DavidNJ

I've seen several sets of analyzes where the thermal transfer rates of the different components seem to be added. However, I'm not sure I understand why that would be the case.

The heat generating components (for this discussion limited to the CPU and pump) are generating a fixed amount of heat. Irrespective of whether it is an air-cooled TRUE or a D-tek Fuzion with a GTX480, the CPU will generate its heat (say 200w) and the pump its heat (say 20w).

The radiator, depending on flows of its two fluids--air and coolant--will reach an equilibrium temperature differential between those fluids that it can dissipate that heat. It shouldn't care if the heat is created by a blow torch, a CPU, or a hamster in a cage.

The thermal resistance of the IHS, CPU water block, and thermal interface material would reflect the temperature difference between the fluid and the core. However, if a fixed amount of heat is generated, I don't see how it can affect coolant temperature.

If that is the case, then we have two separate problems. The involves reducing the coolant temperature, where more is probably always better. Factors would include the speed and quality of the two fluid flows, and the size and effectiveness of the heat exchanger.

The second involves reducing the thermal resistance between the water and the core. This involves the design of the contact between the fluid and the water block, the water block and the IHS, and within the waterblock itself.

Is my understanding correct?

DuckieHo

core -> IHS -> waterblock -> water -> radiator/cooler -> air.

Each one of those steps has a thermal resistence which will affect the core temperature. We can't do much about the core -> IHS (other than removing it). We can improve many of the other ones though.

__________________
WANTED: Socket M Core/Core 2 CPU, SUGO-02 Black, quality 802.11b/g PCI Card
To answer most of your questions: (1) a fridge cannot cool a PC (2) 64-bit OS for over 3.4GB (3) If a PCIe card fits, it will work (4) Resolution, not screen size (5) If you have a question, it is not news (6) Report, not respond to Spam (7) Single/Non-Modular Rail PSUs are NOT better than Multi-Rail/Modular

DavidNJ

yes...but not what I was saying....

This "water -> radiator/cooler -> air" determines water temperature.

This "core -> IHS -> waterblock -> water" with the addition of the TIM determines the differential between the water temperature and the cores.

Which makes them two independent subsystems, subject to different analyses. More so in that for most instances, the radiator's effectiveness is independent of flow rate.

DuckieHo

If it was just "water -> radiator/cooler -> air" that determines water temperature.... then water temperature would always be ambient. No load means no delta.

Radiator performance is dependent on flow rate: http://www.thermochill.com/PATesting/index.php

__________________
WANTED: Socket M Core/Core 2 CPU, SUGO-02 Black, quality 802.11b/g PCI Card
To answer most of your questions: (1) a fridge cannot cool a PC (2) 64-bit OS for over 3.4GB (3) If a PCIe card fits, it will work (4) Resolution, not screen size (5) If you have a question, it is not news (6) Report, not respond to Spam (7) Single/Non-Modular Rail PSUs are NOT better than Multi-Rail/Modular

DavidNJ

In that chart, differences in airflow cause a 400% change in effectiveness, while changes in coolant flow show typically less than 20%, and really less than 10% of a reasonable range of values.

And again, all at a 10° air/coolant differential. As the air/coolant differential drops, I would imagine the difference in perforamance depending on airflow grows.