What actually kills a CPU when OCing?

Ducky Spud

Ducky Spud

Ducky Spud

Associate
Joined
24 Feb 2009
Posts
1,758
Location
Teesside
So like the title says, what is it that kills or damages the CPU? Is it just heat or are there other things to take into account? Do increased voltages cause direct damage (not just increased heat which may later cause damage).

Just wondering as my VCore is already at 1.52V but cant get over 3.2Ghz, though I know I should be able to get to 3.5Ghz easily. My temps are around 19C - 22C idle and never get over 46C when running prime95. Now I know CPU's can be run a lot hotter than that, but Im just wondering if increasing my VCore more(already raised 0.2V) will cause other damage besides the extra heat?

Thanks
 
Ducky Spud said:
So like the title says, what is it that kills or damages the CPU? Is it just heat or are there other things to take into account? Do increased voltages cause direct damage (not just increased heat which may later cause damage).

Just wondering as my VCore is already at 1.52V but cant get over 3.2Ghz, though I know I should be able to get to 3.5Ghz easily. My temps are around 19C - 22C idle and never get over 46C when running prime95. Now I know CPU's can be run a lot hotter than that, but Im just wondering if increasing my VCore more(already raised 0.2V) will cause other damage besides the extra heat?

Thanks
Click to expand...

There are not many reports of failed or damaged CPU's around. Not considering how many there are out there!
 
Stupidly high voltage will kill your cpu eventually. But heat is the biggest killer. But even heat won't kill it quickly. I've got a Q6600, and pushed it so far that the pc throttled and shut itself off. I thought I had killed, it, but I just cleared the CMOS, and waited 2 minutes and started her up again.

The higher your voltage though, something called electro-migration can kick in.
 
Ducky Spud said:
what is it that kills or damages the CPU?
Click to expand...
Dunno? it's never happened to me or anyone I know? :confused:

Been overclocking 15 years and I never bust any hardware due to overclocking. Killed an AMD Palomino by not fitting the heatsink correctly and booting up (me silly!) :o

All the rumours are probably a conspiracy by certain top overclockers or just the results of complete and utter nOObzillas who do silly things like set uBer insane voltages or not fit heatsinks correctly lol! :p
 
* Remembers puting 1.8-2 and a bit volts through Semprons *

Stupidly high voltages will burn them out, probably instantly. Generally high voltages will reduce longevity, but then lower temps increase it. They'll almost always last longer than you want to keep them anyway.

Not many people here have had a cpu die on them and most of us overclock, so go figure.
 
Last edited:
I havent killed a cpu yet through overclocking, my previous E6600 ran most of it's life with 1.525 vcore for 3.6ghz before i sold it, my current q6600 has also been run at 3.8ghz 1.50625 vcore for quite a while and it's going fine, may shorten it's lifespan but ill hardly keep it that long.
 
Electron migration kills CPUs and other electrical components.

As more voltage is pumped through something like a CPU, it generates more heat as the electrons get ever more excited. The cooler on the CPU dissipates this energy, but eventually the voltage will become high enough to overwhelm the heatsink and the tolerances of the CPU's internal circuitry, causing the electrons to start bashing holes through all the little circuits in a processor, killing it.

It's a bad paraphrasing of what actually goes on, but hey.

Stick a CPU in an oven at around 180oC for five minutes and it will probably be ok (this is what is suggested as the way to get the IHS off Core 2 CPUs). Stick enough voltage through the CPU to send its core to that temperature and it will probably shut down or die.
 
In the short term heat is the killer, to much voltage causes some area of the chip to overheat until it literally burns out. This can happen in a matter of seconds and is why manufacturers specify a maximum voltage.

In the longer term once again heat can kill, it causes stress and weakening of joints and components until something eventually breaks. I'd peg this as happening over weeks or months.

The other long term killer is electromigration, here's a nice wiki link on the subject http://en.wikipedia.org/wiki/Electromigration

Edit: one thing worth noting here is that chips can be designed to run at different temperatures and running at or around the designed temperature should not have any ill effect. While running your cpu to the max at 70c may cause harm, your ati graphics board is, hopefully, quite happy around the 80c it was, hopefully, designed to run at.
 
Last edited:
So I should be reasonably risk free for now, if anything goes wrong it should just shut down automatically? Know its done that when voltage hasnt been high enough, but wasnt sure about voltage being "too high".

Any idea on how high the vcore can go before electromigration becomes a major issue? or should I just take it to the point my temps get to around 60 and leave it should still be fine? (which is what I was going to do)

Thanks for all your replies
 
Ducky Spud said:
Any idea on how high the vcore can go before electromigration becomes a major issue?
Click to expand...
Heh Ducky Spud you sound like you have a nervous temperament which may not put you in good stead for overclocking! ;)

The temps will guide you . . . I go by around 65°C full load although I know the chips can run hotter! :)
 
I can't remember the last time I heard of someone killing a chip. But then again I don't frequent Xtreme Systems Forums anymore, and OcUK is relatively mild in comparison.

I can 99% guarantee that heat will not kill your chip. They all have inbuilt thermal shut down nowadays anyway. You can effectively ignore temperatures until you get instability.

High voltages could kill a chip, but unless you're going completely mental you're still unlikely to fry anything. I've seen Opterons at 1.6V their entire lives that have never failed.

Don't fall into the trap of thinking better cooling will reduce the chances of killing a chip due to high voltages. As soon as you boost the voltage a significant amount you run the slight risk of damage. Good cooling will only (possibly) improve stability at very high temps.



edit: my E5200 reaches 70 degrees sometimes and I don't care. :cool:
 
Big.Wayne said:
Heh Ducky Spud you sound like you have a nervous temperament which may not put you in good stead for overclocking! ;)
Click to expand...

Ive overclocked 2 other PC's before this one, and this is already running 1.2Ghz more than stock haha. Just know its capable of 3.5Ghz easily and some lucky people with good chips have got it over 4Ghz! Anyway, this is the first time Ive properly pushed a CPU as far as I can, just want to make sure Im not going to kill it, which it sounds like Im not :)
 
I had a Q6600 it was happy at 3.6Ghz than one day booted it up and it started BSODing a few weeks after a powersupply shorted out. the voltages was not that high. reaplced ram and mobo thinking they took it, it was not it was the cpu :( still alive through but just don't like going past 333FSB
 
not sure I never seen it go over 60c unless muppet here (me) knocked the computer hard enough for the pins to bounch out, but i did that a few times and it just cut out. It seemed fine after as well
 
High VTT is the only thing I know of to kill any core 2 CPUs, despite some people saying its safe I really wouldn't wanna push it past 1.4 (and less is better) personally... older CPUs especially AMD would die from overclocking relatively quickly tho, I've seen early pentium 4s degrade in months from too much voltage... combination of voltage and heat is usually the killer, reduce one or the other and the life span is usually beyond its normal usefulness.
 
You must log in or register to reply here.
Back
Top Bottom