Oil used to cool - not cook, Intel proves the concept

[adonisslanic]

The concept of cooling with oil has been demonstrated by amateur over-clockers before with cooking oil, likewise, oil cooling is frequently used with large power transformers found in substations and underground, where materials such a fibreglass and copper, key components in motherboards, are typically present.

This time around, a year-long study has been performed by Intel with Green Revolution Cooling, revealing what enthusiasts and, with any luck, the industry, wanted to hear.

Intel discovered that with oil cooling, datacentres were able to achieve some of the best efficiency ratings witnessed, resulting in effective cooling at a fraction of the cost when compared to air-cooling. A complete failure analysis was performed on components after a year, revealing that immersion in oil had no negative impact on component lifespans.

So why is oil of such interest? Where enthusiasts are concerned, when the limits of air-cooling are reached, water-cooling is often the next step. Air is a poor conductor of heat and water offers a solution that will extract heat quickly from components and may be pumped, supplying cool fresh water.

Water-cooling can be quite costly, however, requiring special heat-exchange blocks and a network of tubes. If water leaks, it can damage components as it conducts electricity.

In an ideal world, one would simply immerse their entire computer in water, providing optimal cooling with no need for an internal tube network, with a large volume of the liquid offering superior heat dissipation and even cooling of each and every component. The conductivity of the water would, however, destroy the computer and, this is where oil steps in - offering much of the same benefits, without the nasty electrical conductivity.

Intel's study, whilst first likely to affect the HPC market, goes one step to providing hard evidence that oil cooling is both effective and safe. This could spark changes in warranty terms covering the use of oil in over-clocking and eventually lead to oil-based cooling products. We can only hope!

Sursa: Oil used to cool - not cook, Intel proves the concept - Cooling - News - HEXUS.net

Parere personala: Lol, interesanta faza, e una din chestiile alea care te face sa zici "De ce nu s-a gandit nimeni la asta pana acum?"

[phreak]

S-au gandit.

[yoyois]

Dos-ul va mirosi delicios

[ENCODED]

E veche rau metoda asta de a raci cu ulei mineral

[adonisslanic]

Da dar nu exista nimic "oficial" cu garantie cred... Prefer sa se strice ceva care e de la ei, cu garantie and stuff, decat sa fac un sistem DIY

[crs12decoder]

Ce mi se pare interesant e ca in aproape toate domeniile ingineresti, tehnologia nu poate fi aplicata fix din cauza problemelor de racire.

Inginerii nu reusesc sa gaseasca solutii de racire pentru potentialul oferit de celelalte tehnologii.

Procesoarele nu cresc in frecventa din cauza asta.

Bateriile revolutionare nu pot fi puse in aplicare din cauza asta.

Ori vor reusi sa gaseasca solutii bune de racire pentru a da frau liber evolutiei tehnologice ori vor trebui regandite actualele tehnologii chiar din fibra definitorie astfel incat sa nu aiba nevoie asa mare de racire. Ambele variante mi se par greu de atins.

[yoyois]

Am vazut cv la tv cu un inventator .. care si-a facut panouri solare si a inventat si un sistem de racire (care racea panourile de la cateva sute de grade la 50-58 C) care nu era public. oricum metode de racire exista. doar ca nu avem noi $$$ sau valoare. Dupa parerea mea nimeni nu da 2 bani pe utilizatorii de rand.