Comments Locked

36 Comments

Back to Article

  • zeeBomb - Tuesday, December 15, 2015 - link

    Awesome cooler
  • HollyDOL - Tuesday, December 15, 2015 - link

    So, with a SkyLake, you can cool at about 10°C above ambient at 35dB? That's definitely NOT BAD outcome.
  • DanNeely - Tuesday, December 15, 2015 - link

    Almost certainly not; you'll see temps significantly higher. The heater blocks E. Fylladitakis uses to simulate a thermal load that's more stable than you can get with a real CPU have a major failing that makes them impossible to equate to performance on an actual system useless. The temperature reported on your CPU comes from a handful of tiny hotspots on the die that do a lot more work, consume the majority of the power, and are much hotter than the rest of the die.

    The hot spots are why, although TDPs have stayed more or less fixed on desktop chips since the Core 2 era, actual CPU temperatures have been on a steady upward swing. Each new process shrinks the bits of the CPU that do the actual calculations and generate the largest part of the heat into a smaller chunk of the die. The extra die space gets filled up with cache which doesn't have hot spots or GPU space that because they have large numbers of tiny cores instead of a few big ones can spread out the heat a lot better.
  • DanNeely - Tuesday, December 15, 2015 - link

    The CPU waterblock in this kit looks a lot like the EK Supremacy EVO in my custom loop. That block is explicitly intended to be taken apart and reassembled prior to installation (it has several parts intended to control water flow that are intended to be swapped for different CPU models), Unless you're an idiot who can't be bothered to read directions it was strait forward to follow the printed instructions to open it up, swap in the parts for my CPU and then put it back together.

    That experience also has me concerned about your testing methodologies results will correlate to real world relative performance. The intent of adjusting all of those fiddly bits in the block was to direct the flow to maximize cooling of the hottest parts of each specific chip; something that can't be captured by a simple resistive heating block. Testing I saw a few people who were curious if it really mattered showed spreads between different internal configurations of several degrees; comparable to the total spreads you're seeing among decent water cooler kits.
  • londiste - Tuesday, December 15, 2015 - link

    The extra header on the fan controller is with Predator 360 in mind.
  • tabascosauz - Tuesday, December 15, 2015 - link

    This entire article was written with the assumption that the only benefit offered by the Predator is its expandability, as with the H220-X. That's not the point of the kit.

    The point is that when you buy the Predator, you are basically buying EK products, with the added benefit that someone has already assembled the parts for you. The CPU block is the same Supremacy (EVO?) block that you get off of the market for your custom loop, and so are the Vardar F4 fans and the Coolstream PE rad.

    It's not meant to be an end to the endless CoolIT and Asetek AIOs on the market. People aren't going to buy it in comparison to the H100i. They'll buy because it's a nice place to start one's future custom loop system.
  • tabascosauz - Tuesday, December 15, 2015 - link

    I suppose I'm expecting a bit too much from AT reviewers these days. Clearly, they don't do even the slightest bit of reading before starting their reviewers. Block appears to be small and doesn't have a pump like other AIOs? A bit of digging even into EK's product page for the 240 would reveal that the block is a Supremacy MX and obviously wouldn't have a pump integrated into it, like all proper watercooling CPU blocks (exception made to the Apogee Drive II).
  • forgot2yield28 - Tuesday, December 15, 2015 - link

    Your first sentence states that the point is not expandability, and yet your last sentence states that it will be purchased as a start to a future custom loop--ie users will add and modify it, in other words, expanding it. Did I miss something?
  • tabascosauz - Tuesday, December 15, 2015 - link

    Ach, never mind. The review is adequate.
  • owan - Wednesday, December 16, 2015 - link

    Yea, you missed the fact that its *not only* expandability. There are other products that offer expandability, but EK's offers expandability and quality components you'd probably be happy re-using.
  • thestryker - Tuesday, December 15, 2015 - link

    I greatly appreciate the review on this, and it seems to match what I've seen in other places so that's always good. Going over the little pieces is what I've come to appreciate most about AT reviews. Talking about pump noise is also helpful as it seems like the asetek/coolit ones all seem to have issues here.

    For me at least this cooler is worth it simply to not give money to asetek. It would be nice if AnandTech could do a quick writeup or something on which AIOs are using the asetek design, and in turn giving them license money. The patent they've been allowed to use as a weapon is pretty absurd and has put us in a sad state for AIO development.
  • iamkyle - Tuesday, December 15, 2015 - link

    The use of a DDC was an interesting choice...
  • Kid98 - Tuesday, December 15, 2015 - link

    Every 3dB's is a doubling of power....

    Kid
  • DanNeely - Tuesday, December 15, 2015 - link

    Good catch.

    While we're nitpicking the dB discussion though, the human ear has a log response curve; 10 dB is only heard as roughly twice as loud so the 12.3dB spread between the best and worst coolers is only a bit more than twice as loud even though the sound is roughly 17 times as intense.
  • Valkyrierie - Tuesday, December 15, 2015 - link

    The third fan header is meant to be used in EK-XLC Predator 360.

    EK-XLC 240 and 360 share the same power board despite 360 having an additional 120mm fan - Most likely, they stuck to one single board for both units to cut costs.
  • initialised - Tuesday, December 15, 2015 - link

    No mention of the DDC pump or it's specs, power consumption, flow rate...
  • wolfemane - Tuesday, December 15, 2015 - link

    I don't think power is really all that important with this kit. I'm currently using a koolance 450s running at settings for 26w @ 12v roughly doing 3.2gpm with my custom build. I did before and after power draw checks with an at wall kill-o-watt (I know it's not accurate, but it gives a general feel of whT your pulling) and I only saw a power increase of about 15w. Give or take a few watts. Flow rate is a little higher in use as well but not by much. So unless your really desperate to save on power due to over taxing a low end psu, the power draw of the pump/fans is going to be minimal on an AIO kit like this.

    And since the pump isn't an adjustable pump why would flow matter? It's traveling over one surface with a fairly large rad, as long as the kit is designed with decent flow and the chip is cooling, why worry about this uncontrollable spec?

    With these AIO kits I'd be more concerned with how its performance stands up to competition in regards to actual cooling and noise. Power, flow, pump specs just seem irrelevant. If this was a custom built system then I could see more attention needed for these areas.

    I thought it was a great review, I love EKWB and have used them since early Athlon. Nice to see a semi modifiable AIO from them. I even think the price fits the possibility seeing how there is an option for customization compared to other AIOs.
  • DanNeely - Wednesday, December 16, 2015 - link

    I'm surprised the pump's not adjustable. I can adjust my swiftech branded DDC pump using the mobo's fan controller software; it's a bit annoying since it's RPM curve isn't linear; but I was able to tune it to run at ~2/3rds speed (at most 1 or 2 C hotter temps but much quieter operation).
  • londiste - Wednesday, December 16, 2015 - link

    https://shop.ekwb.com/ek-xlc-predator-240
    Pump type: Laing DDC3.1 6W
    that is specific enough to find the rest of the specs.
  • wolfemane - Tuesday, December 15, 2015 - link

    Where can we get a set of anandtech labeled tools? Would love to add those to my tech tool box!!
  • Titillating - Thursday, December 17, 2015 - link

    We gave out a bunch of those AT-branded screwdrivers a while ago! They conveniently feature a reversible Phillips/flathead tip. We don't have a full set of branded tools, but if there is sufficient interest, we're happy to find a way to give away some more screwdrivers!
  • pliablemoosethebanned - Thursday, December 17, 2015 - link

    Put me on the list, I want one :)
  • wolf762 - Sunday, December 20, 2015 - link

    My first though was, cool where can I get one. If they were for sale I'd have bought a few to give as gifts this holiday season, depending on price of course =)
  • Allan_Hundeboll - Monday, December 28, 2015 - link

    I would love an AT branded screwdriver!
  • Dr_Orgo - Tuesday, December 15, 2015 - link

    Does H110 in the figures mean H110, H110i GT, or H110i GTX? If it is just the original H110 AIO cooler, that's a bit disappointing since the H110i GT is a direct upgrade over the original and is the most likely AIO alternative to the reviewed product.
  • satai - Wednesday, December 16, 2015 - link

    Comparison with some high-end air cooler (Noctua NH-D15 comes in mind) would be helpfull.
  • Ninhalem - Wednesday, December 16, 2015 - link

    The EK Vardar fans are comparable to the venerable Gentle Typhoon series as being the best radiator fans on the market. The Vardar fans are extremely quiet even when at high speeds (3000 rpm). I picked up one of the FF5 120 mm PWM fans when they first came out for some viability testing and couldn't believe how quiet they were at full speed. It's no surprise to me that they have excellent cooling ability with low fan noise.
  • eckre - Wednesday, December 16, 2015 - link

    (Facepalm) Why don't you compare this to AIR? Noctua 14/15 blows away almost all of these coolers in performance AND sound.
  • Ninhalem - Wednesday, December 16, 2015 - link

    And why would I want something clunky like the Noctua 14 hanging off my motherboard when I have space restrictions that a water cooling system like this product could solve handily?
  • satai - Sunday, December 20, 2015 - link

    It would be useful to get the comparison. Air can be better for your setup, Water can be better for your setup - details matter. But it doesn't make the performance and noise comparison useless.
  • shaolin95 - Thursday, December 17, 2015 - link

    er no...its the top Air Cooler but no it does not "blow away almost all of these coolers" plus not everyone likes such huge thing on your build.
    I used to my tastes got refined and no longer like that huge chunk of metal there not to mention that its harder to clean between fins compared to a radiator.
  • stren - Thursday, December 17, 2015 - link

    Yes using a lab grade setup like that is great for more accurate system measurements. It's damn hard to test radiators well after all - ask my buddy fast_fate about that. *But* a good amount of performance can be gained or lost from bow of the CPU block and how well it matches the CPU's IHS. One thing you failed to mention in your methodology is how the block was mounted to the heater plate. Which TIM was used? How many mounts did you do? Mounts can produce quite variable results. Typically a 1C variance might be seen in mounting one block to one CPU in one orientation. Different CPUs and orientations will produce a wider spread still. In addition measuring ambient air in is extremely hard to do well even if your sensors are world class. How many sensors do you use, how spread out are they, how far from the fans are they? Is there an intake manifold? Is the manifold restrictive? Is there any other airflow in the room? How much does the ambient change? What's the error in the system? How long do you log for, how many runs do you average etc etc?

    If you were to test CPU water blocks independently on a such a setup the results would be very different to those testing on a real motherboard/CPU. This is particularly true when a small difference in mount height can change mounting pressure and hence TIM thickness. Some block mounts bottom out rather than relying on a set torque to tighten and hence making your own mount will mean a difference in mounting pressure than a real motherboard/CPU combo.

    In essence while you're testing *part* of the system very accurately, you're not testing the whole system accurately. While I applaud the effort to measure AIO's well, I think you'd do well to drop the lower power runs where the margin of error is > than the difference in the data in order to save time and also test with a CPU to try and see some impact of the mount on a real world system. Of course the latter can be frustrating to get the accuracy required. Taking data on ten water blocks accurately on one CPU can take me a month. So I can understand why you'd want to limit the variables, but it might be wise to mention that by limiting the variables in order to be more accurate you are also making the test less accurate by being a simulation of the real thing.
  • r4serei - Friday, December 18, 2015 - link

    i'm lovin my 240. i've already expanded onto it with a block for my gpu and another rad. was surprisingly simple.
  • r4serei - Wednesday, January 13, 2016 - link

    EK has issued a recall for all revision 1.0 Predators due to a faulty O-ring. It's been corrected on revision 1.1 which has been released as of Jan 4th.
  • alexrw - Saturday, January 30, 2016 - link

    > the pressure triples for every increase of 3 dB(A)

    err, doubles ... actually almost doubles (the exact value when it exactly doubles is not 3dB but 10*log10(2)=3.0102999 dB, or conversely the ratio for exactly 3dB is not 2 but 10^(3/10)=1.99526)
  • file2man - Sunday, October 30, 2016 - link

    I recently purchased the ekwb extreme 360 kit- it has a separate pump, reservoir, clear tubing for somwhere around $350 at microcenter. the pump is not variable speed like most other bought outside of kits and can only vary its speed by a fan controller. Im new at liquid cooling but I suspect hard tubing will in the longer term attract less sediment and is much preferred . Have not reached ek yet but not sure if I can use the 360 extreme kit fittings 13/10mm g1/4 for hard tubing literature says soft tubing is 9.5mm/12.7mm (3/8/1/2 inch). Any help is appreciated since is no way to email ek

Log in

Don't have an account? Sign up now