Reasonable maximum vcore for overclocked watercooled Q6700? 1.55v? 1.60v?

[spine]

Hi all, I just wanna get a feel here for what vcore people have had their Q6600/Q6700 overclocked chips at in 24/7 setups, especially when watercooling.


Right now I've got my Q6700 at 3.8ghz and it needs about 1.56v to do it. It's a 65nm chip, I've already lapped the IHS to copper, and it's watercooled. Looking through the [H]ardforum Core2 overclocking database it seems 1.55v is the unwritten maximum reasonably safe voltage.


My gut says no worries, but I just wanna make sure. I'm more concerned for my motherboard burning out. I've added heatsinks to the mosfets and have a fan blowing on them so temperatures are good, it's just that when my test utils (prime95 mostly) fail I don't get errors, instead my PC flat resets which from experience means the cpu is starved for voltage. Sure enough, up the voltage and it's all fine. This indicates to me that the cpu has enough voltage to be stable, but there's not enough for the motherboard so you have to add a bunch more to compensate for the excessive vdroop.



Ultimately, should I be worried? What mad overclocks/voltages have people been running 24/7 in the past with Core2 Quads?

 

[sirmonkey1985]

you should be fine at 1.56v..

 

[BillParrish]

it seems 1.55v is the unwritten maximum reasonably safe voltage.

It is not unwritten, 1.55V is the Absolute Max Vcc according to the Intel Datasheet for the 6000 series. (and most 65nm chips)

A couple of things to keep in mind:

That number is not derived from physical failure analysis but instead computer modeling of the transistor gates and the physical characteristics of the materials used and the manufacturing process and a large amount of experience. It is not a "every cpu will blow up if more than 1.55V is applied" but is an engineering target for designers in that Vcc should never be allowed to exceed that value or all bets are off as to warranty, life of the product and proper operation.

Gate breakdown voltages drop with increased heat.

Intel spec for overshoot by the cpu voltage regulation circuit is .05V.

 

[spine]

It is not unwritten, 1.55V is the Absolute Max Vcc according to the Intel Datasheet for the 6000 series. (and most 65nm chips)

A couple of things to keep in mind:

That number is not derived from physical failure analysis but instead computer modeling of the transistor gates and the physical characteristics of the materials used and the manufacturing process and a large amount of experience. It is not a "every cpu will blow up if more than 1.55V is applied" but is an engineering target for designers in that Vcc should never be allowed to exceed that value or all bets are off as to warranty, life of the product and proper operation.

Gate breakdown voltages drop with increased heat.

Intel spec for overshoot by the cpu voltage regulation circuit is .05V.

Good info there, cheers!

So I guess on air it might be a concern, but on water it's probably fine then.

 

[Zero82z]

It's still a concern on water. Unless you're using sub-ambient cooling, you've still got to worry about voltages. But it's your chip, and it's up to you if you think it's worth taking the risk. Chances are it won't make enough of a difference to significantly reduce the useful lifespan of the CPU, since by the time it dies it'll most likely be obsolete already.