Super Cold X9100; experimental fun. Do not casually attempt!

This is my heatsink which has had copper blocks epoxied onto it:-


It's the norm to use liquid nitrogen (lNO2) in national OCing competitions to keep the CPU happily running at super high OCs. I thought i might as well try replicate the effect. I dont have convenient access to gallons of lNO2 but i do have some compressed air cans! By inverting a compressed air can and spraying the (non-flammable) ice-cold solvent directly onto the heatsink, I was able to get my temps down to a ridiculous 13C & 7C on X9100 Core0 & Core1 respectively. CPU fan stopped spinning after temps dropped below 14C and kicked in again once CPU/heatsink warmed up to 30C. Temperatures measured by RealTemp and Everest Ultimate.

I obviously did not saturate the heatsink in solvent in one huge blast otherwise I probably would have ended up with some internal liquid damage from the excess solvent dripping off. Instead, the heatsink was gradually cooled with short successive bursts of the solvent.

Even managed to get OCCT to run for approximately 18 minutes straight at 4GHz (15x multiplier @ 1.450v) before i ran out of the solvent and the stress test consequently errored out. Max. temp during the 18 minute period only reached a moderate 85C considering the Vcore setting!

It is nigh impossible to get OCCT to even complete 3 minutes on the same OC with regular air cooling (NC2000/Cryo LX) before subsequent auto-shutdown due to TJMax exceeding >102C.


This experiment was just a bit fun. Pics attached as proof:-
Pic 1: Started off on 3.6GHz on 1.3v (my usual OC) for a test run on battery. Managed to get down to 7C.


Pic 2: Repeated test but on 4GHz OC under stress on AC power. CPU cooled to same 7C temperature before OCCT stress test initiated. Due to how i was holding the notebook and air can, I wasnt able to get a screen cap during the stress test @ 100% load. The screen cap depicts the situation immediately after the test errored out (OCCT program still running in taskbar).

 

Looks like TRON to me.

 

Nice one! I wonder if freezer spray would work better? that turns any surface ice cold

 

So at what temperature when running @ 4GHZ, 1.45V did it error out? "EDIT: Looks like 85 C"
A bit over 85 Celsius since the test was going fine at 85? I'm curious how sensitive the chip is to temperature with that voltage and frequency. If the CPU doesn't need a ridiculously cool temperature to be stable it is possible to always run at those speeds if someone could devise and amazing cooling mod for your laptop!

You didn't measure power draw, did you?

What is your desktop wallpaper btw? Very nice
EDIT: It's this: http://www.hddesktopwallpaper.com/3d-wallpapers/abstract/black-cubes-sea.html

 

This was just a bit of casual fun so nothing was accurately recorded or every detail noted. And no, i didnt measure power draw either.

The main point of this was that under normal circumstances, an OCed X9100 @ 4GHz on 1.450v running OCCT or Prime 95 would hit the 105C TJMax in less than 2 minutes and therefore cause auto-shutdown of the 7805u.

With the help of the cooling solvent, I was able to keep the X9100 running at a relatively normal 85C (under load) at the same OC setting which i guess could have lasted indefinitely as long as there was a constant supply of cooling solvent applied to the heatsink.

Prime95 errored out soon after i ran out of cooling solvent, most likely due to the fact that the CPU temp skyrocketed which caused instability.