With big radiators and no vents, closed-loop coolers give us a taste of liquid-cooling’s benefits, sans the risk. While less punishing on the motherboard than big air, can any closed-loop liquid cooler beat Noctua's famed NH-D14 in thermal performance?
Whether we’re talking Porsche engines or performance PCs, surface area has always been liquid cooling’s greatest advantage. You get additional benefits from moving your radiator to a cooler part of your case, though few PC enclosures are designed for this (here’s one). And while it's possible to build an air-only solution similar in size to our favorite liquid-based products, we don't like the thought of a multi-pound copper heat sink hanging from our motherboards. After all, we've seen big coolers devastate processors and sockets when they're strapped in and shipped.
As a result of those logistic nightmares, most boutique builders use liquid cooling of some sort in their overclocked PCs. Of course, those high-end shops have enough experience to know not to ship a filled water-cooled box with an air vent. And they certainly can't expect most of their customers to know how to top-off their own systems, or follow a regular flush and fill routine. Closed-loop coolers are consequently popular because they prevent both leakage and maintenance hassles.

| Closed-Loop CPU Cooler Features | ||||
|---|---|---|---|---|
| Corsair Hydro H90 | Enermax ELC120 | NZXT Kraken X40 | Thermaltake Water2.0 Extreme | |
| Length | 6.8" | 6.0" | 6.8" | 10.7" |
| Width | 5.5" | 4.8" | 5.4" | 4.8" |
| Rad. Thickness | 1.1" | 1.3" | 1.1" | 1.5" |
| Cooling Fans | 1 x 140 x 25 mm | 2 x 120 x 25 mm | 1 x 140 x 25 mm | 2 x 120 x 25 mm |
| Total Thickness | 2.1" | 3.3" | 2.1" | 2.5" |
| Control Type | Motherboard Fan Headers | Motherboard Fan Headers | Software Programmable | Software Programmable |
| Weight | 28 Ounces | 28 Ounces | 32 Ounces | 41 Ounces |
| AMD Sockets | AM2 to AM3+ | AM2 to AM3+ | AM2 to AM3+ | AM2 to AM3+ |
| Intel Interfaces | 1156/1155, 1366, 2011 | 775, 1156/1155, 1366, 2011 | 1156/1155, 1366, 2011 | 1156/1155, 1366, 2011 |
| Web Price | $100 | $90 | $85 | $145 |
If you don't plan to move your PC, you might not be as stressed about the weight of a big air cooler. For you, liquid cooling is all about temperature control. Therefore, the preeminent big air cooler serves as today’s performance baseline in a follow-up to December's Four Closed-Loop CPU Coolers Take On Noctua's NH-D14.
- The Search For A Better Sink
- Corsair Hydro Series H90
- Installing Corsair's H90
- Enermax ELC120
- Installing Enermax's ELC120
- NZXT Kraken X40
- Installing NZXT's Kraken X40
- NZXT Control Software
- Thermaltake Water2.0 Extreme
- Installing Thermaltake's Water2.0 Extreme
- Thermaltake Control Software
- Test Hardware Configuration
- Cooling, Fan Speed, And Noise
- Evaluating Performance Results
- Closed-Loop Cooling: Value Versus Versatility

The justification of buying a closed-loop water cooler (at least for me) was getting the weight of a massive air-cooler off my CPU and therefore removing the stress of that weight from the motherboard. Slightly better temps are just a bonus.
That depends entirely on how airflow is within your case and as such temps vary widely between cases and setup. Believe me when I say that VRM temps are not an issue in my Define XL R2.
http://www.tomshardware.com/reviews/water2.0-extreme-kraken-x40-hydro-h90-elc120,3434-13.html
Doesn't really justify buying one of these over an NH-D14, really only to say that you have "water cooling."
The justification of buying a closed-loop water cooler (at least for me) was getting the weight of a massive air-cooler off my CPU and therefore removing the stress of that weight from the motherboard. Slightly better temps are just a bonus.
I've had my Noctua D14 for almost two years now. It's one of the best investments I've made in my PC. If i needed better cooling, i would just buy a third fan for the D14, and it would drop a few degree's and give better cooling than all but the top to Closed-loop coolers tested here. Plus I'd have lower noise.
If room is not a problem there is NO reason to get a closed-loop system. A D14 with a third fan is $100 and is out performed only in thermals by more expensive and louder CL coolers.
I understand the issue of weight. But saving 2-3 degree's on cpu (aside from the TT 2.0) isn't worth your mobo gaining 15-17 degree's which results in a much hotter system and more stress on all your other components.
Theoretically, the more cooling the space, the more efficient.
You should have compared air coolers to other air coolers, then compared single fan 120mm and 140mm CLC with other single fan CLC and then do a dual fan CLC vs. other dual fan CLC.
I mean, come on, I'm glad you reviewed this, but you should have at least took the best air cooler, single fan CLC, and dual fan CLC in the current market and then tested and compared results.
That depends entirely on how airflow is within your case and as such temps vary widely between cases and setup. Believe me when I say that VRM temps are not an issue in my Define XL R2.
I'm surprised why Noctua hasn't included their newest fans in their coolers. I'm sure its holding many potential customers off.
Allow me to clarify: ALL PRODUCTS SHOULD BE TESTED IN THE SAME WAY. If the H90's fan is reversed to get cold air, then everyone's fan must be reversed to get cold air. Reversing the fan violates the thermal design of the case, and it wasn't done for any other manufacturer.
The NH-D14 exhausts heat out the back of the case. It does, after all, blow directly towards the exhaust fan. Noctua never said "turn your exhaust fan AND CPU COOLER backwards to get colder air into our product, but don't do that for anyone else". Because everyone knows that such a request would be asking the tester to cheat. Except maybe for fans of the company that made that request. Fairness demands that all products be tested in the same way.
In fact, the other case option was Fractal Design's Define R4, which holds a dual-fan radiator in front. But using it would have given an unfair advantage to dual-fan radiators. Intake temperature should be the same for all coolers because, get this, ALL PRODUCTS SHOULD BE TESTED IN THE SAME WAY.
When the H90 got tested IN THE SAME WAY as the other coolers, it lost. Period. That would be the end of the story except that some biased reader like you would have stepped in and said "but you never tested it the way Corsair said to test it". How can I call the review fair and you biased? Because consistent test methods are the hallmarks of fairness.
Next likely question would be "why not test them all with the air blowing in?" One answer would be "top mounted fans shouldn't blow downward, heat rises". Another answer would be "Dust filters are on front and bottom not top and back". Another answer would be "all top and rear fans should be exhaust, because heat rises" Another answer would be "you want the exhaust fans facing away from you so you don't have to hear or feel them". Another answer...
The simple answer is that the H90 lost when it was tested in the same orientation as the other coolers. To test it the other way violates the thermal design of the case. And thanks for bringing that up early so I could address it on Page One of the response thread
Both the H90 and the Kraken have a similar waterblock to the water 2.0 and they both cost less. Seems like you were paid by water 2.0 to give them an easy review...
CPU comes out 1 year later that leaves the overclocked CPU in the dust and costs $40 more after original CPU is sold and old cooler is not compatible with new socket.
Water cooling = no thanks.