Some general tips about AIO's

[BonzaiDuck]

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.

reminds me of a line for Gene Hackman as "Little Bill" in the western "Unforgiven," when he hands a loaded pistol to the paperback-writer who came to town with "English Bob" (Richard Harris) and tells him to go ahead and attempt giving the weapon to Bob, who is behind the bars:

"Hot! Ain't it?!"

Everybody has to "stand their ground" like some west Texas gun-enthusiast. I'm more or less "on the fence."

so much for "escalation."

I'd only offer that the AiO coolers are sealed units, so it shouldn't much matter if the radiator is at lower elevation than the pump. Maybe you could be right though.

 

[Tim]

Seems like more of a "to each his own" kind of thing with every post I see. Are we really talking about such significant differences in performance when it comes down to it? From the reviews I've seen, nope. You can keep your large chunk of ugly metal if it works well for you, and I'll keep my sleek looking AIO, because it's certainly working just fine for me.

 

[Ramses]

I've been extremely pleased with the superbly, nay artfully, designed and crafted Noctua NH-D14 and 15. It made short work of a 9590 and a 4790K. I'm open to an AIO but it'd have to consistently kick the D15's ass by 10-15 degrees out of the box for the same money to be interesting.

Until I have a tiny case...

 

[XavierMace]

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.

It's 100% correct. First off, read again exactly what you quoted. Then, read the article again. Look at the first table, then look at the comparison. The comparison they ran at lower fan and pump speeds for the sake of noise. In the comparison they list the X61 at 36-38c. Then look at the top chart. Those values were with the fans and pump undervolted for the sake of noise. Understandable since that's their focus, but their own chart shows they limited thermal rise to 28c running it at full speed which is a 10c improvement over the lower settings they tested with. Even with the lower settings it beat out all but 4 of the air coolers, when they replaced the fans for something quieter (again their own chart shows the stock fans actually cooled better at high speeds), the X60 topped their charts.

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.

And I've seen cases where they do, what's your point? Reviews are the only way to establish equal testing grounds. We could both go on saying "I'm right" "No I'm right" all day long. I've posted my personal experiences with some screen shots and 3rd party reviews. So far all you've said is "you're wrong". Have any documented facts to back that up? Corsair falls in that bucket of rebranding Asetek's and selling with cheap fans yet the reviews I see of the H100i show it neck and neck with the top air coolers despite it's cheap fans.

 

[BonzaiDuck]

Seems like more of a "to each his own" kind of thing with every post I see. Are we really talking about such significant differences in performance when it comes down to it? From the reviews I've seen, nope. You can keep your large chunk of ugly metal if it works well for you, and I'll keep my sleek looking AIO, because it's certainly working just fine for me.

Actually, that was the point I was making about (what seems to be) the obscure EVGA ACX cooler. It's hardly bigger than a 212 EVO, and none of those air-coolers need be deployed with a fan attached to obtain maximum CFM through the fins to a point where no further increase in cooling can be measured. That leaves the noise factor, which can also be effectively mitigated, and I had shown last year how that can be done.

The ACX will trump an H110i by about 3C degrees, and it can do so without sufficient RPM to cause any more noise than you would get with a dual-fan H110 also pushed to its maximum cooling effectiveness.

But it's also easier to achieve those results with an AiO. At that point, with those sorts of narrow margins, then the issues boil down to minor aesthetics, fan choices, and extra tedium.

Put it another way. I set up my sig-rig in 2011 with an NH-D14. At that time, it was not clear whether the AiO's were made with reliable pumps, a disaster or failure had been reported here and there, and I was able to get great results beyond what the reviews would show. I have no incentive to change it out.

For $55, the smaller ACX cooler was 6C more effective than the D14 generally, and my attentions to it made for 12C more effective than the D14 as tested in the reviews. There was a 3C margin over an H110 -- 65C at full IBT "Maximum" load stress. This second example was as much a matter of curiosity for me. So at that point, choosing air versus Aio would only be a matter of aesthetic difference.

The biggest hassle with the AiOs arises from fitting them to this or that case, choices of intake versus exhaust and so forth. All cases have a rear exhaust fan, so the right heatpipe tower makes an easy fit, even for some "aesthetic" drawback. At that point, you can weigh the shortfall of heatpipes from AiOs without special attention, and the match to AiO cooling with "special attention."

The simplest decision would be to use an AiO, but with an H80i, it becomes "more effective" with the same special attention or tedium you would apply to the air-cooler single-tower. For instance, you might still have trouble with a 240mm AiO radiator in a C70 case, which equals the trouble of fretting over installation of an H80 for stellar performance.

So with those sorts of margins, it actually is an "each to his own" fuzzy area of choices.

I'd say that AiOs -- without more than following the instructions for stock installation -- are better at the margin than heatpipes in stock installation. They're just not that much better.

They can't be, because they're marketed as an easy-installation choice, so they have to fit a lot of different cases. The radiators are only going to be so big and so thick.

If one chose the path of ducted heatpipes, top-end fans and acoustic "mitigation" with them, you're at a different 'starting point" than someone who just bought a good heatpipe cooler and stuck it to their mobo. To get out ahead of the pack using AiO's, the answer is custom-water, custom-chosen radiators to fit certain cases, maybe dual pumps and dual radiators where you can fit them.

 

[A.t]

It's 100% correct. First off, read again exactly what you quoted. Then, read the article again. Look at the first table, then look at the comparison. The comparison they ran at lower fan and pump speeds for the sake of noise. In the comparison they list the X61 at 36-38c. Then look at the top chart. Those values were with the fans and pump undervolted for the sake of noise. Understandable since that's their focus, but their own chart shows they limited thermal rise to 28c running it at full speed which is a 10c improvement over the lower settings they tested with. Even with the lower settings it beat out all but 4 of the air coolers, when they replaced the fans for something quieter (again their own chart shows the stock fans actually cooled better at high speeds), the X60 topped their charts.

Maybe because you are wrong? Posting reviews where the pump speed of the AIO slowed down, and then referring to the difference in thermal rise doesn't help your situation. I don't see a 10C difference between the aforementioned coolers under any real world conditions, that is of course, when you don't go for unfair comparisons.

And I've seen cases where they do, what's your point? Reviews are the only way to establish equal testing grounds. We could both go on saying "I'm right" "No I'm right" all day long. I've posted my personal experiences with some screen shots and 3rd party reviews. So far all you've said is "you're wrong". Have any documented facts to back that up? Corsair falls in that bucket of rebranding Asetek's and selling with cheap fans yet the reviews I see of the H100i show it neck and neck with the top air coolers despite it's cheap fans.

Let's look at the H100i you just mentioned then.

http://www.hardocp.com/article/2013...s_h80i_h100i_cpu_cooler_review/4#.VVZZsYwcHxw

With a 3770K, which is not a high TDP chip, but is a hot running CPU with the no TIM under the IHS, the H100i only manages to beat the SilverStone cooler by a few degrees. If you compare it to a better air cooler, it simply will not beat it, because it's simply not built well enough to. AIO's are cheap solutions with weak pumps and such, but that is not to say they are bad. They do outperform high end air, however if you really understand these things, you would also understand that they aren't "OVER 10C better" or "does better in even a sub-optimal condition".

 

[MrTeal]

Maybe because you are wrong? Posting reviews where the pump speed of the AIO slowed down, and then referring to the difference in thermal rise doesn't help your situation. I don't see a 10C difference between the aforementioned coolers under any real world conditions, that is of course, when you don't go for unfair comparisons.



Let's look at the H100i you just mentioned then.

http://www.hardocp.com/article/2013...s_h80i_h100i_cpu_cooler_review/4#.VVZZsYwcHxw

With a 3770K, which is not a high TDP chip, but is a hot running CPU with the no TIM under the IHS, the H100i only manages to beat the SilverStone cooler by a few degrees. If you compare it to a better air cooler, it simply will not beat it, because it's simply not built well enough to. AIO's are cheap solutions with weak pumps and such, but that is not to say they are bad. They do outperform high end air, however if you really understand these things, you would also understand that they aren't "OVER 10C better" or "does better in even a sub-optimal condition".

Really for showing the difference between coolers only the power dissipation matters. The hot running part doesn't really matter, as that it an effect of the CPU itself. The poor TIM will essentially give you a larger offset at a given power level vs a soldered IHS, but you'd have the same difference between two coolers. IE, using some made up numbers if you're dissipating 100W and cooler A gets ΔT of 40C while cooler B gets 44C on the soldered CPU, you might see cooler A get 45C on the paste TIM CPU while gets 49C for the same 100W. The relative difference between them shouldn't change unless conditionals are wildly different, like if the TIM is really poor and the die starts conducting a larger portion of that 100W through the socket into the MB.


Really for a typical CPU AOI coolers don't offer enough performance benefits over a big air cooler to justify saying they're a must have. CPUs are relatively low power devices, and they have a massive amount of space around them to put a big heatpiped cooler. Dissipating 100-150W there's not going to be a really big difference between a decent CLC and a big air cooler. Where the CLCs really shine is in space constrained applications where a massive air cooler isn't practical, or where the power that needs to be dissipated is much larger. GPUs are a pretty good example. An Asetek CLC might not perform much better than a NH-D15 on a 150W CPU, but then you can't mount an NH-D15 to a GPU. Toss an off the shelf CLC in a G10 bracket on your GPU and there's a night and day difference in thermals and acoustics.

 

[aigomorla]

You can not compare a CLC vs a AIO...

CLC - Custom Liquid Cooling
AIO - All in One..

The ONLY AIO hybrid u can put on near floor plan would be the koolance AIO, or the swiftech one which allows you to fill / customize as required.

But 100% Asetek branded are all AIO's.


CLC's do EXTREMELY well in GPU's... reduction of thermals by 1/2 is the typical result from liquid cooling your GPU.
Full cover blocks, offer cooling on mosfets / vregs and memory as well, which extends the life of your GPU even if you hammering it at extreme conditions.

My main system will always be a CLC.... i seriously cant go back after sitting on this pedistool for so long.
My brother's system has an AIO, because i could not fit any type of massive heat pipe sink in it. I did however swap out all his fans to Scythe GT's because i had a cache of them not in use, and it really did give him a net benifit.
However the fans itself cost about 2/3rds the price of his AIO unit, so a net cost basis for a new person makes no sense.

All my servers, are on AIR.... NO EXCEPTIONS to this, as they are mission critical. Having a AIO in my eyes on a server is a big nono... as u introduce more things which can fail, and cause a downtime on the server.
Also i needed a Air sink which would blow air on the board and keep the IC's on the board relatively cool, which again AIO's can not, so my servers all have Noctuna C14's on them with a GT fan.

 

[MrTeal]

CLC - Closed loop cooler. Just another term for AOI, though I probably should have used the same acronym throughout for consistency. Sorry.

I ended up using a couple Swiftech H220s on my GPUs because I had a dozen of them laying around, but it was a little hairy for the VRMs without a full block. The easy expandability of it ended up being a big plus because I could just crack it open and add in one of those $5 40x40mm Chinese waterblocks for the VRMs, which got me most of the way to full block effectiveness without having to buy a $120 block I can't reuse.

For servers, I'd probably agree. I'm not sure if the CoolIt DCLC parts have extra validation vs their rebranded commercial stuff. I think that's pretty niche in the server space though. Generally if you're doing a really serious server setup you would have a separate water system with a chiller on the roof.
Depending on how you lay out the parts airflow doesn't have to be an issue even with a really powerful system. Two AIO systems with some high spec 200CFM fans will leave no part of a 4U server chassis without some serious airflow.

 

[XavierMace]

Maybe because you are wrong? Posting reviews where the pump speed of the AIO slowed down, and then referring to the difference in thermal rise doesn't help your situation. I don't see a 10C difference between the aforementioned coolers under any real world conditions, that is of course, when you don't go for unfair comparisons.

Are you just ignoring that first chart on the SilentPC deliberately or are you just not seeing it? That chart IS the real world numbers for the majority of people. How many people do you think are running AIO's and then undervolting the pump like SilentPC did for the review?

 

[A.t]

Are you just ignoring that first chart on the SilentPC deliberately or are you just not seeing it? That chart IS the real world numbers for the majority of people. How many people do you think are running AIO's and then undervolting the pump like SilentPC did for the review?

SilentPC, as their name suggests as you also mentioned, goes for absolute silence in their reviews. First of all pick a normal review if you're going to. Let's look at one.

http://www.hardocp.com/article/2014...1_aio_liquid_cpu_cooler_review/3#.VVdSt4wcHxw

It got owned by the Cryorig cooler in non-load temps, which is what the CPU most the time does for most people.

So where is that 10C of difference again? Surely it's obvious here that you're nitpicking reviews and you're providing false information to people.

 

[BonzaiDuck]

SilentPC, as their name suggests as you also mentioned, goes for absolute silence in their reviews. First of all pick a normal review if you're going to. Let's look at one.

http://www.hardocp.com/article/2014...1_aio_liquid_cpu_cooler_review/3#.VVdSt4wcHxw

It got owned by the Cryorig cooler in non-load temps, which is what the CPU most the time does for most people.

So where is that 10C of difference again? Surely it's obvious here that you're nitpicking reviews and you're providing false information to people.

I've stated my own position, and I think it's fairly objective.

I think a reasonable concern about noise can give relative satisfaction compared to an absolutely obsessive concern about total noise-less-ness.

That's why I'd like to see reviews that throw the noise factor to the wind for part of the tests, to assess the actual airflow necessary to make ANY cooler (AiO or heatpipe) reach a point with 0 additional C degrees temperature drop.

Sure -- the noise factor can be part of those reviews, and understandable that the reviews will ONLY use the fans bundled with any of the coolers tested. But if the enthusiast is SERIOUSLY interested in cooling first and foremost, there are ways to eliminate a good part of the noise, and it would be more helpful to see just exactly WHICH of these coolers will prove to have the LOWEST load temperature at the same room-ambient regardless of "noise" and for whatever CFM the beefiest fan will push through the fins. THEN we can worry about your silly noise obsessions.

I actually think that the noise worries have kept Noctua from perfecting their coolers for maximum cooling. They've taken designs that are massive, but don't perform any better than some other single tower heatpipes.

If it were possible, I'd say that somebody should find a noise-meter and set some parameters for a test between my own cooling solutions (which deploy at least one GT AP-30 PWM per system) and any other stock solution -- matched according to processor, thermal wattage, room-ambient, load-TEST, and load temperature. And I'll bet that my systems are just as quiet at -- say -- 60 or 65C -- as any-of-ya's. So there.

[Some of you folks are p***ing so hard in the wind about this "heatpipe-vs-AiO" thing, the noise thing, and all of it -- it's time to change clothes and do some laundry. Geesh!]

 

[MoInSTL]

I retired an i7 2600k and went with an HR-02 Macho in a Fractal R4. It ran cool and quiet. But to clean the fins on both sides meant removing the motherboard. Something I never did. I just hit it with compressed air whenever I had a reason to open the case.

I now have an Asus X99-A with an i7 5820K. I still have the same R4. This time I decided to go with a CLC and ended up going with a small but chubby X41. I mounted it in the bottom front with two Noctua Redux 140 fans in a push/pull configuration. I chose it for convenience. The CAM software is a little bloated but isn't required and reports higher CPU temps than RealTemp does.

I don't fall into either cooling camp. This time around I wanted to try something different and didn't want the hassle of a big heatsink. I read many, many AIO reviews and was using a cheap Hyper 212 EVO (not the Pro version) in the interim as the Haswell-E does not have the Intel heatsink and fan.

I didn't expect it to be any better than air. Just as good as air. I just installed it last weekend. I also picked up a new PSU and ran out of time to do any decent cable management. So right now one case panel is off. So I can't report any temps until I close it back up. I also am running stock speeds since the Hyper 212 couldn't keep up even at stock without running in the low to mid 30s even at idle. In all fairness it only had one 120 fan. So basically I am starting from scratch with the board and the X41.

I like that I can run it in Performance, Silent and Manual mode. I like that I can run 3rd party fans and monitor them in the CAM software. I removed both of my drive cages a few days after I got my case as I just use SSD drives as I also have a Synology NAS, drive dock and a WD My Book. So for me, that open space was a perfect spot to install the radiator. Haven't noticed any ill effects of having it below the block. I can say that if there was a case as sleek and quiet as my R4 I would buy it if it had room for a top mounted X61. I have physical room for it but the hole spacing is wider and does not line up.

 

[BonzaiDuck]

I retired an i7 2600k and went with an HR-02 Macho in a Fractal R4. It ran cool and quiet. But to clean the fins on both sides meant removing the motherboard. Something I never did. I just hit it with compressed air whenever I had a reason to open the case.

I now have an Asus X99-A with an i7 5820K. I still have the same R4. This time I decided to go with a CLC and ended up going with a small but chubby X41. I mounted it in the bottom front with two Noctua Redux 140 fans in a push/pull configuration. I chose it for convenience. The CAM software is a little bloated but isn't required and reports higher CPU temps than RealTemp does.

I don't fall into either cooling camp. This time around I wanted to try something different and didn't want the hassle of a big heatsink. I read many, many AIO reviews and was using a cheap Hyper 212 EVO (not the Pro version) in the interim as the Haswell-E does not have the Intel heatsink and fan.

I didn't expect it to be any better than air. Just as good as air. I just installed it last weekend. I also picked up a new PSU and ran out of time to do any decent cable management. So right now one case panel is off. So I can't report any temps until I close it back up. I also am running stock speeds since the Hyper 212 couldn't keep up even at stock without running in the low to mid 30s even at idle. In all fairness it only had one 120 fan. So basically I am starting from scratch with the board and the X41.

I like that I can run it in Performance, Silent and Manual mode. I like that I can run 3rd party fans and monitor them in the CAM software. I removed both of my drive cages a few days after I got my case as I just use SSD drives as I also have a Synology NAS, drive dock and a WD My Book. So for me, that open space was a perfect spot to install the radiator. Haven't noticed any ill effects of having it below the block. I can say that if there was a case as sleek and quiet as my R4 I would buy it if it had room for a top mounted X61. I have physical room for it but the hole spacing is wider and does not line up.

For that processor, I also would favor either an AiO or custom-water cooling strategy. And while the 5820K or other E-processor could be a temptation to use a full-tower case if only to hold the radiators and reservoir, any project that incorporates a compact case has my attention.

For the 4790K Devils Canyon, I'm inclined to think a top-end air-cooler would work well. Not "best," but good enough for some over-clocked thermal profiles. But -- among the available air-coolers -- you'd want to choose among "the best." You'd do just as well to use at least an H80 or the NZXT X41.

 

[MoInSTL]

I do have an X41. Did you even read my post? I ran on air on my last build. I said this time I wanted to try something differnt.

I may pick up an R5 so I can front mount an X61.

I want a CLC. Not interested in all that is involved with a custom loop.

I don't understand your Devil's Canyon reference at all.

 

[BonzaiDuck]

I do have an X41. Did you even read my post? I ran on air on my last build. I said this time I wanted to try something differnt.

I may pick up and R5 so I can front mount an X61.

I want a CLC. Not interested in all that is involved with a custom loop.

I don't understand your Devil's Canyon reference at all.

Yes -- I did read it -- I was using it as an example. You must have misinterpreted me.

Since we're discussing "tips on AiO coolers," there's a slight difference in acceptable preferences between the 5820K "E" with the indium solder fabrication and the 4790K Haswell with its polymer TIM. But the 5820K is going to put out a lot more heat, anyway -- starting at the stock 140W TDP.

I was saying (or meaning to say) that choosing an air-cooler for a 5820K setup might be adequate for stock settings and even some overclocking, but either AiO or custom-water would be the preference for anything above that.

 

[MoInSTL]

I have no idea yet, but more than likely I have an average chip. They aren't binned the same way like back in the day. I am shooting for a modest OC as it's fast already. I like the clean look of an CLC/AIO.

Mounting the X41 in the front was a bit of a hack. The R4 has two 140 fans in front that snap in a tray that flips out. I just pushed hard enough on the screw from the front side of the fan enough to get a few threads through the top of the radiator. The plastic fan bracket was flexible enough to allow it. The bottom of each bracket have little nubs that also hold in the bottom of the fan holes in place. But the two screws at the top were more than sufficient. Edit: I don't think two screws at the top not fully seated would hold a 280.

I really wanted the X61. Actually I wanted the Fractal Kelvin but they are not sold in the US or Canada due to patent issues. So I have an R5 and X61 waiting for me at Micro Center. The X61 IMO will do a better job as it's a 280 versus a fatter X41 (36mm thick). It's only $20 more. So I will return the X41.

I love my R4 but it's no longer suited for my needs. I prefer the Titanium grey but MC only has black so that's fine. I also got a new PSU and hadn't finished the cable management, so it works out I didn't.

FWIW, with the side panel off, running in Silent mode my temps are 32C. Ambient is 25C. The fans are at 30% and fluctuate around 480 RPM. I am not going to close it up to get a more accurate reading. It may have been better or not with it closed. I just don't think that's good enough.

 

[BonzaiDuck]

I have no idea yet, but more than likely I have an average chip. They aren't binned the same way like back in the day. I am shooting for a modest OC as it's fast already. I like the clean look of an CLC/AIO.

Mounting the X41 in the front was a bit of a hack. The R4 has two 140 fans in front that snap in a tray that flips out. I just pushed hard enough on the screw from the front side of the fan enough to get a few threads through the top of the radiator. The plastic fan bracket was flexible enough to allow it. The bottom of each bracket have little nubs that also hold in the bottom of the fan holes in place. But the two screws at the top were more than sufficient. Edit: I don't think two screws at the top not fully seated would hold a 280.

I really wanted the X61. Actually I wanted the Fractal Kelvin but they are not sold in the US or Canada due to patent issues. So I have an R5 and X61 waiting for me at Micro Center. The X61 IMO will do a better job as it's a 280 versus a fatter X41 (36mm thick). It's only $20 more. So I will return the X41.

I love my R4 but it's no longer suited for my needs. I prefer the Titanium grey but MC only has black so that's fine. I also got a new PSU and hadn't finished the cable management, so it works out I didn't.

FWIW, with the side panel off, running in Silent mode my temps are 32C. Ambient is 25C. The fans are at 30% and fluctuate around 480 RPM. I am not going to close it up to get a more accurate reading. It may have been better or not with it closed. I just don't think that's good enough.

There should be all sorts of options to make fans and radiator "fit" without any modification, "bending," anything of that nature. If I had to myself, I'd use wire ties - the nylon items with buckles -- to secure fans or even a radiator. One would only hesitate if anticipating frequent removal and replacement. If screws are too short, you should be able to find some longer with the same threading.

If those are idle temperatures (32C) for a 5820K processor, there's nothing wrong with that. My 2700K @ 4.7 idles closer to 38 to 40C. Idle temperatures aren't all that significant to anything. Also, some people obsess over noise more than I do, but I do my best to reduce it. If the fans run at 480 RPM with idle temperatures, I think that's a speed as low as you'd want.

 

[MoInSTL]

Started working on the R5 with the X61. Just need to clean up some cabling. Used the stock TIM since it was already applied. Put it in Performance mode to get the liquid moving and check for leaks. In Slient mode, in RealTemp, with the front door closed and ambient temperature of 29C, I am getting 29-30-28-28 on idle. That's more like it. It's very close to what I was getting with a Macho HR-02 on a 2600K Sandy Bridge. With the door open, drops by a couple of degrees but the fans are more noticeable. Btw, if I was going for silent I would go air. I just want it quiet and the R5 helps.

Just remember this whole build is new except for my SSD drives and burner. I needed to wait and see what long term cooling option I wanted to go with. I had a cheap, stop gap air cooler as they no longer ship with any. So I just used it and didn't even pretend to try to OC with it. So now I can do Prime95 and use it without worrying about the air cooler. So yes, I agree, idle temps are next to useless but until a few hours ago I didn't have an option to do much. The X61 performs better that the X41 for obvious reasons. But I tried it for a week and then swapped it out for the bigger brother in a case which better supports water cooling.

I ordered another pair of Noctua fans which will arrive tomorrow. The NZXT fans are better than I expected and others may keep them. I can hear them though and while not horrible like some stock fans are, I just want to use the paired Noctua fans in push/pull. One pair was very quiet in push/pull on the X41 in the R4. I won't have time to install them until the weekend rolls around again.

For any R5 owner considering this CLC, I was able to keep the optical cage in and had just enough clearance to install the radiator in the front. Nothing cramped. Hoses come out nicely. For R4 owners, the X41 fits in the bottom front with a little finessing. I suggest making a trip to the hardware store to get fan/rad screws that are flat as the rounded screws poke out just enough to make it hard to get the dust filter in and out.

-->BonzaiDuck, it's secure and it normally isn't moved around so no need for anything else. It's pretty lightweight.

 

[BonzaiDuck]

I am getting 29-30-28-28 on idle.

[

It's an i7-5820K Haswell-E processor . . . right? With six cores?

Sometimes I'm too quick to focus on nit-picky details.

I recall from setting up some ancient motherboards that you had to "enable" all the cores in BIOS -- for some reason, the default was less than you wanted.

I'll need to get my own X99 motherboard and processor to see what's going on with this firsthand.

But the idle temperatures are less relevant to anything if they're within a respectable range, such as yours. Proof of the pudding will be the temperatures of six cores under stress-tests. The X61 should be more than sufficient up to some higher-than-stock clock speed.

 

[MoInSTL]

AFAIK, RealTemp only supports 4 cores.

Believe me, I checked to make sure everything was on that needed to be on and turned off some obvious things. FWIW, in my 18 years of building my own PC alternating between Abit and Asus boards (except for ASRock Sandy build) I never experienced a hardware problem. I have 4X4 DDR4 configured in quad mode. One DIMM slot was bad. The other slots showed up. Tried swapping it with one that was in The BIOS. Then another. It was just a bad slot. Exchanged it the next day. Well worth paying sales tax sometimes at Micro Center.

In addition to getting your hands on an X99 motherboard and processor, add DDR4 to your list.

 

[YBS1]

You need to look into the RealTemp folder and use the RealTempGT executable instead to support 6 core monitoring. It seems to only support up to six though so I've had to slide over to CoreTemp for the 5960X.

 

[MongGrel]

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.


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


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.

Just several reasons I've never tried a AIO personally.

If I were to do water myself, I'd build a dedicated serious custom loop, and still put some low noise fans in certain areas.

Just my two cents.

Still have a D14 in the main myself.

 

[BonzaiDuck]

Just several reasons I've never tried a AIO personally.

If I were to do water myself, I'd build a dedicated serious custom loop, and still put some low noise fans in certain areas.

Just my two cents.

Still have a D14 in the main myself.

Truth be told, if I choose to build a custom-water rig, it will be a "first." If I build an AiO rig -- it will be a first. So I'm debating whether to mod an existing case -- a CM Stacker midtower -- or simply buy a new case. And debating between AiO and custom.

Per Aigo's points which you quoted, the problem with mosfet, VRM and mobo cooling can be resolved with a water rig of you keep the traditional exhaust fan in place. Put a box over it; fit a cut-out plate to the box so that it covers the CPU area with heatsinks and VRM parts. It could be really simple; you'd have slots for the hoses -- it just has to fit so removal isn't a pain. You could do it with Lexan -- add some "bling" lighting -- or art-board. Just a refinement to the Sabertooth idea, and you wouldn't need to cover much more of the board than between the PCI-E x16-1 graphics card and the upper end of the I/O ports or top edge of the mobo. But maybe you could integrate air-cooled graphics cards into the equation.

If you had radiators at both intake and exhaust -- places with 2x120 = 240mm to fit identical radiators, you could have both intake and exhaust cooling. The rear fan would only be necessary to draw air across the board itself. It wouldn't matter how much "pressure" there'd be in the case -- you'd just have sufficient exhaust where you needed it, in addition to the mobo-cooling and rear fan.

Heck. I think I'm going to look at that Swiftech H240X again.

 

[MoInSTL]

You need to look into the RealTemp folder and use the RealTempGT executable instead to support 6 core monitoring. It seems to only support up to six though so I've had to slide over to CoreTemp for the 5960X.

Thanks for the tip. Where did you find CoreTemp that was adware free? I used to use it but it seems problematic now as some stuff sneaks through even after choosing a custom install. Thanks again.