I've been looking over your thread for a while now, it may be your chipset that is causing you problems, the 680i was known for not playing nice with quads and as far as I'm aware the 780i is only slightly different to the 680i.
1.5v vcore is pretty high for a 3.4GHz oc but then again I had to use 1.55v to hit 3.6GHz on the P5k I had.
What is your vdrop and vdroop like??
vdrop is the difference between the bios and windows idle, vdroop is the difference between windows idle and windows load.
Also I see Absynth Haze (I still think of him as Mattie lol) mention that you're using Orthos to test for stability, are you running two instances and setting the core affinity so all four cores are loaded?
Orthos became a bit of a pain in the butt with quads so I switched to prime95 which has been optimised (version 25.5 onwards iirc) to run on quads.
EDIT: Seems I may be on the right track and while the board in review is an Asus brand the chipset is what matters. Have you checked the bios revisions release documents for your board to see if anything is mentioned about increased overclock stability with 65nm quads?
However, there is a catch. The board works remarkably well with standard dual-core Core 2 processors, but it does have some compatibility issues with quad-core CPUs. And we're talking about 65nm Kentsfields here, not even Penryns. On our test with a QX6850, we only managed to scale up to 390MHz FSB before we encountered problems with POST and stability. ASUS seems to be aware of this and we've recently seen a spate of BIOS updates addressing compatibility, so we hope that it would smoothen out in the coming months. Readers need to note that these issues only pertain to the overclockability of the board, and not general compatibility, which we've not had any issues with at all.