AIO cooling round up

[Makaveli]

http://www.hitechlegion.com/reviews...tech-h240-x-open-loop-280mm-cpu-cooler-review

I've been waiting for a review like this.

Hoping to see more.

Seems the swiftech units are the best of the bunch. And while you can expand the loop on both it seems the newer 240 unit would be better for it. If you are doing cpu only cooling you can even drop down to the 220 unit.

i'm liking the low db on them aswell.

Still have my eye on one of these for my next build.

Thoughts?

 

[aigomorla]

the swiftech unit is probably the only AIO which uses a real cpu block.

The rest of the AIO use a simple copper base with fins, while the swiftech uses a real apogee XT.

The only unit which i feel would come close to it, would probably be the koolance's AIO ICM-PC40 / ICM-PC54 as it has a superior block KL-380A, however is also priced a bit more, and is bulkier.

 

[xthetenth]

I'm actually really impressed by the numbers from the Noctua as well if those graphs are right. It's holding its own quite well as an air cooler against water loops in a case which if I remember right, favored airflow to the graphics card over that to the CPU. I'm wondering what's up with the noise numbers for it because it looks like it got slowed down when the tester overclocked the system.

I definitely wasn't expecting the Noctua's performance to be a superset of the H100i's.

 

[Makaveli]

the swiftech unit is probably the only AIO which uses a real cpu block.

The rest of the AIO use a simple copper base with fins, while the swiftech uses a real apogee XT.

The only unit which i feel would come close to it, would probably be the koolance's AIO ICM-PC40 / ICM-PC54 as it has a superior block KL-380A, however is also priced a bit more, and is bulkier.

I figured it would be the quality of components with swifttech that made the difference. And I know you prefer full custom loops so for this system to get your seal of approval says alot based on how much experience you have with cooling.

I'm actually really impressed by the numbers from the Noctua as well if those graphs are right. It's holding its own quite well as an air cooler against water loops in a case which if I remember right, favored airflow to the graphics card over that to the CPU. I'm wondering what's up with the noise numbers for it because it looks like it got slowed down when the tester overclocked the system.

I definitely wasn't expecting the Noctua's performance to be a superset of the H100i's.

I was also impressed by the Noctua's performance.

Certainly held its own.

 

[AntonioHG]

And yet people say air cooling sucks compared to AIO.

 

[DrMrLordX]

This would not be the first time the nh-d15 matched or beat AiO coolers like the H100i. It has extra performance on tap for those willing to mess with more/faster fans (the same is also true of the AiO coolers featured in the above-linked review, but still).

Speaking of fan expansion: I noticed that the 140mm fans on the H240-x only push ~55 cfm. That's . . . not very powerful. Is the fin layout on the rad dense enough to benefit from higher-flow, higher-static pressure fans?

 

[xthetenth]

And yet people say air cooling sucks compared to AIO.

I've been considering AIO, but it's as much because I get a bit twitchy at the thought of hanging that much metal off my motherboard as for any cooling or anything. It looks like with the pump noise factored in, water loops have to have their fans driven stupid slowly to keep up on noise though. I can't help but think the Swiftech might benefit from some different fans. The review specifically notes that the radiator is a low FPI one, so static pressure might not be that desperately needed, though.

 

[Makaveli]

And yet people say air cooling sucks compared to AIO.

Its alittle more complicated than that.

Water is more efficient at removing the heat than air that is a fact and not debatable.

I'm not surprised this Cooler is doing well they have been in the game much longer than some of the guys making AIO's so they have many years and generations to improve on it.

Most of these AIO's are only 1st and 2nd generation products so there still lots of improvements to be had.

AIO's good ones tend to produce less noise than a standard air cooler and provide much better load temps while doing so.

 

[Cerb]

Except that they really don't produce less noise while doing so, unless you ignore coolers made for low noise, or go for a rig that simply does not have room for sufficient air cooling (I could not fit a NH-D15, or NoFan cooler, in my PC, as examples, and several air video card coolers would need custom mounting, or just not fit). The linked review is quite fair, in that regard, while many aren't, IMO. The temperature improvements are almost always at the expense of noise, not usually even from fans. But, air coolers max out much sooner, due much to limited space, and positioning options. The NH-d15 isn't truly that much better than the NH-D14, FI, but neither are usually reviewed in a way that shows their strengths (I think most cooling equipment reviewers are half deaf, already ).

Heatpipes are basically water moving heat, and the fin spacing issues work out about the same. Of course a low-FPI radiator, even small, can beat a high-FPI heatpipe cooler by a great amount, at moderate noise levels, and that's usually how the comparisons go. Likewise, with high-RPM fans, a high-FPI cooler of either kind will blow a low-FPI one away, with AIOs doing especially well, there.

On a more common midrange OC system, like say an mfenn build from any given week, or a bit better, you could get 90%+ of the performance, with less noise, and at far lower costs, from good air cooling parts. More or less what that review had and shows. The power density needs to be up to a certain minimum, varying by case, GPU, and motherboard choices, to make cooling by water truly worth it, over air (FI, you cannot very quietly cool multiple high-end GPUs, without going water--even the latest Windforce 980s will get you past 40dB@1m, assuming no hot air recycling, or pressure-related dead zones--and suitable single-GPU air coolers can take up to 5 slots!).

Those coolers tested on an OCed LGA2011-3 8C CPU, or OCed FX-8/9 series, now that would be interesting (and, I suspect that then, the Swiftech water units would have a truly commanding lead, rather than being marginally better than the NH-D15).

 

[xthetenth]

That makes sense. Unless you've got a system with enough thermal mass to absorb a whole day's heat (which I would love to see), you've got to get that heat to air eventually, but the liquid solutions give a lot more flexibility in where you get that air from and how much air you can use at once to deal with higher power densities.

And I second wanting to see them really pushed to the limit with a very hot chip. Even one of the six cores would be interesting, considering the crazy low fan RPM the swiftech was running at and still putting up great numbers (would the overclocked numbers be a decent approximation for a 5820 at stock clocks?)

 

[Makaveli]

Those coolers tested on an OCed LGA2011-3 8C CPU, or OCed FX-8/9 series, now that would be interesting (and, I suspect that then, the Swiftech water units would have a truly commanding lead, rather than being marginally better than the NH-D15).

Good post.

And I want to see this also hopefully in a follow up review or if someone else does a AIO round up.

 

[Ajay]

Those coolers tested on an OCed LGA2011-3 8C CPU, or OCed FX-8/9 series, now that would be interesting (and, I suspect that then, the Swiftech water units would have a truly commanding lead, rather than being marginally better than the NH-D15).

Definitely what needs to be tested. An AIO LC system seems pointless for an i7 4770K - just stick with air for a mainstream system. HEDT is where AIO should be entry level and custom LC for larger overclocks (especially if with LC GFX, etc.).

 

[DrMrLordX]

That makes sense. Unless you've got a system with enough thermal mass to absorb a whole day's heat (which I would love to see), you've got to get that heat to air eventually, but the liquid solutions give a lot more flexibility in where you get that air from and how much air you can use at once to deal with higher power densities.)

The way I look at it, the d15 is nothing more than a 2x140mm rad with a high fpi. There are obviously advantages and disadvantages to going straight to air when it comes to heat dissipation. If the motherboard can handle that much metal being strapped directly to the socket mounting mechanism and if the fan airflow complements your overall airflow strategy, having something like the d15 can be a big bonus. I use a d14 to blow directly out of the back of the case (I ripped out the rear fan grill), and the only "intake" fan is a desk fan blowing in through the side of the case (which is semi-permanently removed). Not sophisticated, but it works. An AiO cooler wouldn't necessarily complement the system setup as well.

Since the H240-x is low-fpi, my guess is that going with more powerful fans might not yield much in the way of superior results.

btw, I have seen at least one instance of a 9590 being cooled by a stock nh-d15, so it can handle advanced heat loads. The d15 is going to fall short in any situation where you really need something like a custom loop w/ a 3x120mm rad or better.

 

[aigomorla]

to be fair, i will tell you guys to take this review with a heavy grain of salt.

why?
He didnt list Delta of Ambient - Cores.

A shift in ambient can change the results drastically.
And ambient never stays constant after testing, unless ur doing like 1 test every several hours, or your in a temperature controlled room.


This is also why Anandtech's reviews on heat sinks have listed delta ambient / core temp.

Also on a waterblock, we usually do 3 mounts.
Because of the nature of waterblocks, we average the results of 3 mounts, and not just one mount.

So to me honestly, any review which does not show a difference between ambient with core temp just doesn't mean much anymore.