Ducis' Guide To Benchmark Cooling

ducis

Active Member
Ducis’ Guide to Benchmark Cooling

Benchmark cooling simply defined is a group of cooling methods that would be unnecessary and impractical for any other then extremely shore sessions. Usually at the time you have is to get up to your max clocks and run a few of your favorite benchmarks.

It’s already obliviously that these cooling methods are not for everyone and used by almost no one, but hey more records for you right? That said there’s a reason why most people don’t, Benchmark cooling is; expensive, dangerous at times and always time consuming. You can buy money and make time but without benchmark cooling you’ll never be able to run crysis on anything short an alien mothership :D.
____________________________________________________________________________________________________________________________________________________________________

Sub Ambience: In benchmark cooling that point at which the block makes contact with the CPU die is always below the room’s ambient temperature otherwise it your cooling solution does not count as benchmark cooling whether you break a record or not. Breaking ambience however it tough for the average person because as soon as you do you have a number of problems; condensation will form and damage your components without proper insulation (more on that later), and it requires much more creative solutions that do not rely on air (although you could argue that a condenser does).

In-depth Condensation & Insulation: Condensation occurs when ever an object is cooler then the air around it, the most common example of condensation would appear on mirror as the air around it rises in temperature (say from a shower) and it becomes cooler then the hot air around it. Unfortunately this is true with BCing (benchmark cooling) to an even greater extent because the difference in temperature between your cpu (+the area around it on the mobo), and the air temperature is higher and rises more quickly then it would if you where just having a shower. To prevent this, proper insulation is a must, and there a great guide here to show you how: http://forums.overclockers.co.nz/showthread.php?t=27239

Types of BC:
There are two main categories of benchmark cooling;
1) Refrigeration
Pros:
- Versatile
- Stable
- Reusable
- Fun to experiment with
Cons:
- (more) expensive
- One hell of a learning curve

2) Pre-cooled Compounds
Pros:
- (less) Expensive
- (less) Difficult
- As cold as it gets
Cons:
-notoriously unstable
-more of an art then a science

Refrigeration In-depth:

Refrigeration is a simple concept. Refrigeration is based on the fact that as the pressure of an element increases, the energy needed to change it into its liquid form. In its simplest form: a gas is first compressed by a compressor then the pressurised gas is cooled by a condenser (think of it as an oversized radiator) until it reaches its liquid stage (hence the name phase change), an evaporator uses the liquid refrigerant to cool the area around it until the liquid is warm enough to evaporate, and finally a dyer or filter purifies the gas.
diagram.jpg

Single Stage Phase Change:
This is the simplest form of refrigeration. These systems are next to identical with the phase change systems used in most freezers, dehumidifiers and air conditioners with two main modifications to help them more efficiently cool a CPU die. The first is shrinking the evaporator from something that would resemble a radiator (with the opposite purpose) to something closer to a water block.

This concentrates the cooling power to a smaller area greatly lowering the minimum temperature.

The second mod is to change the refrigerant to something with a lower freezing temperature then Freon (r134 the most common refrigerant), r290 and r22 (both propane with different concentrations) make great candidates because of there availability and sweet spot temps (-56C) for single stagers.
Cascade/multistage Phase Change: A cascade should only be attempted by someone how has already mastered single stagers, if you do it wrong you’ll be building a multi stage bomb that will kill you. Simply put a cascade is like a series of single stage units in a chain cooler the one next to it so the last one can cool a gas that has such a low condensing point a single stage could never come close to (such as R-784 aka krypton). Cascades in reality are much more completed then my synopsis makes them look do your research before trying.
Liquid Chillers: liquid chillers are simply cascades or single stage units with an evaporator uses to cool alcohol (or any other liquid with a low freezing point) as an\ addition tradition liquid cooling set up
Pre-Cooled Compounds In-depth:
Liquid nitrogen:
LN2O is incredibly simple, you pour it into a copper “pot” (resembles more of a tube with fins inside it at the bottom mounted over the CPU) LN2O comes standard at -192C.
Solid Carbon Dioxide:
Solid carbon dioxide or DICE (dry + ice) for short is a little tougher to work with then LN2O because it’s a solid instead of a liquid. They require specialized “pots” and are a little warmer then LN2O but are therefore cheaper to make.

Where to get started:
http://www.refrigerationbasics.com/
http://www.under-the-ice.com/ <retailer for all this sub ambient (though I recommend making/salvaging most of your parts)
www.xtremesystems.com/forum <plenty of knowledgeable people
contributed by just a noob
http://www.rparts.com/index.asp
http://www.xtremesystems.org/forums/...ad.php?t=54668
http://www.xtremesystems.org/forums/...ad.php?t=98591
**keep in mind this is a guide for the laymen and is designed to peak the interest of n00bs, if anyone decides to “go for it” then please pm me and I be happy to provide you with information or even just follow your updates**
 
Last edited:
Condensation occurs when ever an object is cooler then the air around it
A more accurate explanation would be that condensation occurs when the temperature of an object drops below the dew point. If the humidity, and thus the dew point, is low then there is less risk of condensation.

the most common example of condensation would appear on mirror as the air around it rises in temperature (say from a shower) and it becomes cooler then the hot air around it
That is because the humidity increased. The temperature, though it does relate to the relative humidity, is not really the cause.

[Referring to consumable phase change materials]
-notoriously unstable
-more of an art then a science
Actually they are the most stable and trusted method for low temperature cooling in many industries.

r290 and r22 (both propane with different concentrations) make great candidates because of there availability and sweet spot temps (-56C) for single stagers
R290 is pure propane, R22 is chlorodiflouromethane, and to get either of them to boil at -56C would take a pretty good vacuum.

A cascade should only be attempted by someone how has already mastered single stagers, if you do it wrong you’ll be building a multi stage bomb that will kill you.
A cascade system isn't really any more dangerous than a single stage unless the pressures are excessively high in one or more stages. Mixed refrigerant autocascade systems are a bit more dangerous.

LN2O is incredibly simple[...]
"LN2O" should just be "LN2"
 
Actually they are the most stable and trusted method for low temperature cooling in many industries.

with pre-cooled compounds you have little control over how quickly they boil off especially in low sessions there temps very a lot, where as in Refrigeration you know the temps your going to get and keep. pre-cooled compounds are undoubtedly colder but Its proven that there not great for long sessions

I think as soon as you drop below ambient condensation should be a concern whether it likely or not I think its a good habit to get it to

The more stages you add the more there is to go wrong hence cascades = more dangerous
 
Last edited:
ducis said:
with pre-cooled compounds you have little control over how quickly they boil off especially in low sessions there temps very a lot, where as in Refrigeration you know the temps your going to get and keep.
Actually I would contend the oposite. If I built a vapor compression refrigerator and a LN2 pot the temperature and consumption of the LN2 pot would be much closer to my predictions than the refrigerator.

ducis said:
pre-cooled compounds are undoubtedly colder but Its proven that there not great for long sessions
Of course they're not for long sessions at high power, but then we are talking about benchmark cooling here.

ducis said:
I think as soon as you drop below ambient condensation should be a concern whether it likely or not I think its a good habit to get it to
I'm not saying that you shouldn't prepare for it, but understanding the process correctly can only help.

ducis said:
The more stages you add the more there is to go wrong hence cascades = more dangerous
Though I would agree that there may be a better chance that a mistake will be made, the danger comes from higher pressure for certain refrigerants. A mistake isn't typically dangerous unless there is high pressures involved, electrical problems aside.
 
http://totl.net/Eunuch/index.html

he overclocked a 486 SX (dunno why he chose one of those, the DX has better floating point calculations) , maybe it was cheap

he ran it in his fridge and OCed it from 33mhz to 120mhz easialy, but then they decided to go all in and got it to 247mhz for two and a half minutes of halflife (lol)

thats like a 13 times overclock lol
 
sweet

im gonna look out my old pc and try and overclock its pentium1 lol
its 60mhz, but if i can get it to ~100 then it could run windows quite nicely (it has 64mb ram)
 
Back
Top