Concerns with voltage?

[the big dill]

Hey everyone,

I have been overclocking the past few cpus that i had come around and was curious as to what is limiting about increasing the voltage...?

Is it the issue with heat?

Or would it be breaking (damage) due to over-voltage? <<< this would be my educated guess, but then i found videos that people have clocked AMD CPUs a few times over their stock FSB...

Can anyone clarify this for me, as i am interested. :)

Cheers,

Phil.

P.S.

P4 650 cedar mill (stock 3.4Ghz) ----> 4.93Ghz, 1.4875V (stable, don't want to give it any more voltage) :\

Pentium D (stock 2.8Ghz) -----> 3.8Ghz, 1.4375V (and increasing) :)

All on Air cooling.

[Arceles]

There is an effect that Pentium 4 Prescot users know well, that is, "migration"; the more you apply voltage the processor electrons migrate over time, rendering the processor non functional anymore. If you don't care about hte life span of the processor then go ahead, if not, undervolt :D

[the big dill]

There is an effect that Pentium 4 Prescot users know well, that is, "migration"; the more you apply voltage the processor electrons migrate over time, rendering the processor non functional anymore. If you don't care about hte life span of the processor then go ahead, if not, undervolt :D

Yes, i read a paper that made it clear that over volting to ANY processor decreases the life span... makes sense to me. However i can get these CPUs at $5 a pop, i don't think its a major concern to me.

Which brings up another question...

If i am overvolting my CPU, can i damage the motherboard?

Cheers,

Phil.

[Arceles]

Yes, i read a paper that made it clear that over volting to ANY processor decreases the life span... makes sense to me. However i can get these CPUs at $5 a pop, i don't think its a major concern to me.

Which brings up another question...

If i am overvolting my CPU, can i damage the motherboard?

Cheers,

Phil.

That's a quite interesting question because the FETs that supply the processor current need to support all that power that the processor is demanding, but if they hasn't blown in a while I really doubt they will anyway. Those FETs unlike the processor are quite heavy duty.

[Vlad]

There is an effect that Pentium 4 Prescot users know well, that is, "migration"; the more you apply voltage the processor electrons migrate over time, rendering the processor non functional anymore. If you don't care about hte life span of the processor then go ahead, if not, undervolt :D

Electrons are always moving. The higher voltages cause higher temperatures which lead to a greater risk of a breakdown in the metal gate (although ESD can also sometime lead to the same damage).

Overvoltage can always damage components, especially those which are sensitive to heat, which is why they have data sheets describing their recommended operating state. The power MOSFETs which provide the CPU with sufficient current to operate are particularly sensitive since higher voltages will cause an almost linear increase in thermal output. At a point high voltages will cause so much current to be drawn that significant effort must be spent cooling the device. Eventually the electric fields can be so powerful that parts of the processor might be damaged (such as ripping doping ions out of the channel), among other unexpected behaviors. This depends on the manufacturing process technology, the dielectrics used, etc.

[the big dill]

Electrons are always moving. The higher voltages cause higher temperatures which lead to a greater risk of a breakdown in the metal gate (although ESD can also sometime lead to the same damage).

Overvoltage can always damage components, especially those which are sensitive to heat, which is why they have data sheets describing their recommended operating state. The power MOSFETs which provide the CPU with sufficient current to operate are particularly sensitive since higher voltages will cause an almost linear increase in thermal output. At a point high voltages will cause so much current to be drawn that significant effort must be spent cooling the device. Eventually the electric fields can be so powerful that parts of the processor might be damaged (such as ripping doping ions out of the channel), among other unexpected behaviors. This depends on the manufacturing process technology, the dielectrics used, etc.

To clarify, these processors are heavily experimental for me, as i am not afraid to lose a few dollars for the experience.

From what i understand, if i can keep temperatures below recommended average, the increased voltage should be more stable?

I am not running stock cooling, although still air, its doing a good job at it, keeping the temps below recommended at voltages approaching 1.5V.

Nobody answered my question as if there is a large risk for my motherboard to be damaged from this experiment?

[Vlad]

Eventually you wont be able to. The temperatures will be too localized because you wont be able to pull heat away fast enough.

Like I said, there is a risk to the power FETs which provide the CPU with power. If they aren't sufficiently cooled they'll overheat and need to be replaced.