How much Voltage is safe???

[horrain]

hello, im quite new to overclocking, and i am still learning my question is...

i have a AMD XP-mobile 2600+ running at 2600MHz (200FBS x 13) running at 1.9Volts, now is this save, when my systems idleing mbm reports that my cpu is running at 40c, when i am priming this rases to 50-52c.

is this safe or is it worth me droping my voltage down to 1.75 and have my multi at 12? or should i carry on with 1.9Volts and hope my cpu dont melt on me?

[Steve]

My 1.65V rated tbred is running at 1.89V

It runs seti@home all the time and is at about 48ºC today, it's sometimes lower, depending on ambient.

I have no info on the mobiles - what voltage is stock on it?

[blockers]

Would have thought that would be ok, i wouldnt go above it though if i were you. As that is quite a large increase for the mobile chips, and 1.9v is high for any XP's.

[Firelord]

Depends on the cooling people have given the full 2.2v treatment to these chips (a full 0.8v increase!) and they've survived under prommies and 2.1v under water

As long as its properly cooled and doesn't hit 60c and above your fine

[GreenPiggy]

Mobiles will run at the same voltage fine. They're just ubergood cos they will run stock speeds at a lower voltage, there's no other difference in what they will take.
I haven't heard of anybody having a chip die at 1.9 or 2V so you should be fine.

[Swafe]

ultra low v barton 1.35v
mobile barton 1.45v
barton 1.65v

all are esentially the same chips tho afaik, just diff voltages as lower ones are hand picked, better chips, all should be capable of 1.9ish V on air, taking it upto 2.1/2 on water/prommie

[Steve]

Athlon's can take more voltage punishment than P4's - that's for sure!

[eldren]

Ignoring heat, when you up your vcore, electron migration becomes a bigger factor. Electron migration is, as I understand it, when atoms that form the circuits of your CPU get knocked away by the relatively high-energy electrons, effectively eroding said circuits (I found a very good photo once, but Google's coming up empty ). The higher your vcore, the more force the electrons have and the lower your probable lifespan for the CPU. This isn't limited to overclocked CPUs, however; it just tends to happen more quickly on them. That said, I wouldn't go higher than 1.9v.

[gobbo]

My XP 2500+ is running at 2.45Ghz at 1.9v with watercooling. TBH i've found the ammount of voltage is really irrelevent, its the temperature you need to keep your eyes on. I've always tried to keep bellow 50C at full load.

I know some people will say that using too much voltage can "halve the life of your CPU", but what you have to remember, is that most CPUs are designed to last about 20 years.

[SilentDeath]

Power = voltage * current.
Current = voltage/resistance.
Example - a compont can be ran at 1v, or 2v. at 1v, it takes 1A of current. At 2v it takes 2A of current. At 2v it has 4x more power to disipate (as heat) than at 1v.
some 1700+ chips are rated 1.5v, and are rated to 49w. Running them at 2v will increase the power to 82w, and thats before you have increased the clockspeed.
With it overclocked aswell it will be almost doubling the heat output.

Electromigration, is what kills cpus - heat doesnt, voltage doesnt, but they are the cause of electormigration.

It happens when the conductors inside carry to much current - becuase they are too small, it heats up the conductor (because it has resistance).
It can hapen two ways, either the conductors snap, or they join together (short).

The hotter the conductors are, the more likley they are to break - which is why overclocked stuff needs better cooling.

Also, the colder it is, the less resistance it will give, meaning it will make LESS heat in a prommie than if it had an air cooler. this means its using less power. Making chips colder has similar effect as if the cpu's die was manufactuered using a smaller process.
for example my 13nm tbred will overclock a lot better than the 15nm palamino I had before it. Thats why people with phase change cooling can use much higher voltages, AND get a much higher o/c than that voltage would on air cooling.

Another thing to be carefull of is mobo overheating. The motherboards cpu voltage regulators are designed for 1.1-1.85v. Pushing them past this and they *WILL* get hot. This is eveident when going much over 1.85v, there is a lot of undervolting as soon as they are put under load. NF7-S mobos, for instance, need heatsinks, for above 1.9v with any stability. If the power the cpu is getting is unstable, then the cpu wont be happy, and will not overclock as far, and it could also shorted the life of the cpu.

I think that if you have a good heatsink, and the cpu stays below 50c, you can use about as much voltage as you want. Ive tryed 2.1v on AIR, cpu below 50c, it worked well - I managed 2.45ghz out of this chip, which only does 2.2ghz at 1.85v. Im not bothered if this chip dies however - gives me a reson to upgrade. When my cpu gets watercooled I will try upto 2.2v and probably leave it at that 24/7.