Relation between voltage and cooling

mosdapwn

Distinguished
Oct 12, 2007
70
0
18,630
I was just wondering if there is any direct relation between the increase of voltage when overclocking the processor and the amount of cooling provided. I can't get my head around any theories that better cooling could allow for stability on lower voltage, but i could notice that in being able to decrease the voltage considerably after applying better cooling to my current overclock (from 1.44 to 1.38 on a 400*9 for a Q6600 G0) and also for the 425*9 (3.8Ghz) overclock , but i can't remember the voltages.

The reason I'm saying that is that I have a chance at a refrigerator level cooling , where the temperature of the liquid in the tubes wouldn't exceed 2-4C. Temperature wouldn't be anything near a problem , so I'm going to start thinking of mad overclocks .

The one I have in mind would be a 4.8Ghz ( 533*9 ) which would also enable me to run the RAM at 1066Mhz , with 1:1 ratio and keep my 4-4-4-12 timings. While I highly doubt the motherboard would have any problem reaching a 533 FSB, If I ever could get to that overclock , I have a few questions about how my system is going to take it.

Is there a reason why the processor wouldn't respond to such frequency ?

Will the enhanced cooling allow the processor to be stable at a lower voltage ?

If the temperatures are as low as they should be, To what extent would a high voltage (probably somewhere between 1.6 and 1.7 in this case) affect the lifetime of the processor ?
 

Evilonigiri

Splendid
Jun 8, 2007
4,381
0
22,780
First of all, you'd need sub-zero cooling to get a Q6600 to 4.8GHz, not just some 2-4C cooling. Like -50C or so, I wouldn't really know.

And yes, enhanced cooling will allow you to lower the voltage a bit and still maintain stability because colder temperatures reduces resistance, and according to the formula V = IR, if the I, current, stays the same and R, resistance, goes down, V, voltage will also go down.
 

mosdapwn

Distinguished
Oct 12, 2007
70
0
18,630
Don't you think sub zero is rather required for 5Ghz + , I mean they go as far as 5.5 and 5.6 for liquid nitrogen. I used to get my processor to run at 4.1 ghz with some 1.56 voltage on some crapy cooling , it actually booted , and stayed a few mins in a stress test before I got BSOD at 85C.
I realize 4.8 is a little far fetched for liquid cooling , but I'm going to give it a try , and if I can't reach it, I'm guessing my alternative is 533*8 ( 4.26 Ghz ) or just some random FSB like 480*9 (4.32Ghz) , somewhere in that range ... I'm not sure about the temps, I could get them below 0, but somewhere between -2,-4 (its only using an extension from a refrigerator I have lying by in which case I'll use the freezer), I'm also using as much as 2 liters of antifreeze coolant.
 

Evilonigiri

Splendid
Jun 8, 2007
4,381
0
22,780
It's exponentially harder to OC higher and higher, so while you may be able to do 4.1GHz, 4.8GHz is much, much harder than you think. Think exponential. Give it a try, experience is important.

A freezer is a poor idea as a cooler because of condensation which can potentially ruin the motherboard and other components. You might also end up freezing your HD, and to top it off, a refrigerator is not meant for stressful usage such as cooling a system so it will break down within months.
 

iluvgillgill

Splendid
Jan 1, 2007
3,732
0
22,790
there is not board can run such high FSb with a high multiplier.you may reach 533 on 6 multipler.i think if you go and read any review about any top end motherboard.they only break the 500FSB barrier when they set the multiplier to 7 or 6.

plus its not only your CPU will have problem.you run the CPU in that kind of speed.you need VERY VERY good cooling to cool the NB before it overheat and freeze the system.also all the eletric component on the motherboard might mulfunction as there is an optimal operate temp range.if its too hot or too cold they will have problem.

plus run a refridguration system needs alot of preparation on the motherboard to prevent short circuit when condensation occur during operation.and if its long term use you need to mod your case for the pipes etc.

if you want a really "COOL" system use phase change cooling to get more overclock.but bare in mind the high cost involve and the amount of heat it will produce result heat up your ambient temp by more than a few degree.which will have higher graphics card,HDD,NB RAM temperature.which will have another effect on stability.

you have to know that when a thing reach a certain point it needs a big boost to achieve the next stage.take a car for example.a car need about 500bhp to get to 200mph.but to reach 250mph you need an extra 500bhp to achieve that.thats why many ferrari can get upto 200mph with power around 500-600bhp.but to get to 250+ you need 1000BHP like the Bugatti Veyron.

So getting from 4Ghz to 4.5Ghz its not as easy getting from 2Ghz to 3Ghz.
 

mosdapwn

Distinguished
Oct 12, 2007
70
0
18,630
I can deal with the issue of condensation,
and Dude , been running that Freezer 24/7 for years , as I imagine all freezers do , Nothing will break .

After all, I'm not in desperate need of that overclock, I'm just doing something odd with using a refrigerator , wanna see if it works the way I intend it to, All I said Is that I'm going to give it a try , and setting a goal as high as 4.8Ghz would eventually lead me to getting to my maximum.

As far as the FSBs go, I've seen 500*8 before, and even though not sure about 500*9 , I genuinely believe that I'm going to have encountered so many stuff before having to deal with the FSB problem , as you can imagine I'll be getting up from 440-500 gradually using the 9 multiplier.

And I really know that getting from 4Ghz to 4.5Ghz its not as easy getting from 2Ghz to 3Ghz, However, you're right , I definitely need a Veyron.
 

Evilonigiri

Splendid
Jun 8, 2007
4,381
0
22,780
My friend used a freezer for cooling for his FX-60 a while back, and his refrigerator lasted only 4months because it would be constantly running day and night. Brand new fridge btw.
 

iluvgillgill

Splendid
Jan 1, 2007
3,732
0
22,790
it would be a great achievement that i would like to see.

i have never seen a mobo which can run as high as 533x9.i wonder if your board's MOSFET can handle that kind of load?unless a very good quality board.because the NB need to handle not only high cpu speed but also high memory speed as well.the amount of data going in and out the NB is gonna create massive heat issue.

and about the fridge.as a normal fridge dont have a cool a thing that produce more than 80-100C according to your speed most of the time.
 

mosdapwn

Distinguished
Oct 12, 2007
70
0
18,630
I'm cool with my fridge dude , it's really been running 24/7 for over a year.
Neways thanks for your responses , I'll start working on it as soon as I can and post Pics and Results.
Here are pics of my current cooling


The Idea was to fill a container with such an amount of liquid so that the difference between the idle and load temperature isn't as much, since the liquid doesn't get any hotter.
The difference between the idle and the load is about 12-13 degrees as opposed to 25+ on my old gigabyte tank+pump which died on me anyways.
Idle temps at 400*9 (3.6) are 23-33C which are fairly average but the full load is 36-46C which is only 13C higher.
I realize that the whole Idea starting from a fish tank pump sunk in a 2 liters container which I intend to refrigerate is tacky and unprofessional, but I guess that's what makes it fun.