We have a LGA1366 build (about two years old) that would suddenly not turn on about six weeks ago. Power was getting to the motherboard (on-board LED was on) but pressing the power button (whether through the case button or on-board power button) did not turn it on.
I started a process of diagnosis, and eventually found that if I unplugged the 3.5" internal card reader from one of the USB headers, the motherboard turned on fine. So, the problem must have been a short within the card reader. Life goes on.
But two weeks ago, with the card reader still unplugged, the same thing happened again.
This time I stripped EVERYTHING from the build, going as far as to remove the entire motherboard from the case with nothing in it but a stick of RAM, the graphics card and a CPU, and it STILL wouldn't turn on.
In a desperate last attempt, I stuck a couple of paperclips into the correct contact points within the ATX 24v power connector and shorted them to "force" the PC to turn on. It turned on fine, but of course as soon as I broke contact between the paperclips, it would turn off again.
However, if I pressed the power button straight after this, it would turn on normally. Everything works and is rock solid. Everything is plugged into the build. No BSODs. No crashes. If I shut the machine down, the only way to turn it back on again is to short those paperclips again and then very soon after that press the power button.
I do not know what is causing this, and why "jumpstarting" it by doing the paperclip-short would cause the issue to temporarily disappear.