Some general tips about AIO's

XavierMace

Diamond Member
Apr 20, 2013
4,307
450
126
Since it seems nearly every thread on cooling these days gets jumped on by people advising AIO's are junk/worthless/etc/etc (to be fair, some are), rather than cluttering up those threads with pointless arguments, I'll share some tips/points here and be done with it. Please note this is not an attempt to say AIO's are always the best choice. There are many situations where you are better off with an air cooler.

However, in my experience, I've yet to see an air cooling setup that can match a decent AIO setup for cooling. I've been working with AIO setups pretty much since the beginning. I had both Coolermaster's and ThermalTake's early offerings. They sucked. Hard. CM's cooling performance was lackluster at best and I had multiple pump failures on the TT's. But that was a decade ago. Things have come a LONG way since then for AIO's. I've currently got 4 AIO's running at my house, two of which have been running nearly 24x7x365 since they were installed. The last pump failure I had was on the previously mentioned TT's and that was around 8 years ago. Currently my oldest AIO is a first gen Corsair H100 (non "i") and has been running since Dec 2011.

The first thing you need to consider if you are going AIO (or really liquid in general) is your case. Where can you fit a radiator? How big of a radiator can you fit? Can you fit fans on both sides of the radiator?

As far as AIO's go, there's 4 radiator sizes. 120mm, 140mm, 240mm, 280mm. The 120mm and 240mm use 120mm fans. The 140mm and 280mm use 140mm fans. I consider 240mm the minimum radiator size. Smaller than that and you really are better off with a good air cooler. At 240mm even in a sub-optimal setup will usually meet or exceed the best air cooling has to offer.

Now, where to put it. This is an often overlooked question and personally one of my issues with nearly EVERY review of an AIO I have seen. Radiators perform best as intakes. Period. Just like your cars radiator is placed at the front to take advantage of the cooler outside air, your computers radiator prefers sucking in cooler outside air. That doesn't mean you can't mount it otherwise and get satisfactory results. It just means you are either going to have to work harder on the rest of the setup or deal with lesser performance (not to be confused with insufficient performance). If at all possible, you want your radiator mounted at the front of your case as an intake. Years of experience as well as the manuals for multiple AIO kits will tell you this. Yet reviewers rarely do this which really annoy me.

There's two things that often prevent you from doing this. The design of the case or the length of the tubing on your AIO. If you're already planning on buying a new case, take this into consideration. How much does this matter? In my limited testing, I've seen temp increases of 8c-12c simply from moving the radiator from a front intake to a top exhaust. Given the whole goal here is improved cooling, that's counter productive.

That brings us to which AIO kit to get. This is actually an easier question than you might think. Many of the current generation of AIO's are re-brands of the same Asetek kit. Corsair/NZXT/etc just package them with their own fans and software. Corsair's software has a bit more features, but takes a bit more work to setup. Bonus is if you have other Corsair link components, such as a power supply, they use the same software. NZXT's is limited to just your AIO however it basically just install and forget about it. They recently overhauled it and it's a bit more resource friendly now.

You may have seen some people complain that AIO's are loud. That can be true but without more data, that's no different than saying fans are loud. There's two things generating noise on an AIO. The pump and the fans. The fans are always something you can control. However, what you may not have known is the pump is something you may have some control over as well.

If you look at the various AIO's, you'll note some specify they have a variable speed pump. You want that. For one, it's usually just a better pump in general. But, that also means you may have some control over the pump. All the AIO's software usually allows you to select a Silent/Quiet mode or a Performance mode. Across the board, this adjusts fan speeds. However, on some (I only have experience with NZXT's doing this) it also adjusts the pump speed. On performance the pump sits around 3,000rpm. That's certainly audible, although personally I'm not sure I'd call it "loud". On silent, it drops down to 1,000rpm. Personally I can't hear it at all at that speed and you're probably improving it's lifespan as well. Fan's are obviously quieter at the lower speeds as well.

However, just like most air coolers, the fans they include aren't the greatest of fans. From a performance standpoint, most of the AIO's include decent enough fans, but they are loud. Of the current AIO's I have, I found the 140mm NZXT V2's off the Kraken X61 to be the least obnoxious but that doesn't mean you can't do better. On my 120mm based radiators, I replaced the fans with Corsair's SP120's. Quieter than the stock fans (Corsair's or NZXT's) and better performing. On my 140mm based radiators, I've started running Noctua iPPC 3000rpm fans. On the AIO's silent mode they are quiet enough that I can hear them but if for some reason I need more cooling you can crank them up to silly levels of airflow.

That brings us to the fan placement. Most of the AIO's (possibly all, haven't really looked) only include enough fans for a single setup. Meaning fans on one side of the radiator. For best performance fans on both sides is preferred but single side is usually sufficient. Personally, I prefer mine setup in a "PUSH" setup, meaning the air is going through the fan, then the radiator. Especially if the fan is an intake fan as preferred.

Hopefully this sheds some light on the various opinions you see on here.
 

lehtv

Elite Member
Dec 8, 2010
11,900
74
91
Interesting write-up, thanks. Will keep this in mind in case of future overclocking builds (for myself or builds I recommend to others).

Some questions for you:

How much of a performance delta should you expect when comparing the highest end air cooling (Noctua D14, Phanteks TC14PE and such) with 240mm or 280mm front mounted AIO using those same fans at the same RPM? The thing that irks me about most AIOs is that the fans just aren't anywhere near as quiet as those on high end air coolers. With the added cost of buying an AIO in the first place, it seems ludicrous to have to replace the fans if you want it to run properly quiet... but if the performance delta is big enough, it could be worth the hassle.

Does the 8-12°C difference between intake and exhaust configurations also apply to 120-140mm AIOs? This interests me because many 120-140mm AIOs are in the same price bracket as high end air coolers. Corsair H80i and H90, for instance. I'm curious if I - or you - have been too quick to dismiss front mounted 120-140mm AIOs as valid alternatives to air cooling from a perf/dBa stand point.

What do you think about mounting AIOs as intakes on top of the case, blowing air downwards? Is this preferable to exhausting, and if so, is it a good idea to reverse the whole airflow of the case (with front fans blowing air out)?
 
T

Tim

Nice write-up!

I've only once used a radiator as intake, and I'll never do so again. The reason why? Dust. I can certainly deal with a couple of degrees higher temps easily enough. I never noticed an 8-12 degree gap, YMMV there. Taking things apart often just to clean out dust is a time consuming task that I'd rather not deal with, especially when it comes to an AIO that you'd have to take completely off the socket if you really want it cleaned hardcore after months of dusty service. "But theplaidfad," you say "just put a good filter in front of it!" Neat idea, but there goes some of your performance gain. I think the gap between performance and cleanliness is small enough to where it's not a break-breaking decision to make based on which you prefer when it comes to these modern AIO's.

A quick note about pump failures, I've never had one on the 4 different Corsair models that I've owned. I've also never had any pump failure on a pump I've owned, period. I've had an AIO in some form or another for at least the past 10 years now? From strange "Evercool" models with numerous grammar mistakes on the packaging and instructions, to Thermaltake Bigwater systems and so on. The only pump failures I've seen were on some Ibuypower systems for our workplace. Had 5 total in all, all 5 failed at some point after about year 3. There was no branding on them, and looked very cheap in build quality, so no surprise that they died.
 

BonzaiDuck

Lifer
Jun 30, 2004
15,699
1,448
126
We could address another issue here for which I've so far found insufficient explanation.

Xavier notes that there are "Silent" and "Performance" mode for pump speed -- possibly -- fan speed. Supposedly you would want to use the Corsair Link software for the corsair AiOs.

But are these discrete settings for pump speed? Why can't you control pump (and fan) speed using the motherboard thermal control feature? Does it need be a separate problem for pumps versus fans? Is it possible?

I think this issue had come up -- possibly for the H100 or H110 cooler: somebody had a pump that didn't function correctly with the motherboard connection.

Nobody would want to make manual changes between "Performance" and "Silent" if it could be done automatically as CPU temperature rises and falls.

So, given the PWM fan/(pump?) ports on the motherboard, the Corsair Link or other proprietary software, the AiO requirement for using the bundled (Asetek) pumps, and the ability to change out fans with -- for instance -- iPPC 3000's -- how would all this work in the simplest, most automatic way?
 

YBS1

Golden Member
May 14, 2000
1,945
129
106
Nobody would want to make manual changes between "Performance" and "Silent" if it could be done automatically as CPU temperature rises and falls.
I can only say this in regard to the "i" versions of the Corsair units (I have an older non "i" version as well but it's been so long since I've used it I don't recall how it works), but the fans absolutely vary automatically in relation to either cpu load or temp. I've never thought to comfirm which it's based on though. Of course you can bind them to quiet, performance, or balanced as well. I personally do not believe the pumps on any AIO I have seen are PWM though, I've never noticed pump RPM varying on the Corsair units, though I must admit I've never really watched for it either.

One area I don't think people consider though when it comes to AIO vs. traditional HSF (lets just assume both solutions are completely equal in regards to sustained load cooling ability) is water's ability to absorb brief heat spikes and remove them from the source (cpu) more quickly. Most tests you will see comparing AIOs to high end air are temps after lengthy loads are placed upon them both and they become heat soaked. In a typical usage scenario it's rare for a cpu to be loaded 100% for long periods of time, you usually have a lot of peaks and valleys.
 

BonzaiDuck

Lifer
Jun 30, 2004
15,699
1,448
126
I can only say this in regard to the "i" versions of the Corsair units (I have an older non "i" version as well but it's been so long since I've used it I don't recall how it works), but the fans absolutely vary automatically in relation to either cpu load or temp. I've never thought to comfirm which it's based on though. Of course you can bind them to quiet, performance, or balanced as well. I personally do not believe the pumps on any AIO I have seen are PWM though, I've never noticed pump RPM varying on the Corsair units, though I must admit I've never really watched for it either.

One area I don't think people consider though when it comes to AIO vs. traditional HSF (lets just assume both solutions are completely equal in regards to sustained load cooling ability) is water's ability to absorb brief heat spikes and remove them from the source (cpu) more quickly. Most tests you will see comparing AIOs to high end air are temps after lengthy loads are placed upon them both and they become heat soaked. In a typical usage scenario it's rare for a cpu to be loaded 100% for long periods of time, you usually have a lot of peaks and valleys.

I should probably open my ASUS "Sensor Recorder," play some games or use some other software or combination thereof -- and see how temperatures vary. With HWMonitor, I'd noticed that the "Maximum" column would trap an occasional temperature spike -- as much as 50C -- while the "current value" column of four cores seems to hover around an average of maybe 37. How this matters, I can't be sure. But it would be easily attenuated by simply raising the idle "floor" on my CPU and ducted-rear-exhaust fan speeds. Maybe not "eliminated," but more likely "attenuated."

Maybe some other poster has insights from monitoring pump speeds -- which I assume would be easily possible with an AiO pump.

I'm in a sort of cooling "purgatory" with my existing air solutions. With one cooler, I can actually beat an H110i by maybe 3C, but I didn't just slap it on, set it an fahget-it. With the other, it's a wash -- with the same refinements. So I don't have an incentive to shell-out and replace them with AiO's.

But I'm contemplating a build over the forthcoming span of a year, and I can see -- despite folks' assertions that something like a D15 (which I can beat) will work with an i7-5820K -- it will just have to be water in either the AiO or custom flavors.

And I've still yet to make up my mind -- entirely. I've got a CM Stacker Midtower wrapped in plastic, sitting in storage -- just de-commissioned when I replaced my Bro's system a couple months ago. I think it can be fitted with 200mm fans -- with some mods. Of course, these AiO's use at most 140's. I could move the lower drive cage a bit to the rear if there's enough room still to fit graphics cards, and I MIGHT be able to use a 200mm-square front-intake radiator with two such fans in push-pull. If not, another member recently posted his own design for a huge case of a dual 12-core Xeon rig.

But I could save myself a lot of trouble if a good AiO would work. With that, though, I'll want thermal control for . . "as much as possible." If the pump is only going to run at one speed, that, too, might suffice, but I want to know what I'm getting into before I commit.
 

XavierMace

Diamond Member
Apr 20, 2013
4,307
450
126
How much of a performance delta should you expect when comparing the highest end air cooling (Noctua D14, Phanteks TC14PE and such) with 240mm or 280mm front mounted AIO using those same fans at the same RPM? The thing that irks me about most AIOs is that the fans just aren't anywhere near as quiet as those on high end air coolers. With the added cost of buying an AIO in the first place, it seems ludicrous to have to replace the fans if you want it to run properly quiet... but if the performance delta is big enough, it could be worth the hassle.

Hard to say as I'm not aware of any major sites that have done reviews in that manner and I can't just assume my results are standard. SilentPC gears their setups to focus on silence (hence the name) so this applies to your noise setup. They are using an open bench setup which in my experience result in higher temps than a properly sealed/cooled case. But with the stock fans, they show the X61 being neck and neck with the Noctua C14 for cooling at equal noise levels and edging out the D15. Replacing the pair of stock NZXT fans with a pair of Noctua PF-15's lowered the temps by another few degrees at the same noise level. To quote them:

Compared to previous coolers we've tested with dual fans, the X61 takes the top spot, just squeaking by the Prolimatech Genesis for the title. When paired with our reference fans, it drops a few rungs but it's still within a degree or decibel of the very best. [Editor's Note: Insignificant in actual use and likely within the test margin of error.] Against other liquid coolers, well there's simply no competition. The Seidon 240M, Tundra TD03, and Liqtech 120X are't even in the same league.

However, from a pure performance standpoint they note it's only the second cooler they've tested (the first being the X41, it's 140mm little brother) that succeeded in breaking the sub-30C mark on the thermal rise testing but admittedly at a significant noise cost. I mention this because all of their testing is done with noise friendly configurations. They sacrificed 10C of cooling for a 24dBa noise reduction. Neither of those numbers are insignificant. Keep in mind, this is comparing it to the Noctua D-15 which carries a $100 price as well which means the price gap is pretty narrow at this point.

Does the 8-12°C difference between intake and exhaust configurations also apply to 120-140mm AIOs? This interests me because many 120-140mm AIOs are in the same price bracket as high end air coolers. Corsair H80i and H90, for instance. I'm curious if I - or you - have been too quick to dismiss front mounted 120-140mm AIOs as valid alternatives to air cooling from a perf/dBa stand point.

What do you think about mounting AIOs as intakes on top of the case, blowing air downwards? Is this preferable to exhausting, and if so, is it a good idea to reverse the whole airflow of the case (with front fans blowing air out)?

Again, those numbers were from my limited testing. I have no idea what others might get from doing the same thing. But, the same principal applies in regards to pulling cooler air through the radiator so you should notice some difference.

I just got home, so I decided to run some quick burn tests. This is the original stock TIM that came on the X61.

20150513085102-c5beafbe-me.png


20150513085100-d0302c4b-me.png


Only difference between the two runs was running one on silent and one on manual - 90%. On silent, the computer is nearly inaudible while sitting at my feet. At 90% the fans a quite noisy. Maybe not R9 290 reference loud, but definitely loud. On silent, temps topped out at 75c while I think is more than sufficient. Upped the voltage and multiplier a bit more.

20150513101651-7c10eb0a-me.png


Broke 80c this time which is hotter than I normally let them run. But, this was again on silent.

But are these discrete settings for pump speed? Why can't you control pump (and fan) speed using the motherboard thermal control feature? Does it need be a separate problem for pumps versus fans? Is it possible?...Nobody would want to make manual changes between "Performance" and "Silent" if it could be done automatically as CPU temperature rises and falls.

The SIMPLEST way to use to set it and forget it. The NZXT's (and the Corsair's if memory serves) default to silent mode. I never have to touch that button, it's cooling capacity is more than enough on silent mode.

What it really boils down to when you're looking at cooling is what are your priorities? If you want silent, you're generally going to want to replace your fans with most setups. If you want cool, air simply doesn't match liquids peak cooling capacity. Want both? That's where it starts to get grayer. A D15 gives a stock 240mm AIO a run for it's money if you want balance. But the AIO has the option of cooling further if you're willing to sacrifice noise at no additional cost.
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
20,839
3,174
126
Please note this is not an attempt to say AIO's are always the best choice. There are many situations where you are better off with an air cooler.

There only the best choices when u have a height clearance issue.
Because there are air heat sinks which can get close without the requirement of having 2 things which can fail... ie, the pump and fan on rad.
Also i prefer having a heat sink which has the fan blowing on the board like a noctuna C series, because it also cools the mosfets + ram, which an AIO does not.
You have very little air movement near the cpu area which isnt how intel designed these systems.
This is why a stock heat sink will blow AIR down to the board and surrounding IC's.

However, in my experience, I've yet to see an air cooling setup that can match a decent AIO setup for cooling.

i assume your talking about a 240 or 280 system?
because a 120 or 140 system can easily be done by a noctuna d15

Can you fit fans on both sides of the radiator?
you see your now in a push pull config... again... smack a push pull config on a noctuna d15 or even a prometric meglahem and u will get close.

As far as AIO's go, there's 4 radiator sizes. 120mm, 140mm, 240mm, 280mm. The 120mm and 240mm use 120mm fans. The 140mm and 280mm use 140mm fans. I consider 240mm the minimum radiator size. Smaller than that and you really are better off with a good air cooler. At 240mm even in a sub-optimal setup will usually meet or exceed the best air cooling has to offer.

there are 360mm versions as well... and not all 240mm are the same.
a thicker 120mm may be more benifical depending on heat load. Also thicker rads have wider FPI which means they are more efficient at slower speeds on the fan, hence quieter system. You cant just ball park tell people this is the best without knowing what kind of objective they are seeking.

If at all possible, you want your radiator mounted at the front of your case as an intake. Years of experience as well as the manuals for multiple AIO kits will tell you this. Yet reviewers rarely do this which really annoy me.

The main reason for this is because when you top mount a rad on an AIO it causes an air pocket to be spread out across the heat exchange area.
Air pockets are bad, and most AIO's with the exception of swiftech's where u fill it yourself, ALWAYS have an air pocket. By front mounting a rad, you move the air pocket to the top where very little heat exchange is conducting, hence you get more efficiency from the rad.

That brings us to which AIO kit to get. This is actually an easier question than you might think. Many of the current generation of AIO's are re-brands of the same Asetek kit.

Because Asekek sue'd everyone who made there own AIO's to stop making them. This is another peeve i have with AIO's.... but hey we live in a capitalistic world.

On performance the pump sits around 3,000rpm. That's certainly audible, although personally I'm not sure I'd call it "loud". On silent, it drops down to 1,000rpm. Personally I can't hear it at all at that speed and you're probably improving it's lifespan as well. Fan's are obviously quieter at the lower speeds as well.

FLOW is important in a LC system especially with a weak pump.
Leaving your pump at 1000rpm all the time, is not efficient when a system is under load. This is why MOST AIO pumps have a PWM header for 2 purpose.
1. Pump Failure detection.
2. CPU Load/Temp Sense, so the pump can variably ramp up RPM on cpu usage.
 
  • Like
Reactions: dlerious

PPB

Golden Member
Jul 5, 2013
1,118
168
106
I only see the benefit of AIOs on cramped cases where airflow is poor and max ammount of case fans is low. On those situations a front intake AIO is a pretty good solution as you remove the in-case temperature out of the equation. If you ask me, I would always go for a air cooler, safer overall and lately we have seen some REALLY good designs come to market (my beloved TR TS140P being the most interesting considering it is a single tower one, sadly I will sell it soon :( ) which have nothing to envy to the best AIOs...

Saying AIOs come with bad fans is just a lame excuse considering they are already expensive. So now for an AIO to compete in thermals/noise with an air cooler it needs aftermarket fans too? Totally defeats the purpose as the really best air coolers already ship with probably the most sickest fans out there in terms of CFM/noise. Just look at the IB-E Extreme and those two 140mm 2600RPM monsters it ships with, with the caveat that if you dont like the turbine-esque sound you can just lower the RPM and still get a good 80CFM out of them.

I find it ironic, though, that you usually see AIOs on ATX boards that dont have any expansion slot or VRM or RAM clearance issue and in full towers that have good airflow. In those cases an overpriced AIO is kind of fail IMO, unless you have some serious heat coming from your GPUs which isnt exhausted before reaching the CPU area. I usually blame the stylistic fad regarding hardware that have hit us lately for this, I can only think of nerds trying to drool over all their components behind an acrylic window, and then it probably makes sense to remove that big chunk of metal from the sight.
 

BonzaiDuck

Lifer
Jun 30, 2004
15,699
1,448
126
I only see the benefit of AIOs on cramped cases where airflow is poor and max ammount of case fans is low. On those situations a front intake AIO is a pretty good solution as you remove the in-case temperature out of the equation. If you ask me, I would always go for a air cooler, safer overall and lately we have seen some REALLY good designs come to market (my beloved TR TS140P being the most interesting considering it is a single tower one, sadly I will sell it soon :( ) which have nothing to envy to the best AIOs...

Saying AIOs come with bad fans is just a lame excuse considering they are already expensive. So now for an AIO to compete in thermals/noise with an air cooler it needs aftermarket fans too? Totally defeats the purpose as the really best air coolers already ship with probably the most sickest fans out there in terms of CFM/noise. Just look at the IB-E Extreme and those two 140mm 2600RPM monsters it ships with, with the caveat that if you dont like the turbine-esque sound you can just lower the RPM and still get a good 80CFM out of them.

I find it ironic, though, that you usually see AIOs on ATX boards that dont have any expansion slot or VRM or RAM clearance issue and in full towers that have good airflow. In those cases an overpriced AIO is kind of fail IMO, unless you have some serious heat coming from your GPUs which isnt exhausted before reaching the CPU area. I usually blame the stylistic fad regarding hardware that have hit us lately for this, I can only think of nerds trying to drool over all their components behind an acrylic window, and then it probably makes sense to remove that big chunk of metal from the sight.

I guess some of us are thinking that a D15 or ACX heatpipe is just not enough for clocks we'd like to see on an i7-5820K. Just for starters, we're talking about 140W of TDP thermal power, without adding voltage and boosting the multiplier. On the up side, the IHS is still mated to the die with indium solder.

Even the i7-4790K could offer a challenge if one really needed to overclock beyond the 4.4 Turbo speed, even with a TDP much lower. No indium solder; smaller die-size; less effective heat transfer surface overall. But we've seen an H80 cooler keep the temperatures below 70C with higher clocks, and below 60C @ 4.4 Ghz.

It's possible that the air solutions (Heatpipes) will suffice for some of this hardware. It's likely that custom-water would be more worthy for the 5820K or other "E" chip. It seems fairly certain that you could swing either way -- heatpipes or AiO -- on the Devils Canyon without missing anything.
 

PPB

Golden Member
Jul 5, 2013
1,118
168
106
I can aasure you that what an NH D15, or a IB-E Extreme cant cool, an AIO wont either, at least with bearable noise
 

XavierMace

Diamond Member
Apr 20, 2013
4,307
450
126
i assume your talking about a 240 or 280 system?
because a 120 or 140 system can easily be done by a noctuna d15

Yes, I thought I made that pretty clear. :)

you see your now in a push pull config... again... smack a push pull config on a noctuna d15 or even a prometric meglahem and u will get close.

Define close. Both of those were tested and specifically mentioned in SilentPC's and their own review showed they left 10c of cooling on the table with the X61 using the stock fans for the sake of noise. You've already spent $100 on cooling with a D15 or most of the 240mm AIO's. You could spend a fraction of that and get sufficient cooling for most scenarios. We're spending $100 because we want more than that. If another $40 on top of that gets me a few more degrees, personally that makes sense to me.

there are 360mm versions as well... and not all 240mm are the same.
a thicker 120mm may be more benifical depending on heat load. Also thicker rads have wider FPI which means they are more efficient at slower speeds on the fan, hence quieter system. You cant just ball park tell people this is the best without knowing what kind of objective they are seeking.

Yes, I should have specified generally. I have a hard time seeing a thicker 120mm keeping up with 240mm but that's somewhat moot as MOST of the AIO's have very similar radiators. That would absolutely be a valid argument if doing a custom setup where you are picking your own radiator.

FLOW is important in a LC system especially with a weak pump.
Leaving your pump at 1000rpm all the time, is not efficient when a system is under load. This is why MOST AIO pumps have a PWM header for 2 purpose.
1. Pump Failure detection.
2. CPU Load/Temp Sense, so the pump can variably ramp up RPM on cpu usage.

Flow is important to keep your liquid temperature down. That's the reason for having a pump in the first place, to keep liquid flowing through the radiator allowing it to be cooled. But if your liquid temps are fine at lower pump speeds, I don't see how increasing the pump speed is going to make any useful difference. My liquid is holding at 30c in the tests I ran above on silent. It holds at 28c on max. That 2c liquid temp is not worth maxing the pump out IMO. I ASSUME the reason Corsair recommends keeping the pump on full to to keep liquid temps down assuming most people don't monitor them (or maybe Corsair's software doesn't show it). If heat isn't a factor, I'm not aware of any benefit to running the pump at max. If there was, why bother with a variable speed pump in the first place?

In regards to the ramping, I ASSUME it would still ramp up on the silent profile if needed, but I could be entirely wrong on that.
 
T

Tim

I can aasure you that what an NH D15, or a IB-E Extreme cant cool, an AIO wont either, at least with bearable noise

...and to hit on your little pet peeve about then costing too much to be a good value, not if you replace the fans with better ones. I replaced the fans in my H100i GTX for two reasons, looks and sound.

I'm not quite sure what your huge problem with all of that is, but to each his own.
 

rchunter

Senior member
Feb 26, 2015
933
72
91
I'm in the process of replacing 2 of the fans on one of my kraken x61. They have a noisy clicking sound when turned to performance mode. Maybe I got a couple dud fans but i'm replacing them with these. Noctua NF-A14 iPPC-3000 PWM.
My other kraken x61 doesn't have the problem so i'm leaving it alone. These AIO's seem to cool fine. I pretty much like them a lot better than the thermalright ultra 120's I was using before I got them.
 

BonzaiDuck

Lifer
Jun 30, 2004
15,699
1,448
126
...and to hit on your little pet peeve about then costing too much to be a good value, not if you replace the fans with better ones. I replaced the fans in my H100i GTX for two reasons, looks and sound.

I'm not quite sure what your huge problem with all of that is, but to each his own.

The only shortcoming I see in reliable lab-test review comparisons for coolers of various kinds (heatpipe, AiO) is this.

They take the whole bundle in the shipping box as-is. Of course, this only makes sense -- in a "product" comparison: the fans are part of the "product."

It might be useful to pick a fan solution based exclusively on airflow without any regard to the noise factor, and see where these coolers top out in terms of airflow. That is, at what point in a schedule of airflows does the cooler(s) show no additional improvement in cooling?

That's not going to happen in published comparison reviews. And since enthusiasts have limited budgets, even they come away with limited knowledge based on their own custom configurations of this or that cooler.

Users can deal with noise effectively with DIY solutions and careful fan choices. Or -- users can simply accept the product-bundle they choose, as-is.

Now about the NH-D15 and the IB-E. I hadn't looked closely at the SilverArrow lately, but it's nice to see ThermalRight is "still in the running" among these high-end coolers. But there are coolers -- one that I know of -- which will best the D15 and likely the IB-E. Without special attention, they may actually match an H110i. With some special attention and no such thought to the 110i, they can actually beat the Corsair by a few degrees.

But I just don't hold out any further prospects or optimism that heatpipe coolers can be much improved next year over last year's. And you'll only get so much mileage out of AiO's when the imperative with them is to fit a radiator to more than just a handful of computer cases.

I will say this. For the price, the prospective user can avoid the tedium and "attentions" some of us put into our heatpipe installations by purchasing an AiO. He can possibly improve the AiO performance by applying the same tedium and trouble -- by how much, I can't say. And -- it's true: for the most part, you'll spend little more on an AiO than you might spend on top-end, dual-tower heatpipes.

But even in that comparison, I've found exceptions.

And just as a footnote. Each to his own! I won't replace fans because of looks and noise; I'll replace them for airflow first and noise second. And I'll deal with all the consequences at the "edge of the envelope" on my own. Because -- I can.
 
T

Tim

Not that I'd know, since I haven't used air cooling for my CPU in a looooonnng time, but doesn't having those huge hunks of metal put some wicked structural stress on these motherboards?
 

PPB

Golden Member
Jul 5, 2013
1,118
168
106
...and to hit on your little pet peeve about then costing too much to be a good value, not if you replace the fans with better ones. I replaced the fans in my H100i GTX for two reasons, looks and sound.

I'm not quite sure what your huge problem with all of that is, but to each his own.

The problem is that you, even though already paying a high price for a "high end AIO", need to shell even more cash because the stock fans arent good. This is mostly not the case with the top end air coolers.

Saying that buying new fans make them a better value (????) when you are actually PAYING even more because of that, is just insane. AIOs on stock configurations are already poor value, having their fans replaced makes them even lower value. And as I said, this is NOT the case with air coolers. The top air coolers are mostly shipped with the top end fans of their respective brands. Corsair wont even ship their AIOs with their SP fans which are actually optimized for radiators. It just shows how those AIO brands are milking their uninformed customers.

PD: Air coolers do put stress on the motherboard. Then again, backplates were invented and that problem is almost totally solved. Saying you dont use air coolers because they are heavy when most, if not all of them ship with a backplate, is just asinine. I would rather have you tell me that you just use for the looks and to have your enthusiast hardware being visible behind an acrylic window.
 
Last edited:

BonzaiDuck

Lifer
Jun 30, 2004
15,699
1,448
126
Not that I'd know, since I haven't used air cooling for my CPU in a looooonnng time, but doesn't having those huge hunks of metal put some wicked structural stress on these motherboards?

Myth. Mostly myth.

Instead, the weight of beefy fans -- plural -- contributes more to the stress.

I use the word "stress" because the issue is not simply one of "weight." Your average 120mm or 140mm fan weighs in at around 7 oz. The issue is, instead, one of "torque." A measure of torque means the weight multiplied by a distance from a pivot-point.

What's a large part of the weight of a heatpipe tower? It is the heatsink base. The base is essentially flush to the motherboard PCB -- effectively at a zero distance from its mounting point. So it literally hangs on the mounting holes, exerting minimal or no stress to the board, since it literally rests on the edge of the board at four mounting points.

The fins and heatpipe towers don't add that much to the weight, and the torque would probably be an estimate of the assembly's center-of-gravity times the weight of those components, something like half the height of the heatpipe tower times the effective weight excluding the base.

So any stress -- resulting from torque -- will arise much or mostly from adding the fan or fans. And that can be eliminated, anyway. You could have a "fanless" heatpipe tower such as mine, with the rear of the fins ducted to the case exhaust fan.

Again, to repeat my earlier statement -- I'm in sort of a "limbo" or "purgatory" of cooling choices. My existing rigs have excellent cooling -- from heatpipe towers. For my 32nm processors mated to the bottom of their IHS' with indium solder, the approximately 140W overclocked thermal power is well-matched with those coolers, and I match or exceed what an AiO would provide -- albeit with some extra "attentions" and tedious modding.

Those solutions may still work for a Devils Canyon build, and I'd also embrace the idea that I won't get much of a performance boost for OC'ing a 4790K to 4.7 as opposed to running the turbo at 4.4 "on all cores."

I can probably use an excellent heatpipe tower on something like the i7-5820K, but the overclocks will leave me envious of others who use custom-water configurations.

So I'm watching these discussions, looking at threads like "My first custom-water rig," etc. and trying -- in my mind and on paper -- to fit an available case with choice options, taking my time before I pull the trigger and actually build it.
 

XavierMace

Diamond Member
Apr 20, 2013
4,307
450
126
The problem is that you, even though already paying a high price for a "high end AIO", need to shell even more cash because the stock fans arent good. This is mostly not the case with the top end air coolers.

Saying that buying new fans make them a better value (????) when you are actually PAYING even more because of that, is just insane. AIOs on stock configurations are already poor value, having their fans replaced makes them even lower value.

First off, you don't NEED to shell out more cash. You can choose to. Like most things the further you go to an extreme, the less cost effective it is. Do I think the AIO's should include better fans? Absolutely. Of the ones I've owned I find NZXT's bunded fans to be the best all around which is one of the reasons I recommend them and SilentPC noted switching them out get a minimal improvement.

But you're being completely biased here. We are buying these for cooling. The more cooling you want, the more you pay. The AIO's are following the same basic price curve set by the air coolers. The Phanteks PH-TC12DX is $50. The Noctua D-15 is $100 and gets you about 3c at equal noise levels over the Phanteks. The Kraken X61 is $140 and gets you another 3c over the D-15. You can't say it's worth $50 to go from the Phanteks to the Noctua for 3c but it's "insane" to pay another $40 to gain another 3c over that. Plus you at least have the potential for more from the AIO. The Noctua is tapped out.

Plus, personally, I'm happy to pay a bit extra for a cooler that doesn't block plugs and slots on my motherboard like the Noctua's do.
 

rchunter

Senior member
Feb 26, 2015
933
72
91
Plus, personally, I'm happy to pay a bit extra for a cooler that doesn't block plugs and slots on my motherboard like the Noctua's do.

That's one of the main reasons I chose to go with AIO coolers in my rigs. I can promise you it wasn't so I can look at it though an acrylic window. lol.
I'm just not really sold on having a huge block of metal hanging off my MB blocking plugs.
 

BonzaiDuck

Lifer
Jun 30, 2004
15,699
1,448
126
Just for comment responding to Xavier and rchunter.

There ARE anomalies or exceptions in the general pattern of test results among a copious assortment of AiO's and heatpipes.

This one -- which I'm now using with a 2700K build @ 4.7 and 138W thermal power -- is an example:

http://www.newegg.com/Product/Produ...&cm_re=EVGA_ACX_cooler-_-35-288-004-_-Product

I'm looking at a choice of reviews for the Phanteks PH-TC12DX . Some of the comparisons -- from "LegitReviews" and "TechPowerUp" -- show maybe as many other cooler results as you can count on one hand. I'd also seen reviews a few years back from Maximum PC magazine which raise a suspicion of "Payola-in-kind:" They cherry-pick coolers for a comparison, or seem to; the winner gets a "Kickass-9"; two pages after the review, you see a two-page color ad from the winning entry. So you can wonder if the ad-purchase was a tit-for-tat in the cherry-picking.

Frosty-Tech and Hardware Secrets offer more comprehensive comparisons, and enthusiasts can apply axiomatic logic to a "merging" of several reviews: "If A > B, and B > C, then A > C" where C wasn't in the review that covered A.

I wouldn't throw the research method on the junk-heap when it comes to picking an AiO, Picking AiO over a heatpipe, or any other objective.

So . . . the PHanteks prices in at ~$50. The EVGA SuperClock -- today called the EVGA ACX -- is about $55. I suspect that pricing in that competitive market is based on manufacturing cost.

Given relative performance, the cooler-maker could actually add something, but they probably figure they'll make more money with aggressive pricing:

http://www.hardwaresecrets.com/article/Phanteks-PH-TC12DX-CPU-Cooler-Review/1727/6

But that's an 8C difference between the ACX and the Phanteks. You can look at the fans and dBA ratings (part of the review comparison), and I would say -- without even looking at a noise comparison in the review and based on immediate experience -- "throw that LED fan away!"

You then have two choices: buy another fan (which could be a Noctua iPPC 3000 for $30), or buy a ThermalRight Accordion duct for $5. But while you could say that case hardware is "separate," you'd also be inclined to replace the case exhaust fan for a similar amount -- say -- $25.

I ALWAYS chuck the bundled case exhaust fan, and I OFTEN chuck the heatpipe cooler fan. So no matter what you do, you might increase the actual cost of a $52 heatpipe cooler by another $30. You may be less inclined to do that with a $100 Noctua D15, because the fans may be quite serviceable.

And in the end, if I were to buy an AiO cooler, I'd replace those fans, too. That's why the review that will never take place -- to assess the maximum airflow at which temperature declines no further -- would otherwise be a good idea.

Otherwise, splitting hairs over a couple Andrew Jacksons doesn't give me much of a rise. If you're just slapping together a system and pinching pennies on what you'd "like" to say is a "custom-build," it makes a difference. But "Custom-build" can mean a lot of things. To get ahead of the pack in cooling effectiveness, you might spend $300 or more. To just "get an edge" on heatpipe coolers, you still might spend more than $140.

Suppose I buy the Kraken, which comes well-recommended? I'm very likely going to replace the fans anyway.

But back to the Phanteks versus ACX. Do I want cooling? Or some "acceptable-very-good" cooling that balances bundled-fan-noise and price?

I want cooling.
 
Last edited:

A.t

Member
May 11, 2015
50
0
0
Define close. Both of those were tested and specifically mentioned in SilentPC's and their own review showed they left 10c of cooling on the table with the X61 using the stock fans for the sake of noise. You've already spent $100 on cooling with a D15 or most of the 240mm AIO's. You could spend a fraction of that and get sufficient cooling for most scenarios. We're spending $100 because we want more than that. If another $40 on top of that gets me a few more degrees, personally that makes sense to me.

This is incorrect. Anybody with basic knowledge about these things would tell you that there is not a 10C difference between a top end air and AIO, most likely a few C instead, and looking at the review you mentioned...

http://www.silentpcreview.com/article1408-page6.html

You can clearly see that there is not that much of a difference.

AIO's do not outperform high end air as well as you make them out to. Like how you said "at 240mm even a sub-optimal AIO will reach or exceed the best air" is not true as I've seen cases where it was the other way around. Asetek units are rebadged under a lot of brands that some just sell them with cheap fans at high prices, under which of these cases one can do better with a good air cooler and a fan swap.
 

MrTeal

Diamond Member
Dec 7, 2003
3,554
1,658
136
Well, this escalated quickly.

I won't chime in much here, but I will say that there are some disadvantages to the front intake position. Primarily, that position usually puts the CPU block and pump above the rad/res. Especially with fillable/expandable kits like the H220 or Eisberg, any small bubbles which might be in the loop will migrate to the pump where they can give issues with pump noise.