A few weeks back our Fedora HTPC suddenly died & refused to start again. Basic spec is:
Motherboard - Gigabyte MA78G-DS3H (Rev. 1)
CPU - AMD Phenom X3 8750
Graphics - Asus Radeon HD 4670
Case is a HFX passive-cooling AV case - basically heatpipes & radiators. System is not overclocked & temps have never been a problem.
I concluded that the FSP Zen 400W PSU had failed, as I could boot using a known-good Enermax 650W PSU. The problem seemed to be that the 5VSB rail was low, at approx. 4.2V - seems that is too low for the power-on circuitry to work, though external USB devices got power & I could start the PSU using the "paperclip trick". I got a SilverStone ST40 replacement PSU & all seemed OK...
Then after a day or so the system died again. With either the SilverStone or Enermax I get the green power LED (sometimes the red HDD comes & stays on as well) but no further; however I can still get absolutely nothing from the FSP so conclude it is dead. I get no monitor output (on either PCI-e card or onboard GPU) but the following happens:
CPU heatsink gets hot
Chipset heatsink gets hot
LEDs on DIMMs flash for a few seconds but then go out; DIMM heatsinks stay cold.
In the region of CPU power circuitry between socket & back panel there are 6 approx. 1 cm black cubes labelled "R60" - I understand these are the heatsinks over ferrite core chokes. These get very hot (considering CPU must be at idle) and on one I measured surface temp. of 70C with a digital probe. The blue lacquer on the underside of the board has been discoloured by the heat, centred round the hottest of the R60s.
My take is that one of more of the chokes are overheating & have fried neighbouring capacitors & other power-conditioning components, specifically those supplying power to RAM. I didn't pay any particular regard to this area when the PSU problem arose, so I don't know exactly when the heat damage started & whether it was progressive or not.
So my questions are
Is the M/b really fried or can it be rescued?
What is the likely causality between PSU and M/b problems, or are they completely independent?
In either case what are possible underlying causes so I can stop it happening again
I have tried all usual thinks like stripping out expansion cards & peripherals, bread-boarding etc. As noted above, case is passive cooling but has reasonable convective airflow & temps have always been well within default shutdown thresholds. If the M/b really is dead it will be a problem as I will probably have to upgrade CPU & RAM as well - MA78G-DS3H seems unavailable now and AM2+ / DDR2 boards that will fit in the case are rare these days...
Heatsinks get pretty warm; unless you burn your finger when touching one of them, the cpu/chipset may be fine. I would also try a different video card, and each ram stick one at a time in dimm slot one. Could be a bad ram stick. But if you still can't get it to work, it's probably time for a new board. If you have the funds, I would go for a new board, cpu, and ddr3. Frys has 8 gb for only $25 after rebate today, if you live near one. If you need a low cost board, this http://www.newegg.com/Product/Product.aspx?Item=N82E168... may work well for you if you can find one. It can run either am2 or am3 cpus up to 95 watts, depending on which ram you use.