Comments Locked

62 Comments

Back to Article

  • Edkiefer - Friday, April 26, 2013 - link

    really nice review with air cooler verse the water coolers . One comment though, looking at the air temps of PWM and 100% . They seem so close I would think you should get better results on 100% fan . Maybe the case still doesn't have good airflow for air coolers ?

    Just saying, I have the hyper evo and with stock MB fan profile verse modified (maxes to about 80% ) I say a least few c with app like prime95 .
  • biostud - Friday, April 26, 2013 - link

    You could do a nice xy diagram with noise and delta temp on the axes.
  • truprecht - Friday, April 26, 2013 - link

    "You could do a nice xy diagram with noise and delta temp on the axes."

    Yes this 100x. It's so obvious - why is it not SOP for cooler comparisons?
  • DanNeely - Friday, April 26, 2013 - link

    I put one together for a prior cooler review; it wasn't nearly as useful as a I hoped. With very few exceptions everything bunched up in a fairly narrow diagonal line.

    Noise is vertical; temp horizontal: http://orthogonaltonormal.com/midden/fans.png
  • JCheng - Friday, April 26, 2013 - link

    DanNeely: On the contrary, I find that extremely helpful! Having to jump back and forth between the data points and the legend is kind of a drag, but to see the two dimensions really helps!
  • DanNeely - Monday, April 29, 2013 - link

    The reason I felt it was of limited value was that excepting some really bad performers on the quiet end of the range (designed for low power CPUs in SFF systems?) almost all the coolers fell along a relatively narrow horizontal line; meaning the best to worst ordering in the temp and noise tables was mostly equal with no major outliers.
  • nail076 - Monday, April 29, 2013 - link

    I agree, an X-Y chart of these values would better show the best performers in a sea of coolers.
  • buhusky - Friday, April 26, 2013 - link

    Why do fan & cooler reviews NEVER put the OEM fan/cooler numbers in there as a baseline? Never! It would be much better for me to compare to see how different of an upgrade the item(s) would be compared to what came with my stuff, not just as compared to each other....
  • A5 - Friday, April 26, 2013 - link

    The OEM coolers are terrible. They would make the graphs unreadable because all the aftermarket stuff would hardly look any different compared to the OEM cooler.
  • matagyula - Friday, April 26, 2013 - link

    In this case I don't think the OEM solution would keep the CPU cool enough -> i7 @4,4GHz, as written on the "Testing methodology" page.
  • DanNeely - Friday, April 26, 2013 - link

    In the last test it was 6C hotter than the worst cooler in this review. That would put the core temp in the high 90s and possibly result in thermal throttling.

    http://www.anandtech.com/show/6830/cpu-air-cooler-...
  • tsponholz - Friday, April 26, 2013 - link

    It would be nice to see this a baseline.
  • Hrel - Friday, April 26, 2013 - link

    agreed, just mention the numbers outside of the graph, so it doesn't fuck up the comparative look.
  • Torrijos - Friday, April 26, 2013 - link

    It's a little sad not to have the H110 on that test since it tends to be quieter than the X60 for the same level of performance.
  • JustMoreFun - Friday, April 26, 2013 - link

    It's very sad that you up to now didn't test one of the Thermalright Coolers, as they are commonly referred to as being the reference when it comes to air coolers.
  • davidthemaster30 - Friday, April 26, 2013 - link

    If the NH-U14S, was mounted so that it pushed air towards the top of the case, would it still block the PCIe slot?
  • marc1000 - Friday, April 26, 2013 - link

    I don't believe it would fit inside the case in this position... it looks like it would go past the backplate.
  • epoon2 - Friday, April 26, 2013 - link

    You should check with Noctua, they have an FAQ listing motherboard compatibility
  • spidey81 - Friday, April 26, 2013 - link

    First I'd like to say that I thoroughly enjoy your articles. In your recommendation of the U12S/U14S over the D14 I think you may have missed something. At your current 4.4 Ghz overclock the smaller/cheaper heatsinks performed, let's say, more efficiently. However, wouldn't the D14 be able to handle a higher thermal load that come with higher overclocks? So it's kind of like you said, it depends on your usage. It would be interesting to see at what point in overclocking would the D14 become worth purchasing over it's smaller siblings.
  • epoon2 - Friday, April 26, 2013 - link

    it scales in the same direction for all coolers as temperature increases, common sense or wrong?
  • spidey81 - Friday, April 26, 2013 - link

    This article was to test the performance of the coolers at a certain clock speed. However, I think you would find that when pushing the overclock higher you would see some coolers wouldn't be able to keep the temps in check while others could. There's a difference between efficiently transferring the heat and the amount of heat capable of being transferred. That's why the D14 and TRUE are popular among overclockers who push higher than what we see in this test. To that end, the 240mm or 280mm closed loop coolers will also be able to handle higher overclocks.

    In this (http://www.overclock3d.net/reviews/cases_cooling/s... review of the H220, you can see that on the socket 2011 cpu the air coolers weren't able to maintain the extreme overclocks like the closed loop coolers could.
  • spidey81 - Friday, April 26, 2013 - link

    sorry, link broke. http://tinyurl.com/cu24bwk That should work better :)
  • epoon2 - Friday, April 26, 2013 - link

    You meant that as the load/watt/heat produced by CPU increases, the order in this graph http://images.anandtech.com/graphs/graph6916/54390... changes?

    but in the above test, the fans are at constant, max speed already ... I know what you mean but I just can't figure out the science/reason behind why your scenario would happen.
  • BrightCandle - Friday, April 26, 2013 - link

    Two key factors are at play:
    1) the amount of watts of heat the solution can dissipate at a given delta temperature to the air.
    2) the thermal interface performance to the CPU itself and hence what temperature the CPU is running at.

    Water is cooling the water block but then water itself is above ambient temperature. This often means the ambient temp CPU effectively sees is higher compared to what it was with air. But water warms and cools much faster than air by volume, and it can hold a lot more energy. Combining water with a substantially larger area to exchange its heat to air and you get a solution that cools better given a large number of watts of heat but does less well on low wattage where air cooling fits and is sufficient.

    A 2600k pulls about 170W over clocked. A 3930k can be over 300W. 170W is about the point to consider water but you need an overspecced loop to show any advantage at all. Passed 200 watts however and water takes a clear lead and air stops being able to do the job.
  • epoon2 - Saturday, April 27, 2013 - link

    Assuming the water pipes dissipates negligible heat, the role of the water system is equivalent to that of the thermal compounds: to transfer energy and heat to the sink and radiator.

    The difference in rate of heat exchange between the metal pieces and the external environment becomes the key.

    In an open air test, the rankings should remain the same as amount of heat generated by the CPU increases. In the real world , the performance of air coolers depends on the effectiveness of the case's heat exchange system..
  • A5 - Friday, April 26, 2013 - link

    For future reviews, I wouldn't mind seeing the Xigmatek SD1283 tested. A lot of people recommend it for people who want something better than the CM Evo but don't want to spend Noctua money. It typically runs ~$50 so it fits that niche price-wise as well.
  • A5 - Friday, April 26, 2013 - link

    I guess I meant the silly-named "Dark Knight 2" when I wrote this. Didn't know they had 2 different SD1283s. Guess I wouldn't mind seeing the cheaper, non-coated "Gaia" as well.
  • Dr_b_ - Friday, April 26, 2013 - link

    Does the U12 have socket 2011 mounting capability?
  • epoon2 - Friday, April 26, 2013 - link

    http://www.noctua.at/main.php?show=productview&...
  • Dr_b_ - Friday, April 26, 2013 - link

    thanks 4 link. NH-D14 is really too big, blocks a socket in my x79-UP4 gigabyte. U12 looks like it might work here.
  • epoon2 - Friday, April 26, 2013 - link

    wow that was fast, maker updated with editor's choice logo already btw
  • tsponholz - Friday, April 26, 2013 - link

    I don't understand why the Thermaltake WATER2.0 Series doesn't seem to make it to these lists. I have the Performer and love the performance and noise. The review I have seen (never in a round up) put it on top of the Antec and Corsair offerings.
  • karasaj - Friday, April 26, 2013 - link

    Any good places to get the NH-U12S in the US for that 65$?
  • Hrel - Friday, April 26, 2013 - link

    Man, that 30db noise floor is REALLY becoming a problem. You guys would CLEARLY get value/use out of better equipment. I would REALLY REALLY REALLY like to see that happen.
  • Razorbak86 - Friday, April 26, 2013 - link

    Feel free to make a contribution of said equipment to the site's standard test setup.
  • DanNeely - Saturday, April 27, 2013 - link

    When you're buying something to donate; keep in mind getting ambient noise levels much below 30db is easier said than done and an anechoic chamber large enough not to turn into a hotbox during testing isn't going to be cheap.
  • TrackSmart - Friday, April 26, 2013 - link

    Great review. Kudos to Cooler Master for giving us cheapskates something to play with at $33 that will handle some significant overclocking, while also staying quiet. I've usually dismissed aftermarket coolers as being a poor return on investment for mid-range builds (i.e. it is more cost effective to spend the money on a beefer CPU than a bigger cooler). But $33 is a price point where it might make sense to invest in better cooling over a modestly higher-end CPU.
  • crimson117 - Friday, April 26, 2013 - link

    Not sure how you can give a bronze award to such an expensive item - price/performance value simply isn't there. It's 3-4 times more expensive than the Kraken X60 or Noctua but doesn't offer nearly that much improvement.
  • Razorbak86 - Friday, April 26, 2013 - link

    Huh? Where are you getting your pricing? o.O

    In the US, the Swiftech H220 is $139.99, and the Kraken X60 is $109.99. Maybe my math is a little rusty, but I calculate the price premium to be 27% higher (i.e., [$139.99-$109.99]/$109.99 = 27%). If it was 3-4x higher, it would be priced at $329.97-$439.96.
  • Treckin - Friday, April 26, 2013 - link

    Dustin! you promised to include the Antec Kuhler pieces in your closed loop reviews!

    Hope thats still planned!
  • Rogerdodge1 - Friday, April 26, 2013 - link

    are these really the delta temps in Celsius above room temp? if they are then even at a room temp of 21.11C (70 F) you have gone way beyond the temp spec on the processor(72.6C at the heat spreader) with several of your air coolers....even if these are absolute temps Celsius some of the air coolers are dangerously close to max temp. I understand pushing the overclock to test the coolers, but subjecting the chip to those temps is bound to kill it a lot quicker. i suppose if this is all you use it for it doesn't matter much, but damn i would hate to do that to a chip i paid for.
  • Bobs_Your_Uncle - Friday, April 26, 2013 - link

    I've remained extremely intrigued with both the design & the potential efficacy of a prototyped cooler noted within Anandtech ( http://forums.anandtech.com/showthread.php?t=21782... ) & elsewhere ( http://www.tomshardware.com/news/cpu-cooler-sandia... --- https://www.youtube.com/watch?v=JWQZNXEKkaU ). It's been quite some time since initially publicized & I wonder if you might know where to research for an update on it's status?

    While the concept has many possible applications, the thought of it's implementation for PC component cooling is enough to excite one into debilitating nervous ticks. (Well, maybe it's not quite that exciting, but ..... I'm still interested on where it is in development!)
  • politbureau - Friday, April 26, 2013 - link

    Hey Dustin, where's the H110? I understand you'd like to limit the test to 'current' coolers, but the H110 is still readily avaialable at retail, and it seems coutnerproductive not to include what should be the top performaing CLC is this roundup...
  • Dustin Sklavos - Friday, April 26, 2013 - link

    The H110 and Kraken X60 are almost the exact same product, just different fans (understanding that Corsairs fans are superior) and different control (the Kraken X60 uses USB and software, the H110 relies entirely upon your motherboard). But they're the same Asetek radiator and I suspect the same pump.
  • Guspaz - Friday, April 26, 2013 - link

    I recently tried to replace my Shuttle's 92mm fan with a Noctua NF-B9 PWM, which was a very disappointing experience. Yes, it was quiet, but it also spins extremely slowly at max speed, with a low static pressure and a very low airflow. The end result was an increase in CPU load temperatures of 40 degrees (celcius) or so, which is pretty damned massive. Since the CPU fan in a Shuttle doubles as the case fan (the CPU has a heatplate which connects via heatpipes to a large heatsink on the rear of the case, onto which a 92mm fan blows air through it, cooling the CPU and exhausting air from the case at the same time).

    The problem is that the stock fan, while decently quiet up until 35-40% speed, has this point around 40% speed where it very quickly gets very loud (motor noise, not airflow noise). Some research showed that there are only a tiny handful of fans on the market that spin anywhere near as fast (Delta had one, that's about it), and they're a tad pricey to buy on a "hope it's quieter at medium speeds".

    Is it really so much to ask for a fan that is quiet at lower speeds, but can still spin fast (loud if required)? I don't care how loud the thing is at max speed, because if I'm gaming I've got headphones on anyhow. But during normal use (or while trying to sleep) the occasional CPU spike pushes it just past the 40% mark and even the 5% speed bump causes an audible revving.
  • epoon2 - Friday, April 26, 2013 - link

    to your point on quiet at lower speeds, but can still spin fast (loud if required) -

    It would be more expensive, and the fan may be larger, and the original target audience would flock because their original wish was getting a quiet fan at all times.
  • FH123 - Sunday, April 28, 2013 - link

    I'm a bit out of the game, but I think the Noctua fans that only have 6 blades, spaced quite far apart, have a reputation for low air pressure and don't work well when the CPU heat sink has lots of closely spaced fins. I also use a Noctua fan (see below), but it's a model with more blades and more closely spaced ones.

    In general I agree with you. Fans that run at low speeds by default are a silly marketing phenomenon, which are only useful when you don't have a fan controller, no motherboard fan headers or can't get to grips with the freely available Speedfan program. What you really want is a fan that will still actuate and won't stall at low speeds (say 600RPM), but can ramp up to (much) higher speeds, if needed.

    Another phenomenon I find with my own fan / case combination is that there are certain speeds where the fan noise takes on a distinct and annoying pitch. I don't think it's motor whine, but perhaps some sort of whistling effect produced by the fan holes in the case. This happens at 1,300RPM and around 1,050RPM for me, but neither at 900RPM nor 1,200RPM. I therefore find it very important to be able to regulate the fan speed.
  • StitchExperimen - Sunday, April 28, 2013 - link

    I would like to see the replacement of fans with 9 blade fans and see if a lower rpm and higher air flow makes a difference. I took two 9 blade fans in push pull on a Corsair 80i on a i7 3770 because the stock fan noise was so loud/bad. Possibly with this is what I used in push pull from newegg >>> COUGAR CF-V12HPB Vortex Hydro-Dynamic-Bearing (Fluid) 300,000 Hours 12CM Silent Cooling Fan with Pulse Width Modulation (Black)
  • FH123 - Sunday, April 28, 2013 - link

    Are there any fan ducts still on the market? I built my PC years ago using a Thermalright HR-01 with the optional fan duct that connects the heatsink to the rear 120mm case fan. This arrangement seems to work very well, which perhaps ties in with your observation that the case fan is very important for air coolers. Fan ducts also seem to be common in commercial designs from the likes of Dell. I don't even use a CPU fan, only the case fan.

    If it's of any interest, my system is built in an Antec Solo case, optimised for low noise. The only fans are the rear case fan, the power supply fan (at the top of the case) and the GPU fan (ATI 5850 radial fan). There is no front intake fan. The Q9650 CPU has a mild overclock from 3 to 3.6GHz, while remaining at stock voltage. The case fan, at 1,200RPM, keeps the CPU within 45 to 50C over ambient during a Prime stress test. It runs at merely 900RPM during lighter loads.
  • flemeister - Sunday, April 28, 2013 - link

    You can easily make your own with cardboard and tape. I've found a three-sided duct to be the best for connecting a CPU heatsink to a rear fan (with the open side facing the motherboard). It still pulls plenty of air through the CPU heatsink, but also allows air to be pulled over the motherboard VRMs.
  • disappointed1 - Sunday, April 28, 2013 - link

    With all due respect, this testing methodology is now completely flawed:

    "For air coolers, I added a Noctua 140mm rear exhaust fan and used the ultra low noise adaptor to ensure it didn't affect acoustics in any meaningful way. This is in line with the usage cases air coolers are designed for, and should be representative of the kind of airflow most users will have from their exhaust fan."

    You FUNDAMENTALLY can't compare coolers on the same charts with different testing conditions. The closed-loop coolers are just as much designed, and will be operated, with proper/equivalent case ventilation. Just test them under identical conditions and let the liquid coolers pay any penalty with higher idle noise readings.
  • epoon2 - Monday, April 29, 2013 - link

    I tried the article again, couldn't find where Dustin mentioned his testing method for Water. On the page where he shows the Seidon, it's clearly installed inside the case. I do not believe there is a strong bias towards either air or water coolers in this test.
  • disappointed1 - Tuesday, April 30, 2013 - link

    "...I'm now using that exhaust fan for testing air coolers. Closed loop coolers continue to do without."

    He added an extra case fan for the air coolers, which was not present for the liquid coolers. This will have the effect of biasing the results and renders them void. This is readily apparent by the author's own admission that "the differences in performance were pronounced" and "now liquid coolers aren't the juggernauts they used to be" and cannot be compared with previous results.
  • Alvar - Monday, April 29, 2013 - link

    Today we have something a little special on the table. If you have previously been reading our CPU Cooler reviews you probably saw our recent review for Silverstone. We reviewed the Silverstone Heligon Series – HEO1 CPU Cooler....
    more details:- http://tinyurl.com/c5czh4b
  • bobbozzo - Tuesday, April 30, 2013 - link

    Spam link above goes to a site for a women's magazine, nothing about coolers.
  • TheStranger81 - Tuesday, April 30, 2013 - link

    It would be a review if it actually had any charts....WTF is going on ? Where are the charts ?
  • Wwhat - Saturday, May 4, 2013 - link

    Look at how this site does such things:
    http://www.computerbase.de/artikel/gehaeuse-und-ku...

    (language is irrelevant for the subject of graphs)

    You can deselect items in the list and when you select a line you can see the position and details as you move the mousepointer over it.
    And in their bargraphs it uses mouseover to show the percentage and relative percentage.
    Like shown here: http://www.computerbase.de/artikel/gehaeuse-und-ku...

    PNG's for data are a bit.. outdated really aren't they? (But perhaps you need to dumb down again for tablets and phones these days?)
  • Death666Angel - Saturday, May 11, 2013 - link

    Can you adjust the air and clc results so that the x-axis is the same unit length? That way it is easier to compare between the two cooler types. :)
  • hooner - Tuesday, July 2, 2013 - link

    Great review...thanks!

    I have a quick question...I am thinking of buying a cooler master N200 and is front or rear radiator cooling best? I am presuming the front fans are intakes, rear and top are out.

    My thinking is rear takes heat straight out the back from CPU, where as front means air is drawn in, cooled and then blow back through case. Surely venting the heat straight away is better?

    Cheers
  • SloppyFloppy - Wednesday, September 17, 2014 - link

    You keep recommending the really well performing Swiftech H220, but after some research they appear to have server quality control issues with their pumps failing and/or making lots of noise as well as some of their fans being noisy.

    Now I don't know what cooler to buy that performs well without sounding like a lawn blower.
  • balazer - Sunday, January 26, 2020 - link

    I believe these noise measurements are flawed. Some of these air coolers should be much quieter on an idle system with PWM than at 100%, but the measurements show them to be within 1 dB of the same cooler at 100%. And quite a few of the different coolers are all listed at the same 30 dB. It suggests the ambient room noise and/or noise from other components in the system is around 30 dB whereas some of the coolers can be quite a bit quieter than that, thus making all of the measurements around 30 dB suspect. Also it wasn't mentioned whether the noise measurements were A-weighted or not. (To be useful they probably should be)
  • balazer - Sunday, January 26, 2020 - link

    Or maybe the noise measurement device has a high noise floor.

Log in

Don't have an account? Sign up now