Watercooling Review @ Anandtech

qdemn7

Supreme [H]ardness
Joined
May 2, 2002
Messages
4,531
Swiftech H2O-120 Compact and Corsair Nautilus 500: Is Water Better?

This review is going to start some debate last page.

We asked when we started this review "Is water better?". The clear answer is NO, after comparing the Corsair Nautilus 500 and Swiftech H2O-120 Compact to the top air coolers. Water cooling comes close to top air performance in current designs but it did not outperform top air cooling in any test results. Upcoming improvements in block design may bring water performance closer to the best air coolers but water is not likely to provide better performance than air in kits priced at under $300. Air is also much quieter than these water kits. For most enthusiasts air cooling is the better value, better performer, and lower noise system. Unless you plan to also water cool your video card and/or chipset there is no real reason to buy water cooling instead.
 
Well it pushed my button. :eek:

Targeted at those to two entry level systems I can see them getting the results they did. I do not argue their results as it applies to those two systems ONLY.

But to say you have to spend $300+ for a quality WC system is incorrect. I submit the system in my sig for this machine. I wish it had been reviewed, For $165ish bucks (mine was $149) you get real rads (1x120 + 2x120) decent pump and a decent block, 2 x120 whisper fans and 1 x 120 adjustable speed fan. And while I did not like the retention mechanism for the block, its not bad, but I get crazy about spring mounting any heatsink/block and proper clamping force, about $15 of materials and an hour in the garage fixed that.

My point is, anyone considering water at anything above the very basic entry level should take that review with more than a grain of salt. For those two and only those two systems the conclusions can be supported. However to paint such a broad stroke on other systems (and most entry stuff is crap, but there are jewels as well) under $300 is flat wrong.

I want to see the tuniq or ultra that will maintain a 56C max core temp on my setup while running loops of 3D mark with Orthos (and thats with a GPU cooler in the loop) I do not think it can do it and remain whisper quiet. Not to mention the high OC I can get on the video card with no additional noise. Total cost $200 including the GPU block and ramsinks. Now, is it worth the price increase over a top of the line air cooler + fan + aftermarket GPU cooler, to me it is because it is just as quiet at idle as it is playing 12 hours of DAOC. However I can see going with top quality air if you are on a restricted budget, have no mechanical skills or tools. etc. It does cost more, no argument about it. Oh forgot to mention, women love the glowly stuff, nice mood lighting while you fire up some vintage Frank Sinatra mp3s and uncork the wine. Try that with your Thermalright. ;)

/start mini rant

I have been water cooling since NF7-s days, in fact it is still sitting right here as my secondary computer with "old school" water cooling. Heater core + surplus pump + DD maze 2, total cost about $100 with a highly OCed Barton mobile. (I will have to admit the array of 6 x 80mm fans on the heater core, even on medium speed, makes a nice background hum, need to 7V them one day. ) Anyone who knows which is the sharp end of a screwdriver and with a few basic tools can put together a decent WC system for $150 or go with a good kit from Petra's etc. If you don't want to mod your case then I recommend highly the kit in my sig. Best bang for the buck out there. My installation, see link, cost $4 for a 3/4 metal hole saw and $1 for some long screws. All I hear is the fan in my power supply as I have never needed to turn up the speed on the adjustable fan on low its as quiet as the low speed ones.

http://mysite.verizon.net/ressdxka/id3.html

http://www.thermaltakeusa.com/product/Liquid/DIY/cl-w0076/cl-w0076.asp

I wish to make it clear, the above is the only Tt water cooling product I would even consider using/recommending. (And the Tt case used in my build was a surplus one from a client order obtained at no cost, its OK, but if it had not been free, taking up space to store, and huge, I would have gone with something else. )
 
why debate ??? its no secret that high end air is better then low end water cooling . And if you want high end water cooling parts its gonna cost you way more than high end air . So no surprising news at all .
 
I don't think anyone will be or should be surprised by this, low-end water cooling doesn't beat high end air. So?. It just re-states the answer you see people get all the time around here, cheap systems aren't going to impress you. If you want performance you need to spend a bit more money. The only reason to get a small cheap WC loop like one of those is silence, no one buys them expecting them to out-perform their huge HSFs.

also, $300?
$60 d-tek fuzion
$50 triple rad
$90 Laing pump (ddc or d5)
$30 for addditives, distilled water, and tubing
$20 for 3 fans (could go cheaper if you get yates from jab)
$250
if you want watercooling, that would tip the scales over 300, but I think that statement you quoted is totally wrong.

Petrastech has amazing kits for well under $300, and swiftechs 220 kits are pretty nice for the price.
People who get into watercooling and spend $150 should NOT expect stellar performance, watercooling isnt for everyone the performance/price ratio compared to air is NOT the same. You pay a price premium, but you CAN get much better cooling from even a mid-range watercooling loop. Nothing will ever beat are in price/performance, air is ridiculously cheap.
 
why debate ??? its no secret that high end air is better then low end water cooling . And if you want high end water cooling parts its gonna cost you way more than high end air . So no surprising news at all .

I don't think anyone will be or should be surprised by this, low-end water cooling doesn't beat high end air. So?. It just re-states the answer you see people get all the time around here, cheap systems aren't going to impress you. If you want performance you need to spend a bit more money. The only reason to get a small cheap WC loop like one of those is silence, no one buys them expecting them to out-perform their huge HSFs.

also, $300?
$60 d-tek fuzion
$50 triple rad
$90 Laing pump (ddc or d5)
$30 for addditives, distilled water, and tubing
$20 for 3 fans (could go cheaper if you get yates from jab)
$250
if you want watercooling, that would tip the scales over 300, but I think that statement you quoted is totally wrong.

Petrastech has amazing kits for well under $300, and swiftechs 220 kits are pretty nice for the price.
People who get into watercooling and spend $150 should NOT expect stellar performance, watercooling isnt for everyone the performance/price ratio compared to air is NOT the same. You pay a price premium, but you CAN get much better cooling from even a mid-range watercooling loop. Nothing will ever beat are in price/performance, air is ridiculously cheap.



Both posts are very on point ! Although I do want to try watercooling soon,and will buy the new Swiftech 120 Compact,I came to a shocking realization recently..... You dont have to go with water cooling to overclock and still get complete or relative silence,or near to it.

You dont have to spend alot of cash to get low noise and still have mid range oc #'s.My overclocks may not be as impressive as turtletrax or tirumasu,but they do not make any noise,and for what I get out of it,I at least think they are impressive.I dont see Wesley's
article as controversial at all,merely an eye opener for the masses.I knew this ages ago.Heat pipe technology has advanced recently by massive leaps and bounds. :cool:

Still, I have installed a Swifty 120 Compact in a clients system (g0 Q6600) and know it can do better,then what was shown in that review.The pumps do make more 'noise',but that 'noise' is far from annoying.Its all somewhat subjective anyway,as some fans are fine for some,but raging screaming turbines for others.

I have turned to my new build on a few occasions and wondered if it was on.It was only in the last week after I got the Raptor X,did I notice any real noise coming out of it.Still thats
something I sheepishly admit to liking. :eek:

Since my recent system re-build (Tr U90I / Lian Li PC65B / Tr HR-03+ / IC Diamond 7) I have come to quickly see that water cooling is not the only path to maintaining a silent but healthy oc.I stupidly assumed it was.I have a x3210 B3 that I plan to water cool in the coming months,but just to say I have on a personal system,as well as I do want a new project to tackle in my spare time.
 
Before I had a Tuniq Tower and today I have a Swiftech 120 Premium on my E6600 @ 3.2 and my idle/load temps are far lower than my Tuniq ever was.

I have seen 18C temps from my kit granted it was cold as f*ck in my room but still I used to argue that air was cleaner and more effiecient, Today I can say after 6 months on the Swiftech 120 (great kit) I am happy and a fan of Water.
 
All I can say is thanks but no thanks.......

I have a Nautilus 500 coupled with a Swiftech Apogee GTX......it beat down my ArcticCooling Freezer Pro 7 which is a pretty good cooler.
Who can comment on noise, when you have extra fans in the case and two GTXs with fans running at 2300 rpm as well.
All I can say is I cant even tell my Nautilus is running.

The external system keeps warm air out of the case, and my case has absolutely no room for a WC system anyway.

For me the choice was easy.....I can cool my CPU and my NB chipset, have watercooling, run quiet, and beat air cooling temps from a good little air cooler and bonus in the NB. Plus it was fun putting it all together versus plugging in the ArcticCooler.:cool:

Cost: 145+65+50. (I liked the looks of the EK NB cooler,heh:D)
 
Wasn't the worst review ever,

hammered down a pretty simple point, mid to high end air cooling will beat low end water just about any day. Mid to higher end water on the other hand is still (usually) the more quiet and best cooling solution. Unless you get into extreme cooling.
 
I could go on further why at least the H2O-120 Compact should have beaten even the Ultra eXtreme, but most notably, they had it set up as the main exhaust for the entire system - with no sign of radiator air intake/exhaust temps. Anand hasn't done many water reviews, and it's not too surprising. For one, they aren't loading the X6800 that much, and higher heat loads are where water really shines. In addition, despite (what I believe to be) pretty faulty testing conditions, those two units don't have the size radiator most would actually recommend for any dual core OCing nowadays. They made no mention of testing multiple water block mountings - something nearly everyone else does, and something that in general takes a fair amount of practice to get an optimal mount. There's also no mention of properly bleeding either unit - it should happen with either unit's reservoir, but you never know - and we can't tell from their review. This could also explain the noisy DDC levels they complained about.

Water cooling needs a more in depth review than another HSF - including testing the actual water cooling components for their merit, not based on biased conditions. They didn't try either kit with a different fan, much like they don't with every HSF they test, which doesn't give a fair assessment of each heatsink itself, if the end user wants to change the fan. On another point, their OCing levels make absolutely no sense. Only two HSFs, the Ultra 120 eXtreme and Monsoon II Lite are able, by their standards, to make it to 3.96GHz, yet at 3.90, both the Ultra 120 and Ultima 90 are running cooler - yet they can't make OC just as high as a weaker performing HSF?

Once again, these two kits are not designed for extreme OCing either, and like most water, it gains value in cooling a CPU and a GPU (or two) and/or NB. Not to mention, Anand's temperatures in general are incredibly low compared to other reviewers. Take this with a grain of salt
 
The line of "takes 300 bucks to match air cooling" proves they know squat.
 
I dunno about water cooling but what I got from that article was that I need a Tuniq Tower 120 :D
 
Dumbest article ever. Their results contradict all the results done on the Swiftech Compact on other sites. VR-Zone used a Quad and yet still posted better temps on water than on air by a significant margine. Yet somehow, a C2D runs hotter in Anandtech's test?

Not to mention the author seems to have no idea how to correctly test. Far Cry as a test for load? WTF? How about Orthos next time.

People are mistaken. A 150$ setup consisting of DB-1 pump, 2x120mm Heatercore, and a Swiftech GT would perform significantly better than high end air and get within 80% of the performance of a high end loop. This is coming from someone who spent over $1.5k on just one computer's modding/WC'ing.
 
Back
Top