i dont think more than 1.3v vtt kills cpus, i know some people reported that, but i think it was something else that killed the cpus, or the cpu was damaged because they ran much higher vtt than that before settling for 1.3v, or vtt was unstable, or vtt was bound to vmch and that was the problem... i think on some boards you shouldnt use more vt than vmch or else it might damage the cpu or something? i think that was on gigabyte boards, but im not sure.

and dont forget your using a 45nm cpu, which means setting 1.3v in bios will results in 1.2v.
vt is crucial in reaching high fsb speeds, unless you wanna fiddle with cpu gtl, and even then you will need high vtt to reach high fsb speeds with a quad.
and fsb seems to be your problem, you cant get all cores stable at a high fsb, right?

about vcore, could be that your cpu just doesnt like more vcore?
there shouldnt be any safety features... lots of people run 2v without issues for benching on ln2.
what psu are you using?
what do you mean with relays wont latch?