Self-contained CPU water cooler?

[dauss]

When you said that you ran it at continous load, how long did you exactly run it for? Also what about people that are considering abuot buying it, but don't know about the mod. Too bad you didn't get to give us the unmodded results.

[real.genius]

In answer to your question dauss, I also have the unmodded data tests. I ran both modded and unmodded tests for about 4 hours each. The computer was warmed up throughly so my test data could be compared. This data if for a xp2100 chip on a asus a7v8x mb with ddr400 and an 8xagp albatron video card with a kt400 via chipset. I used the xp2100 since it is one of amds hottest running chips.

Water-x test data unmodded, stock fan
5% cpu loading
cpu 63c/mb 33c/ air temp 28c
100% cpu loading
cpu 67c/ mb 33c/ air temp 28c

Water-x modded with 84 cfm tornado top mt fan
5% cpu load
cpu 60c/mb 37c/ air temp 28c
100% cpu load
cpu 64c/mb 39c/air temp 28c

It ran cooler in the modded form. It is very easy to install the clips work well and easy for the amd chip. It is my first use of a water cooling system so its $60.00 price was more attractive then the $99 and up water cooler systems with many parts. Here is another test I did with water x. I mounted a tornado fan directly above the stock unit and matched the air flow direction with the stock unit. To my surprise, the cpu temp dropped but the mb temp went up.

100% load with fan mounted above stock unit
cpu 63c/mb 41c/air temp 28c

apparently heat from the onboard water-x fan and the water pump were being pushed down to elevate the mb temp while reducing the cpu temp. a little.
I also ran the water pump out of the case without a cooling fan and found that the water block surface temp went up to 33c without a cooling fan blowing on it while the air temp was at 27c. So the water pump itself also produces a little heat.
So the cooling fan has to cool the pump as well as the water in the system and the heat from the fans own motor. So there you have it, two different mods and a pump test as well as stock data.

[Bullion]

I am sorry if this appears rude BUT,
64deg is absolutly atrocious for a stock processor using an aftermarket hsf. Whilst I am happy you have posted the actual die temp and not the socket temp There is NO way you can recommend this product. My personal experience with watercooling is that a high end air hsf will preform similarly at a greatly increased noise. I have a thermaltake aquarius 2 kit on a 2100xp BUT it is overclocked with 1.85V @ 2250mhz (2800+) and it keeps the die temp below 60deg at a claimed 29db (100% load). On stock operation the load temp is 50deg. when one considers the price difference there is no way anyone can reccomend that product. which is probably why nobody has!

[dauss]

Interesting data. My Thermaltake Spark 7 running at 4200 RPM with an air temperature of 27c keep my Pentium 4 3.06gHz at 49c idle. Don't have any temps for stress load yet, but you can see where this is going.

[Phynix]

For a real world test, your numbers seem a bit odd. Currently the air about 4 feet from my computer is at 21.5C, My motherboard is showing a temp of 30C, and my processor is showing 43C. Now, I don't have an XP2100+, but I do have an XP1700+, and I'm using the stock retail AMD fan that came with it. Now, my outside air is quite a bit cooler, but not to the degree that your spiffy fan is showing.

My air is 6.5C cooler, and my processor is 20C cooler. Oh, and I'm at full load, 100% usage, and I have been for, lets see, about 36 hours since the last reboot (power outage). However, these numbers are exactly the same as 3 days ago after 2 weeks straight of 100% load.

63C is VERY hot. That's 27 degrees C away from the Athlon XP's top limit, which I believe AMD rates it for 90C.

What I'd really like to know, is what numbers you had with the stock heatsink from AMD before you replaced it with the this one, or whatever heatsink/fan combo you had before. That would show us if it's actually useful or not. A 3 degree change from stock fan to super charged fan is not uncommon on any heatsink, never the less one with a not-so-spiffy watercooling feature.

Also, keep in mind that often 'real world' tests such as yours tend to not factor in things. For example, you don't have any numbers for what your system did with a stock fan. You need a test sample to go by. Now, if your processor ran at 75C with a stock fan, and this watercooler unit brings that down to 63C, then yeah, it would be a fairly decent unit. Also, did the heatsink come with thermal compound? I'm sure you had to use something different then what was on the previous heatsink. Did you move any of the wires around when you installed the heatsink? Different airflow is a possibility as well.

This is why a system such as I posted before, with a heating element, works. The only thing that changes from review to review is the heatsink. Same thermal compound is applied, etc. It's not using a measerment system that you can easily recognize, but it is a measurement system. The results on his scale were very high compared to other heatsinks, and certainly MUCH higher then similarly priced heatsink/fan combo's. Heck, I just bought a $40 heatsink/fan combo that works 33% better then the watercool unit, and it cost me 33% less.

Watercooling is benefiecial due to the large radiators the water goes through. However, they are complex to set up. These single unit watercooler heatsinks are a gimmick, that's it. So far, no one has come out with one that performs any better then a mid range heatsink/fan combo, and they cost around 2-3 times as much. To any of you thinking about getting one of these, I suggest you skip it and find something cheaper that performs better. You'll be happier in the long run.

[dauss]

If I bought that water CPU cooler, even at 67c it is too hot. My motherboard reduces the power to the processor to keep the temperatures down when it hits 70c. 70c is the top limit for a p4 I believe.

63C is VERY hot. That's 17 degrees C away from the Athlon XP's top limit, which I believe AMD rates it for 90C.

If 90c is the top limit, 90-63=27, not 17.

[Phynix]

Doh! Edited and corrected. Nothing like a good math mistake to put you in your place, eh? Too bad I don't drink, I could have blamed it on beer or something.

Still, 63C is significantly hotter then the average AMD processor. If mine was that high, I'd be looking to change something. I believe the rule is every 10 degrees hotter halves the life of the processor.