There's been a lot of talk about flowrates....

[Graystar]

But all this talk hasn't seemed to materialize into a useful process that we can use to test and compare. So, as an admitted (or is that committed??) know-it-all, I came up with a procedure that might give us some useful information for testing. So far, all the test I've seen tests an entire system but never the individual components. (Is this kind of info somewhere I haven't looked?)

My proposal is to separately test the waterblock and radiator. The test procedure is similar for the two parts, but reversed.

To test the waterblock I suggest running 10 gallons of room-temperature water through the block at a time for various flowrates, and then measuring the temperature of the water that exits the waterblock.

To test the radiator, start by blowing through room-temperature air, then run water of different temperatures, 10 gallons at a time through the radiator and measure the temperature of the exiting water.

The first test tells us how much heat is being transferred to water and tells us how well the waterblock works. The second test tells us how much heat can be removed by the radiator.

By knowing the delta Ts for given flowrates, we can better match waterblocks, radiators, and pumps.

Opinions, critizisms, outright slander?

[Iron Hawk]

We need someone with $$$. they could buy all of the popular pumps on the market, get all of the top waterblocks, and get the most popular radiators and heatercores. then they could test all of these different things in an organized way. when done they could sell off the parts. I would love to do the testing and stuff, and it would make an awesome science fair project. unfortunatly I dont have the type of cash need to attempt this:(

[nikhsub1]

Well you're on to something but it seems like it would be difficult to test that way. I have wanted someone to do H20 reviews with different components like you mentioned. Here is an example: First, use a MOBO that reports the on die sensor in the CPU, NO EXTERNAL PROBES PLEASE:D

When testing the Maze 3 (for example) use:
RAD (A) and PUMP (A), this is test 1.
RAD (B) and PUMP (A), this is test 2.
RAD (B) and PUMP (B), this is test 3.
RAD (B) and PUMP (A), this is test 4.
RAD (A) and PUMP (B), this is test 5.

You get the picture. When reviewers test just the block in a given system, yes you can test other blocks in the same system and see which one performs better IN THAT GIVEN SYSTEM but why not go a step further and test all components to see which EXACT setup for a particular block yeilds the best results. OK, I'm done.:eek:

[Graystar]

For example, we know that PC100 memory is not going to work properly on a PC133 bus. We don't need to test it. My concept is that by rating the components we can simply pick the correct radiator to match the cooling capacity of a waterblock for a given CPU.

So it would go something like this:

From testing, it has been determined that my AMD 1.4 needs a waterblock rated at least +20 (the rating being the increased temperature of the water as it exits the block during testing.) I also have an AMD 2.4 processor. Because it runs hotter, it needs a waterblock rated at least +30 (hotter water means more heat removed.)

The first setup needs a radiator with a -20 rating (the decreased temperature drop of the water as it exits the radiator.) The second setup needs a bigger radiator with a -30 rating.

This is just off the top of my head, there probably is a better way to rate this stuff...but you get the idea.