EKWB Predator 240 XLC Performance Testing and Review

rubix_1011

Contributing Writer
Moderator
EKWB 240 Predator XLC – Performance Testing and Observations

Rbjtgmd.jpg


q0bGNxk.jpg


Hey everyone, welcome back. I’ve recently had the opportunity to run the EKWB Predator 240 XLC through testing on the watercooling test bench. This has allowed me to get some very firm results on load testing, performance, noise levels and what might be the biggest question – how much can this cooler handle when expanded to include a graphics card in the cooling loop before you need an additional radiator?

I have been testing the 240 (2x120mm) version of this cooler, so while reading this, remember that the big brother, the 360, is going to offer even more potential due to 33% more radiator cooling area.

2Be51XW.jpg


Overview
The EK Predator 240 XLC is shipped as an enclosed cooling loop, but has the benefit of being capable of being expanded to include other components into the watercooling loop, effectively allowing you to buy a boxed cooler and later expand it into a custom cooling loop. EK has (wisely) opted to use standardized, G1/4 fittings to allow someone to disassemble and re-assemble the cooling loop to use common watercooling components with this cooler if and when the user wishes to do so. It utilizes a Laing DDC pump at its heart, which is a compact workhorse that is often employed by custom watercooling loops for both power and effectiveness. The pump is securely and discreetly mounted to the radiator itself and utilizes shiny, chrome EK compression fittings. Matte black EK-ZMT rubberized tubing is utilized for carrying coolant to and from the EK Supremacy water block.

qVtFJzv.jpg


The radiator housing is finished with a textured, matte black finish. The PE rad is a full copper core design, which should offer great heat dissipation and is also finished in matte black, although simply painted, not textured like the housing. The radiator housing thickness is right at 40mm, but the housing itself is almost 10mm thicker than the actual core itself, so while it does appear to be a nominally thicker format radiator, the core is approximately 30mm thick, which is fairly standard when compared to most available radiators used for custom loops. Including the fans, the entire cooler width is right at 65mm, so make sure you have enough clearance when you install the unit.

CrU6x6l.jpg


Fin spacing is right around 20 folds per inch (FPI) which makes for a fairly dense radiator fin layout. I will say that over the entire radiator field, all of the fin rows are very consistent and paint coverage is thorough. If you’ve ever purchased watercooling radiators, you often see bent fins, variance in fin spacing or areas of thin paint or exposed metal.

xGezez5.jpg


The EK 240 Predator XLC comes equipped with two Vardar 1900RPM 120mm fans mounted in pull. They are very neatly loomed (the term surgical comes to mind) as the PWM fan headers run through small rubberized retainer rings that are held down by the fan mounting screws, which incidentally, require an Allen key to remove. The fans themselves are matte finished black and grey and the entire color scheme is very industrial + technical in presentation. Silver EK compression fittings are used for every interface and rubberized EK-Tube ZMT ([3/8” or 9.5mm ID] & [5/8” or 15.9mm OD]) matte black tubing is used throughout.

Lzmsvfe.jpg

nuGwoLU.jpg

2Be51XW.jpg


The EK Predator 240 comes with the entire unit assembled and filled with glycol coolant, so it is ready to install out of the box, much like your typical closed loop cooler. It includes a tube of EK thermal paste, rather than a factory-installed ‘splotch’ already on the waterblock itself. You get the typical mounting screws and SATA power adapter with PWM support, which needs to be connected to the power distribution block, as well as an Allen key for the fan screws. The installation manual for most hardware is often not of much help or benefit, but in this case, EK has provided some excellent instructions and a multitude of possible future expansion possibilities when it comes to adding to your cooling loop down the road. It is very refreshing to see detailed descriptions for outlining these steps for someone new to watercooling loops.

YOETSH0.jpg

7lY5TyJ.jpg


A Look Inside
The EK Predator 240 XLC is obviously designed to be expanded and used as the base for a watercooling loop, if the user decides to do so. The EK compression fittings can be unthreaded and the tubing removed to add length to the tubing runs, to include a graphics card block and even an additional radiator or reservoir if you choose. Since the EK 240 uses G1/4 fittings commonly used for most watercooling components, this means that you can customize your Predator XLC with different fittings to match a color scheme or to find additional tubing and fittings to match your existing tubing and fittings.

YIWRf1B.jpg


I found the compression fittings on this unit to be very tight and required a pair of pliers to loosen. If you encounter this, be careful as you can cause damage to the exterior of the compression fittings – it shouldn’t cause functional damage, but no one likes scratches on their shiny fittings. I wrapped mine with tape before using pliers.
The coolant used appears to be a clear glycol that is almost certainly EK brand as they offer and recommend using EK coolants in all of their EK watercooling products. I found that the EK 240 holds just under a 300ml volume worth of coolant.

0dQxyeC.jpg


Closer inspection of the pump with the tubing and fitting removed shows a sticker denoting the specs of the pump: DDC3.1 PWM, 12vDC, 6 watt draw. DDC pump mods are commonplace in the enthusiast watercooling space, so I would expect this pump would be no different, but as that voids the pump warranty and falls outside of the boxed cooler tests I am doing, we’ll table that for now.

qVtFJzv.jpg


Flow rate testing shows just under half a gallon per minute flow, which is actually a little less than I was expecting, considering an unrestricted DDC nets almost 1.5 gallons per minute on the same test meter. This tells me that either the DDC might be slightly undervolted or that the EK radiator and Supremacy block might be more restrictive than I thought…or both. A little digging leads me down this path – the EK Predator 240 DDC pump has a sticker stating 6 watts, so I decided to test the load draw with the Kill-a-Watt meter. The power supply itself draws right at 6.6 watts without any components running from it. I setup the EK 240 to only run the pump by disconnecting both fan leads from the power block. The result? 9.6 watts of draw, or only an increase of 3w of draw by the pump of the EK 240. The standalone DDC pump? 15.5 watts on the Kill-a-Watt meter – a load draw of 8.9 watts and almost 3 times as much power draw, which would mean almost a tripling of power potential. Looks like pump load/draw (in watts) is the culprit and not necessarily component restriction.

fq2h71X.jpg


Thermal Test Setup and Methods
The EK Predator 240 XLC is meant to be a CPU + full watercooling loop, cooler. Knowing this, I expanded the thermal load tests this round to include loads that would simulate a single CPU at stock speeds, a single CPU at overclocked speeds and both of these tests including an added graphics card to determine the overall benefit of this system. My watercooling]http://www.tomshardware.com/forum/id-2829608/rubix-1011-watercooling-lab-equipment.html]watercooling lab equipment[/url] offers me the ability to simulate thermal loads accurately for different points of data to report on.

In this case, we’re testing thermal loads in watts at various intervals:

95w
130w
150w
175w
200w
250w
300w
350w
400w
450w
500w

Each of these load tests is done with the fans set at speeds 1200RPM, 1600RPM and 1900RPM (full speed).

Also, each thermal test and each speed test is done with the fans in both push and pull configuration.

All tests are run for 20 minutes, and a total of 64 individual tests were run. I also utilize a 20 minute thermal ramp-up time between thermal load changes; ex. 200w --> 225w load increase. While I know that the write up seems to be simple and directly to the point, there is a lot of work behind the scenes to collect all the data, in addition to charting and graphing the data, as well. (It's 32 hours just for the load testing; I know you're over there doing the math on those numbers.)

These thermal load tests in watts represent the TDP or thermal design power, of the components being cooled in the system. Whether that is a CPU, or a CPU and a graphics card together in a loop, the total TDP is what needs to be evaluated since you have the potential for hundreds of processors and their clock speeds as well as graphics cards and their clock speeds to be cooled by this cooler.

Cooling performance of watercooling is determined by the difference in coolant temperature against the temperature of the ambient air. This is called the temperature delta. The lower the delta or DT, the better performing the cooler at that load. In the world of water cooling, a 5C delta is considered excellent, a 10C delta is considered good and a 15C delta is average.

For more information, please see Understanding TDP and Understanding Temperature Delta or DT for more information.

Here is where having all those collected points of data comes in handy – a visible performance graph:

jwk67Nm.jpg


In every fan speed – 1200/1600/1900 RPM, there is a very linear set of tight clusters for each data point. The interesting thing to note is the negligible deviance between each load test and the set of push or pull tests associated – there is hardly any difference in performance between the fan orientations, but pull does have a slight overall edge across almost every test. However, it is typically only 1C with the largest gap coming at 1600rpm at 500watts of load…and this only results in a difference of 1.72C.

uRUknK3.jpg


Also, you’ll see that the radiator and fans at full speed can sustain a 10C delta at 300 watts of load. This means, boys and girls, that you could potentially add a graphics card and run a loop with an overclocked i7 4790k and a stock GTX 970 and still have yourself a 10C watercooling loop without the need for an additional radiator. You could even get away with staying within a 15C delta if your loop was 450w running fans at full speed... meaning you’d be able to run a GTX Titan or an AMD 390 with an overclocked CPU and still have yourself some decent temps. However, if your plan includes the latter, I would be one to strongly suggest going the route of the EK Predator 360 XLC or adding an additional 120mm or 140mm radiator to your EK 240 XLC.

Fan Noise Levels at Testing Levels
Fan noise is often a matter of personal perception, but we can measure the sound level output of the fans in a quantitative way: decibels. Measurement of sound using the decibel is a logarithmic factor, so as the value of the measured level increases, it increases as a steady multiplier. This means that every increase of 10dB means a doubling in perceived noise/sound level.
Using my digital decibel meter, I captured the noise levels at 0 RPM, 1200 RPM, 1600 RPM and 1900 RPM at a distance of one foot (30cm):

j06nvZU.jpg


Fans completely off shows a room ambient sound level of < 30 dB, which the meter is unable to register. At 1200rpm the meter reading still results in levels less than 30dB. At 1600rpm, the meter registers 33.1dB and at full speed of 1900+ RPM, we get a value of 36.2dB. For comparison, many decibel charts list 30dB as the levels in an audio recording studio (obviously silent) and 40dB would be similar to the sounds in a quiet bedroom. The levels taken on this cooler fall below 30dB and don’t even reach the levels thought to be inclusive of a very quiet room at full fan speeds. Compare this to the Corsair H110i GTX where the lowest fan speeds tested (1200 RPM) were 38.2dB and the highest fan speeds were 64.2dB (2300rpm). 70dB is considered sound levels of a passenger car on the highway.

Conclusion and Final Thoughts
The EKWB Predator 240 XLC is an excellent cooler built from excellent components which are well-known in the watercooling world. It is a bit more industrialist in design, which gives it a slightly larger, blocky profile which might turn the casual PC cooling enthusiast away. However, understanding the potential that lies within the cooler is where it truly shines as a complete solution. Most boxed all in one coolers do not allow for the user to expand to include other components they wish to cool as this voids the warranty. Furthermore, most of those other solutions are not built to withstand the added flow requirements and head pressure needed to support an expanded cooling loop; the EK Predator series is, though. It also hits the sweet spot of mounting options as it is a 2x120mm mounting option while some of the larger closed loop coolers are moving to the 2x140mm option.

Like the Swiftech H220/H240-X coolers, the EK XLC cooling solution is meant to tackle the cooling needs of the budding watercooling enthusiast – a simple to deploy boxed solution for a zero-day setup as well as allowing the flexibility to expand as they become familiar with watercooling and wish to add to their loop without having to subtract extra cash from their wallets in order to start over again. I’ve helped so many users in the forums who have started out with a closed loop boxed cooler and then wished to branch out to bigger watercooling endeavors, only to be discouraged at the thought that it would have been simpler (and cheaper) to have started with a beginner cooling loop than a closed loop cooler.

The only negative that I took away from working with the Predator is one of the fans, the wiring harness loops and the mounting screws. The hex-key allen screws are very short and require that you align the Allen key through one hole in the fan in order to thread the screw into the housing to connect fan to radiator. This meant that I needed a pair of needle nose pliers to help align the screw threads and get them started correctly. Most common fan screws run through both holes of the fan mount and then into the radiator, which makes them easier to attach. Since the Predator's fans are already mounted in PULL orientation, and with this actually being the slightly advantageous airflow that I found in testing, you can be relieved that you shouldn't have to remove and re-attach the fans.

lNi3VgI.jpg


However, what is different than many other coolers is that the way in which the Predator's fans ARE attached, the cooler expects you to then mount the fans to the top or side of your case to exhaust the air, rather than the fans being mounted in a way that pulls cool air from the outside, into the case. It would seem that EK has designed the Predator to work mainly as a PULL/exhaust cooler to help keep the rest of your case cool, rather than being a PULL/intake cooler, which would benefit the Predator's radiator, but would also dump the warmer radiator air inside the case, leaving you with a need to otherwise exhaust it back out. Now, those strange fan screw mounts make a bit more sense as it leaves you those fan mounting holes in which to mount the Predator XLC to the case itself. This is shown as a diagram in the Predator manual as one of the ideal mounting orientations:

re2OgwO.jpg


The EK Predator XLC solutions offer the performance and future customization of a custom watercooling loop AND the simplicity of a closed loop cooler solution – built of quality watercooling components by EK; a company well-known for their attention to detail, performance and quality.


**Update - As I concluded the write-up for this set of testing, it was announced that EKWB is issuing a recall for all XLC Predator 240 and 360 Rev 1.0 coolers produced between October and December of 2015. Supposedly there is an issue where they discovered a problem batch of o-rings that were installed and shipped with the units during this time. There is a chance that these units could leak and cause system damage.

EK has offered a full RMA; either replacement or complete refund for all units within this time period. They will also pay for return/replacement shipping as well as compensation for any damaged components due to leaks from the product. Please see the announcement from the Tom's Hardware home page for more details and information on replacement/reimbursement.

It should be noted that the unit I had for testing does meet these criteria as listed by EKWB but I did not encounter any issues with the cooler during my tests, which included several disassemblies of the system during the testing period.
 
Nice work! :)

The noise levels of those fans are very low for a 20fpi rad, they must use a revolutionary blade pitch design to increase static pressure.

Will you be testing any higher static pressure fans or adding additional fans to create push/pull to see if the cooling of the radiator improves even further?

The quality of construction of the predator is just what I was hoping to see, Thank You for all your hard work, excellent pictures, and attention to detail.
 

rubix_1011

Contributing Writer
Moderator
I had thought about the 'optional' fans idea on testing and was originally going to use my 3000 RPM Ultra Kaze's on it, but then stopped to remember that I didn't have any similar fans for the 280 rad on the Corsair H110i GTX.

However...I might create a supplement set of tests later on with some additional testing with them, just to see. I wanted to see if I could test within the boundaries of what the retail box held in order to present level playing field comparisons without spending extra money on variables. But...I do know that it is of interest to many people, including myself. One of the biggest challenges in doing this is simply the amount of time to run a complete suite of tests for push & pull at all load levels. At least with running 3000 RPM fans at full tilt, you're only running 1 test for push, 1 for pull and then changing the load.

The biggest thing I was puzzled about was the choice to undervolt the DDC pump. When compared to the DDC that I have on my test bench for, well, testing...there was a very noticeable difference in flow. Would be interesting to see the differences when using a fully-volted DDC on this cooler; and makes me wonder why they opted against it.