pins you can short to trick the board into thinking the BIOS is corrupt. Short the two pins as you're applying power. Remove the short as soon as the floppy engages. It will boot a floppy if you're lucky and from there you can re-flash the BIOS (uniflash.exe should work). My board has some sort of issue where this condition will return from time to time and I'll have to use this procedure each time to get it to boot normally again. You'll have to source a backplane to do this, and swap your vga cable from the on-board VGA to an ISA VGA card. If you can't get a backplane, you'll have to either find a way to get the boot disk to automate flashing the bios or memorize the keystrokes required on the keyboard, because the on-board video will likely output nothing.

I didn't know there was such a method. I'm grateful that you told me the specific steps! I'll give it a try.

J1 seems to have been correct. I tried removing it, but the operation remained the same.

I tried the trick of shorting the address pins, but accidentally shorted WE and A17, which (probably) corrupted the ROM data (a long 2-second beep was repeated many times).
It was a W29C020, so I'll prepare to do the hot swap trick on a machine I have (buy an adapter, etc.) (maybe I should take this opportunity to buy a TL866?)

Sorry to hear about the incorrect shorting.

Regarding J1, it shouldn't ever be completely removed, just adjusted from 1-2 to 2-3 when you want to clear CMOS. Since you're going battery-less you don't really have to worry about clearing the CMOS, it does it on it's own when you power off for more than a minute or two.

https://advdownload.advantech.com/productfile … Manual_ed.4.pdf

" The PCA-6276 CPU card contains a jumper that can erase CMOS data
and reset the system BIOS information. Normally this jumper should be
set with pins 1-2 closed. If you want to reset the CMOS data, set J1 to
2-3 closed for just a few seconds, and then move the jumper back to 1
2 closed. This procedure will reset the CMOS to its default setting."

Why the pins on the fan connector on top-right (second photo) are shorted?

These SBC's can take a 3 pin cable from the power supply or backplane to simulate ATX soft power-off functionality. It's the same connector as the fan headers, but totally different function. You jumper those pins to make the board power on as soon as power is supplied if you don't have the special connector. And sorry kagura, the angle of your photo was making me thing half of J4 was J1 and I see now that you have J1 properly set to pins 1-2 and J4 (which I see now is 6 pins) unpopulated. You may want to try the bootblock procedure again with the correct pins (same chip as mine, 2 and 3 should work) just in case it's still salvageable. Won't hurt at this point.

I finally contacted support yesterday. I asked to be provided with the BIOS file, but was told that it no longer even exists at their headquarters.
I think I'll put the board away or sell it.
Thank you to everyone who helped me so far.