Welcome Guest ( Log In | Register )

Outline · [ Standard ] · Linear+

 Long term effects of overclocking., Please share your experience.

views
     
empire23
post Aug 1 2006, 12:10 AM

Team Island Hopper
Group Icon
Staff
9,417 posts

Joined: Jan 2003
From: Bladin Point, Northern Territory
Electromigration sayang is the key word in this topic for Processors.

Capacitors suffer from Dielectric degradation, slippage, contamination and cummulative ESR effects as the voltage gets bumped up.

Traces can burn out due to overcurrent (bridge wire principle), but generally heat degrades components by allowing materials to react faster with other materials, oxidization is good example and such.

But generally there's nothing to be afraid off once you stave off the short term problems like explosive loads and heat since most semi conductors are made to withstand decades of use.


empire23
post Aug 4 2006, 04:05 AM

Team Island Hopper
Group Icon
Staff
9,417 posts

Joined: Jan 2003
From: Bladin Point, Northern Territory
QUOTE(RIGmaster @ Aug 3 2006, 10:27 AM)
It is very important to never let your Processor reach 50 celcius once biggrin.gif

thats my Key biggrin.gif

Ah .... electron is such a cute fellow but also create such problem when migration and tunneling happens biggrin.gif

No way to prevent a Component for not ROSAK at all . Thats why Superconductor is so important ... The current flow without Resistance . Since no resistance , no work against the unwanted force and hence no heat being dessipated and hence no thing will be delocalised biggrin.gif

The Most stable structure is still Subatomic Particle biggrin.gif

It form from the first universe start with all the 6 Quarks , electron , positron , neutrino and anti neutrino . Things never change since then .
Haha , unless you can break quarks into even smaller particle tongue.gif
*
First of all heat isn't related, since it just damages the physical lattice and increases environment reaction rates.

Problems due to electron flow is called electromigration, as you raise the voltage, the force behind and electron continues to grow with it, the electron is already moving at C which is as fast as the laws of physics allow for, but then it doesn't stop it increasing it's kenetic energy, and that kenetic energy will eventually thin out and break those mircoscopic traces.

Resistance is another point, semiconductors require resistance to operate. It's by the laws of physics that for any action, some energy must be put it, so the switching of a transistor (which uses heat or energy to allow the source and drain to connect) must have resistance. Generally it's the use of logic, good placement, less redundancy, higher IPC and better processes that allow for less heat. Since generally all transistors are the same and switch using the same amount of energy, it's just a problem of needed switching and other factors.

Particles do not break unless you can overcome the electrostatic forces detailed in Coloumb's law and the strong force (one of the 4 fundemental forces) that binds the atoms together. There is decay, but the amount of time it takes for such particles to decay is insanely long


QUOTE(lichyetan @ Aug 4 2006, 01:53 AM)
there are some theory frm electronics which increase the voltage will make the proc die faster even with good cooling, i dunno hw to explain though but it seems related to atoms... and also all proc have a lifespan, overclock too high will greatly reduce the lifespan...
*
Generally if kept within the confines of the process limitations, procs can survive a very long time. Semiconductor based OPAMPs already have an MTBF of 10^12 hours and that's damned long lol. Same goes generally with processors, unless you take them far past their process specifications (Eg; Intel 90nm, AMD 90nm SOI, TSMC 80nm)

This post has been edited by empire23: Aug 4 2006, 04:06 AM
empire23
post Aug 4 2006, 10:04 AM

Team Island Hopper
Group Icon
Staff
9,417 posts

Joined: Jan 2003
From: Bladin Point, Northern Territory
QUOTE(ikanayam @ Aug 4 2006, 04:25 AM)
Not sure what you mean by this. AFAIK transistors vary greatly in terms of switching characteristics.
*
I meant All transistors on the same piece of silicon lah tongue.gif. I know SOI and Strained Silicon have far different switching characteristic and there are tons more processes with different variables.

 

Change to:
| Lo-Fi Version
0.0159sec    0.19    6 queries    GZIP Disabled
Time is now: 30th November 2025 - 04:46 AM