Overclocking Temperature

[AdmiralFace]

I've thought about this, and the fans are gonna be removed in my project.
a simple pump system will keep the oil moving.

 

[cbn]

I've thought about this, and the fans are gonna be removed in my project.
a simple pump system will keep the oil moving.

How loud would that pump be?

 

[AdmiralFace]

almost slient? i have a noise reducing case for it. <--clever chicken

 

[Tlkki]

There are only 3 things you need to do something like this.

I'm someone with too much time and money on my hands

And enthusiasm on top of that. If you feel you have them all, then good luck and have fun!

 

[Idontcare]

If you're interested in the formulas:

static power consumption: transistors are not ideal and do not switch off completly, resulting in a small leakage current that flows even in the off state
dynamic power consumption: gates and wires have capacities that must be charged or discharged during the switching process

f.. frequency
C.. wire capacity

Pdyn: 1/2 * C * Vdd^2 * f
Psta: Idd * Vdd

P = Pdyn + Psta

Hey Voo, didn't you get the memo?

Modern CMOS processes result in dynamic power-consumption that scales to the cube of the supply voltage.

http://forums.anandtech.com/showpost.php?p=27459192&postcount=20

 

[cbn]

There are only 3 things you need to do something like this.


And enthusiasm on top of that. If you feel you have them all, then good luck and have fun!

http://www.pugetsystems.com/aquarium_computer/V2/both_module.php

$461 for the V2 kit, but that includes the case, pumps and radiator.

They are also claiming 4.6 Ghz for a QX9770 Core 2 quad with some type of passive heatsink is mounted on the CPU.

Also I think it is interesting that all the components on the mainboard receive cooling (in contrast to H20 cooling where only parts in the loop benefit from heat transfer). The downside, of course, is once the parts have been dipped in mineral oil Puget claims cleaning them off is virtually impossible.

 

[cbn]

http://www.pugetsystems.com/submerged.php

More interesting reading from Puget. The original computer they made is obviously much cheaper than the V2 or V3.

 

[muskie32]

Wow. That is quite interesting!

What are the temps like?

And can it damage the hardware at all?

 

[cbn]

Wow. That is quite interesting!

What are the temps like?

http://www.pugetsystems.com/mineral-oil-pc.php

If you scroll down to the bottom of the page it says the 4.6 Ghz Core 2 quad had temps of 50c under full load with the radiator fans on low (keep in mind they also have a 280 GTX video card in the mineral oil as well).

 

[muskie32]

Wow... I so want to do one of those now

 

[cbn]

http://en.wikipedia.org/wiki/Thermal_conductivity

If anyone is interested in thermal conductivity values here they are:

Air .025
Mineral Oil .138
Water .600

 

[cbn]

Wow... I so want to do one of those now

I was surprised to see 4.6 Ghz for a Core 2 quad. AFAIK that is better than what can be accomplished with H20.

The big question in my mind: Reliability

How does a water system compare to a mineral oil system with radiator as far as cost, leaks and failure rates go?

Pump lifespan? How does submerging the pump affect this? How about restriction differences between water and mineral oil systems?

 

[muskie32]

Go all the way to the bottom of the page:
http://www.pugetsystems.com/submerged.php

 

[cbn]

It would be interesting to hear Aigomorla's opinion on this.

According to his previous posts, he mounts a external heatsink to the bottom of his DDC.

But with a submerged pump the entire unit would be surrounded by mineral oil (as compared to a DDC dumping heat into the higher thermal conductivity of H20 and transferring the rest of the heat to ambient air).

 

[muskie32]

So, i am guessing you can do the same kind of system but with liquid nitrogen? That would be cool to see how far you can OC it ^_^

 

[cbn]

Go all the way to the bottom of the page:
http://www.pugetsystems.com/submerged.php

Thanks for the link. Apparently the 2 year update for the V2 also has it running strong without failures.

http://www.pugetsystems.com/mineral-oil-pc.php

But I just wonder did they really need such a large case and high volume of mineral oil?

Also according to this quote it sounds like they could have used a smaller cooler.

We also used a larger heatsink, though our testing still showed that the Freezer7 is an ideal pick.

 

[muskie32]

Yeah,
But 4.6 with a MAX temp of 50c sounds amazing!

 

[Voo]

Hey Voo, didn't you get the memo?

Modern CMOS processes result in dynamic power-consumption that scales to the cube of the supply voltage.

No missed that.. really interesting results - oh why can't easy physics models work at least sometimes? But no, reality always has to has the last word and complicate everything

But I'd love to see an explanation for that behavior - though I doubt my basic electronics lectures would be enough for that.

 

[AdmiralFace]

Air .025
Mineral Oil .138
Water .600

It's a shame water conducts electricity.. it would be great for this