Water Cooling ..... All Hype??

[BigBadger]

This should start some lively debate:

Over the last few years I've used both water cooled* and air cooled systems in roughly equal measure. Ive pretty much come to the conclusion based on my experience and what I've researched that water cooling* is a waste of money. In my experience a standard re-circulating radiator type system wont give you temps significantly lower than top notch air cooling but it will be much more expensive and require more maintenance. To demonsatrate this my current rig uses an XP2500+ Mobile running at 2650MHz on 1.93V and full load (summer) temp is 44-45C on my NF7-S with D24 bios. I've yet to see any water system* that can do significantly better, certainly the ones I've tried could not.
"Ahh... but what about noise" is the usual standard response: well that really depends on what fans you use. High performance water cooled systems still need fans to cool the radiator and those fans are exactly the same as the ones you could use on your air cooled system so why should there be any difference?. Again my experience is that the 2 x 120mm fans on my water rig were just as noisey as the fans on my air cooled system. Unless you take your rad and stick it in another room your noise is not gonna be much changed. You can of course get passive cooled water systems but like passive air cooled systems their cooling performance is poor.

I firmly believe most people run water cooling* for the novelty value or they just like the look of day-glow fluid filled pipes through their case window... or more likely because they have been taken in by the hype.

I'd be interested so find out everyone elses view on the subject, especially if you have water cooling experience.

*I'd like to make it clear that this is directed at standard re-circulating radiator cooled water systems.. not chilled water or refrigerated systems which I fully except can produce extreme cooling results.

[Hawk]

I disagree, I had great results with my water setup. I'm actually thinking about using it again for the chipset and video in the machine and let the Vapo do all the CPU work.

Here is some awsome info posted by EVA2000 on the Corsair Hydrocool unit:

http://www.bleedinedge.com/forum/showthread.php?t=2292&highlight=corsair

[BigBadger]

Hawk,
Not bad results for that hydracool system.... 43C with the CPU at 1.87V. However I'd say my 44C with the CPU @ 1.93 V on air compares pretty favourably. There is no point comparing overclocks because one system is a Intel P4 and one a AMD XP, but generally the two chips put out similar heat.
I'm not saying that air cools better than water, just that water offers no significant improvement and cannot justify the extra expense and hassle, and I think the result you linked too favours my argument.
Of course your quite entitled to desagree ;)

[Hawk]

Everyone is entitled to their own opinion. ;)

And you cannot compare the P4 and AMD chip. It's not apples to apples. Compare your AMD on air and water and I gurantee that you will see different results between the two.

[BigBadger]

Everyone is entitled to their own opinion. ;)

And you cannot compare the P4 and AMD chip. It's not apples to apples. Compare your AMD on air and water and I gurantee that you will see different results between the two.
I could not agree more... it's very hard to compare one system to another and when they are very different systems comparing temps has very little value... but it was not me who posted the link for comparison in the first place ;).
The thing is, as I explained at the start; I have run my system water cooled and I did not get measurably lower temps. I'm curious to see what others have found running air and water back to back... which is why I started this thread.

[pointreyes]

P4 2.8e
air cooled with Zalman 7000Cu using AS5 cured
idle - 55c
full load - 70c

water cooled with 300w water block using AS5 cured
idle - 38c
full load - 41c

Need I say more?

[BigBadger]

P4 2.8e
air cooled with Zalman 7000Cu using AS5 cured
idle - 55c
full load - 70c

water cooled with 300w water block using AS5 cured
idle - 38c
full load - 41c

Need I say more?

Thats a pretty big temp drop... but I've got to say that if you were getting 70C load then there was something seriously wrong with your air cooling. I'm certain you could have got similar improvement simply by improving your air cooling setup. I built a P4 2.6 @ 3.2GHz for a mate at work just a few weeks back and with the stock Intel HS it never went higher than 55C. I never rated Zalman heatsinks... maybe ok for noise but not the best for cooling.

I'm sure water cooling gives far better results than poor or middle of the road air cooling... but the real test is comparing it to top draw air coolers in a well designed system.

But you dont need to take my word for it. Overclockers.com have done testing on all the top heatsinks and water coolers.
Air coolers here: http://www.overclockers.com/articles373/socketA.asp
Water coolers here: http://www.overclockers.com/articles373/waterkit.asp

Notice the top air cooler (Thermalright) and the top water kit (Corsair) share exactly the same performance C/W ratio of 0.22 . IMO thats pretty conclusive.

When you think about it there is really no reason to believe water cooling should be much better than air... Because really water cooling systems are air cooled...the radiator is after all just an air cooled heatsink.

[Snafu]

... but I've got to say that if you were getting 70C load then there was something seriously wrong with your air cooling.

The temp is 'cause it's a Prescott-cooker ( :lol: ). Sorry bud I could not resist :D .

[pointreyes]

BigBadger. Noise. I hate it, hate it, hate it. I actually had that proc running at 52c under load but the noise of all the stupid fans is nuts. And I was using 25db fans, along with a 25db PSU, and a Zalman 7000Cu. I have six computers in one corner of my house and when I was running that Prescott with all the fans it was 3x times louder than all the rest of the systems on. I treasure my haaring and prefer to keep it. The reason for the higher temps with air cooling was due to the case; however, I then tried another case with 120mm fans and once again the noise was too much for me and the blasted proc actually got hotter (that was very strange to me).

As for the Zalman with cooling. I also happen to have a 2.6c oc'ed to 3.25 with a Zalman 7000AlCu and that system is quiet. It does not reach over 50c under load with that system but it's in a CoolerMaster case with quiet fans, so your assessment of Zalmans not being that good for air cooling appears to be false with that system. Oh yeah, I originally has the retail HSF on that oc'ed 2.6 and it was louder and did indeed run hotter than with the current Zalman. It's all in the case and the case I picked for the Prescott system is simply not that great for air cooling so I worked around that issue. I don't understand why this Prescott did not work as well with air cooling unless I build a loud system. I think part of the problem is that the wattage of the Prescott releases a lot more heat than the wattage of a 2.6c.

[RotorHead]

Bigbadger, I think in general, your assessment on the water cooling is good and I do have to agree with you on some of the high end air coolers. They have come a long way on the air cooling technology over the last year. If you look at it from a purely performance point of view, most all of the well built water cooling systems will produce lower temps. Having said that, at what cost do you make the decision to cross over, price, noise, temps, coolness factor??

[BigBadger]

Bigbadger, I think in general, your assessment on the water cooling is good and I do have to agree with you on some of the high end air coolers. They have come a long way on the air cooling technology over the last year. If you look at it from a purely performance point of view, most all of the well built water cooling systems will produce lower temps. Having said that, at what cost do you make the decision to cross over, price, noise, temps, coolness factor??
That's pretty much my thinking... but I believe that the temp difference is very small indeed, probably less than 2C if anything, which for me just is not worth it. Of course some people may think that spending an extra $250 to drop 1 or 2C is money well spent... that's up to them and their bank manager :rolleyes: Personaly I'd rather spend $500 on a refrigeration system and drop 50C :thumb:

[pointreyes]

For those of you that would like to see the fun I went through with my heat generator Lian-Li case: http://www.abxzone.com/forums/showthread.php?t=72961 .

BTW: I actually thought of returning the case to the vendor I purchased it from but after discovering that another quiet case that I purchased locally was actually worst convinced that I needed to try something different. Eventually, I will get my father-in-law to dremel the front of the case to get more air flow through the system but currently it is doing good with what it's getting since I have basically moved the heat off of the proc through the radiator instead of trying to force the air from the proc through the case. I'm just happy the I'm not getting hot air coming out of the back of the case and that if I open the side panel I don't get hit with high heat either.

[pointreyes]

That's pretty much my thinking... but I believe that the temp difference is very small indeed, probably less than 2C if anything, which for me just is not worth it. Of course some people may think that spending an extra $250 to drop 1 or 2C is money well spent... that's up to them and their bank manager :rolleyes:

Now that I can agree with. It's obvious that I hava a case that was good for air flow if the proc does not release that much heat. I say that because with only liquid cooling the proc and drawing the heat out without to much release in the case was a huge improvement for the whole system. The rest of the system does not need anything more than air cooling.

[BigBadger]

Some food for thought... the temps I was quoting for my system was with the fan on quiet mode running 2500rpm. If I crank it up my full load temp is around 35C... but then it is a bit noisey... But if anyone wants to convince me that water cooling is way better than air this is what you need to beat:

[pointreyes]

Some food for thought... the temps I was quoting for my system was with the fan on quiet mode running 2500rpm. If I crank it up my full load temp is around 35C... but then it is a bit noisey... But if anyone wants to convince me that water cooling is way better than air this is what you need to beat:

You are using a Mobile XP2500-there is not need for water cooling it. I had an XP2500 unlocked running at 2.2Ghz with a Zalman 6000Cu. Only a Prescott made me consider water cooling. No one has a reason to convince you to switch to water cooling with the highly efficient very cool running mobile Athlon XP2500. Heck, I'm using the retail HSF on my current XP2500 (it's superlocked :( ) and it resides in an Antec Overture. No way in the world would I water cool it.