Why I don't Believe Cooling Works

1. I've never seen the percentage of heat that is reduced from the computer marked on the sales box for marketing reasons, and I would expect for the percentage of heat reduction to be listed for marketing reasons if cooling works.

2. Once the water reaches a certain level of heat the water in the tube holds heat in the pc box longer than air holds heat. So the first few hours of using the water pipes may cool the computer faster than air, but once the water in the pipes reaches the temperature that is close to as hot as the air in the computer the pipes may actually transfer hot liquid through the computer.

3. If water coolant devices really cool a computer it can be tested by running a test on a air cooled and water pc and find with temperature measurements devices the temperature of each computer at different hours of computer usage, and I've never heard of any test like this being done.

It would have made things clearer if you'd specified water cooling in the thread title, as it initially sounded as if you were criticizing cooling in general. You are certainly not the first person to bring the issue up, I might add - it's something of a perennial argument, as you can see here - though in this case you seem to have a few misunderstandings. Perhaps I'm wrong, I am not an expert in this field.

First, there is no way to show a percentage rating without knowing the heat generation rate of the system it is cooling. All of the water coolers I've see show a heat dissipation rating in either Watts (Joules per second) or BTUs per hour, reflecting how much heat it can move out of the box in a period of time, which is something that can be determined without knowing the overall configuration of the case.

Second, the whole point of water cooling is to transfer the heat to an external radiator, condenser, or Peltier junction). Otherwise, it would be as pointless as you say. The same applies to any other cooling method, however: the heat has to be dumped somewhere outside of the box. Air cooling systems do so by forcing the air out of the box and pulling cool air in, but the process is the same thermodynamically speaking.

As for the test you mention, you can find innumerable examples of such tests in the many overclocking web sites; oc'ers tend to be very empirical about this sort of thing. A casual search turns up this review fairly soon, for example; it mostly discusses the matter in terms of the CPU performance, but it does give temperature results towards the bottom. Note that even with a significantly higher overclocking rate (3.55GHz vs 3.2GHz), the water cooling did indeed maintain a lower core temperature for this configuration.

However, the effectiveness of water cooling versus air cooling will depend greatly the processor type, the case design, the layout of the drives and other components in the case, etc.

That having been said, I tend to agree that water cooling isn't very useful, not because it doesn't work but because it isn't necessary. Modern CPUs are heat-efficient enough (compared to earlier models of comparable clock speed) that in a modern case designed with air-flow in mind, the complexity of a water cooling system is rarely justified except with the most extreme overclocking. Also, water-cooling seems to be a bit passe days; it made something of a comeback a year or two ago with the introduction of several improved, easier to use kits, but AFAICT has fallen off again (I may be wrong in this, as I'm not an oc'er and don't follow the developments in the field). It seems to me that oil cooling also made a comeback in popularity around the same time, though the same arguments against it apply IMAO. This doesn't stop a lot of people from using them, admittedly. Interestingly, the relatively quiet operation seems to be as big a selling point as the higher heat dissipation rate.

Comments and corrections welcome.

References:

• Wikipedia: Computer Cooling
• Wikipedia: Water Cooling - Computer Use
• Swiftech H2O-120 Compact and Corsair Nautilus 500: Is Water Better?
• Core 2 Duo e4300 vs. e6400: performance on Air and Water Cooling
• How to Water Cool a Computer
• Hands Across Four Water Cooling Systems
• Water cooling - all hype?
• Two water coolers compared
• A Beginner's Guide for Water Cooling Your PC
• Water-cooling roundup

heatsinks don't "reduce" the heat, instead they make it easier for the heat to move from the CPU to out of the case (and into your room). The CPU is a small device that generates a lot of heat, it also happens to be sensitive to heat.

A heat sink is a device that uses something that is good at conducting heat (copper, aluminum, water) to move the heat to a large area where it can transfer to the air over a very large area, the larger area means that the CPU needs to be a lower temperature to dissipate the heat it generates. air is then pushed over the fins to keep the air that absorbs the heat cool (and thus allow a higher rate of transfer), water is good because it can be pumped over a small area and absorb all the heat very efficiently, this heats up the water which then needs to be cooled to work correctly, unlike a heatsink, water can be spread out over a very large area using pumps and pipes to make a very effective heatsink (called a radiator in this case), basically because the heat is carried by a liquid its thermal conductance is much less important as the heat can be moved using pumps to the coldest part (where it can cool quickly), and the size of the radiator determines its cooling capacity (where with an air cooled heatsink the heat is moved by relying on the thermal conductance of the material to get the heat to the fins where it can be cooled, and this somewhat limits the size of the heatsink, and thus cooling ability)

as for the ratings of the heatsinks, well if you look hard the company will often post the thermal conductance of their device, but in real life the heat your CPU sees is influenced by too many factors, air temperature, CPU, thermal compound, fan setting, dust, they all make a significant difference to the temps you will so, so they won't advertise it (because then they would have to prove it, and thats hard)

This is not like UFOs (whether you believe or not) but for simple practicalities :-

- How were the Israeli overclockers able to achieve a 4.1 Ghz aircooled Quad Core?

- How did Intel set a new record with an overclocked 5.9 GHz Quad-Core with the CPU super cooled to minus 60°C when idle?

O.K. "Water Cooling" you dont use "water" you use somthing like ethelyne glycol a non conductive or semi conductive coolant which will pull the heat off of your CPU/GPU then it goes throug a radiator kinda like a car but smaller and a fan as the air blow onto the coils in the radiator the solution is cooled then it cycles back through. I bought a BFG 8800 ULTRA at idle it was running at 69c under load my screen was tearing i checked the temp it was running a whopping 86c wow that is way to hot I RMA'd the card and got Danger den liquid cooling setup. for my GPU and CPU now both are running at 46-54c depending if I am playing vidio games or not.
so yes "Liquid cooling" does work

BTW i had some pics but i am a new user and cant post my desktop using speedfan and NVeiw monitor and a pic of rig

its been a while since I pulled out the heat transfer and thermdynamics references. Cars run glycol because its anti-freeze, and cars get exposed to freezing temperatures. Glycol hurts the specific heat of the fluid. If you want the most heat transfered, use distilled water. But you generally want some anti-corrosion stuff in the fluid to prevent rust and other bad stuff.

All heat transfer is about what engineers call "delta T" or the difference between the temperature in degrees of the hot thing and the cold thing. The more delta T, the more heat is transfered. If you have a low delta T, you need to move more cooling fluid.

This is true with air or water as the cooling fluid.

What makes "water cooling" work is that you can run cooled water from the radiator to the hot part and water has a higher specific heat than air. If your room temperature is 70 degrees F, you can try to get the return water as cold as 80F. which makes for a nice large delta T to a CPU or GPU running at 80C.

IBM mainframes used liquid cooling from the 1950s on up into the mid-1990s when I stopped paying attention. There is nothing new about using liquid instead of air.