NO POST Nightmare - DFI G7L330-B Industrial Motherboard Socket 775

[techpointhull]

Hi Guys,

Newbie here to board repair so bare with me. I have a bunch of Socket 775 motherboards from arcade gaming machines. The brand is DFI, and it's their industrial motherboard range model number G7L330-B

All boards have the same issue, no POST. I've re-capped the boards completely, can't see any shorts across the 12V rail and I've removed all MOSFETS from the board to test them. The boards have been ultrasonic cleaned as well as an alcohol bath, doesn't seem to have made a difference

BIOS is also correct, I do have a working board for reference and the BIOS chip does indeed work. I'm also testing with a working CPU, and RAM. Using my thermal camera, I can't see anything getting hot on the board that isn't normal.

I have checked for expected voltages, and I am seeing 1V, 1.6V, 1.8V, 3.3V, 5V and 12V across the board so I believe the regulators are working correctly.

This motherboard is pretty restricted so there isn't any schematics online from what I can see, just wondering if anyone could give me some pointers and ideas as I'm stuck now. Sadly, its not a case of just replacing the board, the software the arcade machine uses is locked to these motherboards. (Something in the BIOS).

I'm happy to repeat any tests, just in case I've not done the correct procedure. I have attached a picture for reference of the board. My PCI Post Analyser card just displays 00.

Any help, things to check is greatly appreciated.

James

[прямо]

Re: NO POST Nightmare - DFI G7L330-B Industrial Motherboard Socket 775

Does it complain with memory error beep codes if you leave the DIMM slots unpopulated?

Have you tried memory sticks from a good working board? Some mobo doesn't like SS/high density (chips only on one side) memory sticks. Some are also picky with the SPD detection, and won't POST like for an example if the CPU only supports 100MHz times a multiplier to achieve the desired CPU clock speed, and you're using a faster or a slower memory speed.

How did you know the BIOS is working? Did you dump the BIOS from the board you're working on and compared it with a good BIOS image?

Plenty of mobo doesn't support the PCI POST Analyzer card. You could try a different PCI slot though and see if that makes any difference, but I doubt it.

[techpointhull]

Re: NO POST Nightmare - DFI G7L330-B Industrial Motherboard Socket 775

Thanks for the reply. The board does not beep with or without memory populated. I have an identical working board, and the BIOS chip is not soldered so I can swap the chips easily between the boards, and it works fine on my working board.

The PCI card also works fine on my test board.

Does it complain with memory error beep codes if you leave the DIMM slots unpopulated?

Have you tried memory sticks from a good working board? Some mobo doesn't like SS/high density (chips only on one side) memory sticks. Some are also picky with the SPD detection, and won't POST like for an example if the CPU only supports 100MHz times a multiplier to achieve the desired CPU clock speed, and you're using a faster or a slower memory speed.

How did you know the BIOS is working? Did you dump the BIOS from the board you're working on and compared it with a good BIOS image?

Plenty of mobo doesn't support the PCI POST Analyzer card. You could try a different PCI slot though and see if that makes any difference, but I doubt it.

[techpointhull]

Not sure why previous responses have disappeared from this thread since the site was updated, but has anyone got any suggestions?

[RJARRRPCGP]

Could very well be a variation of "bad socket syndrome".

[CapLeaker]

Take that square chip out of that brown socket and put it back in (reseat). See if she POST's now. Also make sure the bios chip is in the socket correctly.

[techpointhull]

Take that square chip out of that brown socket and put it back in (reseat). See if she POST's now. Also make sure the bios chip is in the socket correctly.

Yeah the chip in the brown socket is the BIOS chip, this has been re-seated as well as replaced from a known good motherbord, contacts also cleaned.

[techpointhull]

Could very well be a variation of "bad socket syndrome".

I am starting to think this is the case with these boards... they were running 24/7 for years in an arcade many of which overheated. 😕

[CapLeaker]

I am starting to think this is the case with these boards... they were running 24/7 for years in an arcade many of which overheated. 😕

I think it is the socket at the moment.