Noctua NH-D14 Review

ccokeman - 2009-11-12 20:45:45 in CPU Cooling
Category: CPU Cooling
Reviewed by: ccokeman   
Reviewed on: December 10, 2009
Price: $87.95

Introduction:

One improvement most of us make to the system we have or a new system that is being built is the cooling system for the processor. Whether it is to just drop the operating temperatures to a level much lower than the OEM solution is capable of on your stock clocked processor or so you can try to reach some righteous clock speeds to improve performance, keeping the CPU cool is the ultimate goal. There have been many designs over the years, some successful and some, well, not so successful. The current crop of successful coolers are all tower-style heatsinks that use heatpipes connected to the contact plate and run through a large fin array to maximize the amount of surface area available. Bigger fin arrays and more and larger heatpipes seem to be the current trend. Noctua has one of the top performers in its NH-U12P SE 1366 that compares favorably with one of the other "King of the Hill" contenders, the Thermalright TRUE. Noctua is not only known for its cooling performance, but for the fact that its solutions offer performance without the noise penalty commonly associated with high end air cooling, thanks in part to the Vortex Notches used on its fans to smooth out the airflow, help build static pressure, and reduce the acoustics.

The Noctua NH-D14 makes use of the tower design, but instead of a single tower, it has two large fin arrays interconnected with a total of six heatpipes to increase the cooling performance. That's one additional fin array and two more heatpipes than the NH-U12P is equipped with. Instead of two fans of the same dimension, you get a single 120mm NF-P12 and single 140mm NF-P14 FLX used in a push/pull configuration. If the results delivered by the NH-U12P are any indication, this cooler should eclipse the performance of Noctua's previous top-of-the-line cooling solution.

Closer Look

I received the NH-D14 with packaging that was not the final revision, so I have chosen not to show that as a representation of the final product. Suffice it to say that the cooler was packed in a foam enclosure that was more than adequate for the task. The first impression of the NH-D14 is that this thing is huge! It measures 160mm tall (just over 6.25 inches) and 140mm wide (just over 5.5 inches), and with the fan on the front, you get just over 6 inches in depth. That's a lot of real estate to cover on a motherboard. The NH-D14 is shipped pre-assembled with the fans and mounting brackets for the SecuFirm2 mounting kit installed. The fan used to push air into the heatsink is the NF-P12, with the NF-P14 used in the middle to force air through the second tower. Notice that the 140mm fan blows air under the fin arrays to offer additional cooling to the motherboard components.

 

 

The base of the NH-D14 is smooth and contains some fine machining marks – something visible in the right light, but nothing that would impact cooling performance. There are a total of six heatpipes that run from one fin array through the contact pad and out to the second fin array. This arrangement should offer exceptional load capacity, allowing you to push the limits of your hardware.

 

 

This heatsink is just as imposing when you remove the fans. The two towers are just massive, each one almost as large as the fin array on the TRUE. The front of each tower has a total of four vibration dampers to keep the fans from vibrating against the fins of the tower. These are much thicker than those used on the NF-P12 and look like they will be more effective, long term. The heatpipe termination looks much better than some heatsinks I have seen. The top view shows the toothed surface of the fin array. This gives a substantial boost to available surface area to improve cooling. Six large heatpipes carry the thermal load from the contact surface in the fins to be dispersed. The mounting brackets to attach the NH-D14 to the SecuFirm 2 mounting assembly are pre-installed, so you have one less thing to install, which of course lessens the install time.

 

 

 

Some say that size matters when it comes to performance. To give you an idea how large the NH-D14 is, I've compared it to the TRUE and Noctua NF-P12. No matter which way you look at it this thing, it is absolutely massive! Left to right, you have the Noctua NF-P12, the NH-D14 and the TRUE!

 

 

Impressive? Yes! But that's not all that comes as part of the package!


 

Closer Look:

Noctua's fans are an important part of the performance equation. Both the 120mm and 140mm fans are a nine blade design that feature Vortex Notches to improve the airflow and static pressure without increasing the noise the fans make. If you look carefully you will see that the notches are offset from one blade to the next. This helps to spread the airflow noise over many different frequencies so that you do not end up with any annoying fan noise. Another of the hidden bits that you don't see, but quickly appreciate, is the use of SSO (Self Stabilizing Oil Pressure) bearings to further reduce the noise and improve longevity. The NF-P12 pushes around 55CFM at its rated maximum speed, while the NF-14 FLX pushes right around 65 CFM at just under 20dBA. The quality build of Noctua's products are clearly evident. The wiring for each fan is sleeved with a rubbery-textured sleeving that has held up well on the older Noctua fans that I have.

 

 

 

 

 

 

 

 

 

Each of the two fans come equipped with a captured mounting clip to mount the fans to the NH-D14. The clip is held in place on each fan through a unique clip assembly that locks the clip into place. At first glance you have to wonder how it comes apart to use them as case fans if you choose to use the NH-D14 as a passive cooler. One end of the clip is bent with two 90 degree turns at the end that runs through the post assembly, while the other end uses a single bend so that when you lift the clip up, it will slide out and the clip can be rotated out of place. To lock the fan onto the NH-D14, you just lay it into place and pull the clip into position and lock it into the notch on the side of the fin array. This arrangement is much easier to use than the non-captured clips used on the NH-U12P. One detail I had not noticed before was the directional arrows on the outside of the frame that tell the direction of the fans rotation as well as the direction of the airflow.

 

 

 

The bundle of accessories that Noctua provides is top notch. You get pretty much everything you need, including the tools to install the NH-D-14. There is a manual for both AMD and Intel setups that includes very detailed instructions on how to install the heatsink. You get a bag with the AMD components, a bag with the Intel components, and a third bag that includes shared components. What you get in the shared components bag is a tube of Noctua's finest thermal paste NT-H1, a metal case badge to show your support, screws to mount the fans to your chassis, vibration isolators that accomplish the same function, a Y-harness to connect both fans to a single header, and the Low Noise and Ultra Low Noise fan adapters.

 

 

 

The SecuFirm2 backing plate for the Intel kit is a pretty sturdy piece and is insulated well so that there are no shorts on the back of the motherboard PCB. The backplate and brackets are drilled for use with multiple Intel sockets. You have compatibility with socket 775, 1156 and 1366 systems with only a single kit. This kit is a breeze to use with its use of positive stops on the attaching hardware, preventing you from over-tightening the mounting mechanism.

 

 

Fitting the NH-D14 onto the motherboard is not challenging by itself. The problems come when you stuff it into the chassis. Much like the Sapphire HD 5970, you need a larger chassis to take advantage of the NH-D14. It is large and takes up a large amount of space. In the Thermaltake Armour+ Full Tower I use, it was a tight fit against the top rail. Additionally, if you plan on using memory modules with tall heatspreaders, you may be out of luck if the fan overlaps the modules. If you have room, you could always relocate the front pusher fan onto the backside as a puller to help out. My Mushkin modules were fine in this case, as they use a shorter heatsink.

 

Let's see if the size of this cooler makes a performance difference over its contemporaries.

Specifications:

Socket compatibility     
Intel LGA1366, LGA1156, LGA775 & AMD AM2, AM2+, AM3 (backplate required)
Height (without fan)
160mm
Width (without fan)
140mm
Depth (without fan)
130mm
Height (with fan)
160mm
Width (with fan)
140mm
Depth (with fan)
158mm
Weight (without fan)
900 g
Weight (with fan)
1070/1240* g
Fan compatibility
140x140x25 & 120x120x25mm
Scope of Delivery
1x NF-P12 premium fan
 2x Ultra-Low-Noise Adaptor (U.L.N.A.)
 Y-Split Cable
 4x Vibration-Compensators (for using NF-P12 as case fan)
 4x Fan screws (for using NF-P12 as case fan)
 NT-H1 high-grade thermal compound
 SecuFirm2™ Mounting Kits
 Noctua Metal Case-Badge
Warranty
6 Years

 

Fan specifications
Model
Noctua NF-P14 & Noctua NF-P12
Bearing
SSO-Bearing
Rotational Speed (+/- 10%)
1200 / 1300 RPM
Rotational Speed with U.L.N.A. (+/- 10%)
900 / 900 RPM
Airflow
110,3 / 92,3 m³/h
Airflow with U.L.N.A.   
83,7 / 63,4 m³/h
Acoustical Noise
19,6 / 19,8 dB(A)
Acoustical Noise with U.L.N.A.
13,2 / 12,6 dB(A)
Input Power
1,2 / 1,08 W
Voltage Range
12 V
MTBF   
> 150.000 h

 

Features:

 

 

 

Allinformation courtesy of Noctua @ http://www.noctua.at/main.php?show=productview&products_id=34&lng=en&set=1
 

Testing:

To put the latest performance heatsink from Noctua to the test, I will be making a comparison of the temperatures at idle and under load, both while the CPU is at the stock voltages and clock speeds, as well as when the CPU is overclocked and over-volted to show what kind of cooling performance this monster of a heatsink has to offer when compared to other socket 1366-compatible high performance heatsinks. These heatsinks will be tested head-to-head as they are delivered from the manufacturer. I could throw in a bunch of testing variables, but it is not what the products are capable of as delivered. To test the idle temperatures, I will allow the computer to stay idle for 30 minutes and take the idle temperature at this point. For the load testing, I will use Prime95 version 25.9 and choose the blend testing and allow the processor and memory controller to heat up to the maximum temperatures. The time frame in the 14k test is usually a four hour run to allow the temperature to peak. I will use Real temp 3.0 to take the high and low temperatures and average the temperatures generated over the four cores as my reported temperature.

Testing Setup:

 

Comparison Heat sinks:

 

 

 

 

 

 

The NH-D14 is the heatsink that delivers the lowest temperatures in the testing that I performed. The idle temperatures were all pretty close and don't really tell a true story. The load temperatures are the best indicator of performance. What you get is about a 4 degree Celsius drop from the already high end performance of the comparison heatsinks, the Thermalright TRUE and Noctua NH-U12P SE 1366. I really was not satisfied with a 4 degree margin and did a couple of remounts, but the results were repeatable. What I really wanted to know was how well the NH-D14 performed with a much heavier heat load, so I loaded up a 4.2GHz overclock on the Core i7 920 to see what kind of temperatures I would get, as well as seeing if any reduction in operating temperatures gave me Prime 95 stability at this level.

 

What I was able to achieve with the NH-D14 was prime stability at 4.2GHz with load temperatures that peaked at 70 degrees Celsius and ran most of the time in the mid to high 60's. I was not prepared for this result, but am now thoroughly impressed with the performance of the NH-D14. All that, with no noise. Absolutely sick!


 

Conclusion:

I went into the testing knowing that the performance of the NH-D14 was going to be good. I just did not realize how good. When I ran it through the testing, it was far and above superior to the comparison heatsinks. Of course, there is no reason for it not to be, simply based on the size and construction of the NH-D14. The surface area available to remove the heat is huge. It easily looks like the dual-fin arrays each encompass as much or more surface area than the previous top dog in Noctua's arsenal, the NH-U12P SE 1366. Add in an additional two heatpipes and you have a winning recipe. With load temperatures in the high 40's Celsius (49°C to be exact) when overclocked to 3.3GHz, I figured I would take a shot at some big numbers with an overclock up to 4.2GHz. My DO stepping 920 is not a bad clocker and could previously do 4.2GHz on air, but the temperatures would keep the processor just stable enough to only run this speed for short benchmark runs and did not offer any kind of stability. Something I needy when running a distributed computing project. Enter the Noctua NH-D14. I cranked up the bclock to 210 with a 20 multiplier for an even 4.2GHz, adjusted the vcore to 1.375v in the BIOS and crossed my fingers. Of course, hyperthreading is on. Temperatures at 4.2GHz with any of the other heatsinks in my possession would start pushing 80°C+ at this speed. Not so with the NH-D14. The highest temperature I saw was 71°C on two of the four cores, with the heatsink for the most part keeping the temperatures in the 65°C to 68°C range. Pretty spectacular results from this heatsink. It delivered results that averaged four degrees Celsius better under load both at stock and overclocked speeds. Four degrees Celsius does not seem like a big drop in temperatures, but it is when you are comparing top of the line heatsinks.

The SecuFirm 2 mounting package is simple to install and does not leave any doubt about how tight you need to tighten the mounting nuts. The spacers and studs with positive stops allow anyone to install it without fear of getting it hooked up incorrectly. The Intel bracket assembly comes with an insulated backplate that is drilled for multi-socket compatibility. 775,1156 and 1366 sockets are covered, so you do not have to worry about your next upgrade if you are on a socket 775 setup. The bundle that Noctua sends is mighty generous. You get all the bracketry that you could need for most popular sockets, a fan power splitter so you only have to use one header or 3-pin connector to power both fans, and low noise and ulta low noise adapters to bring the fan speeds down to a lower level to cut down on the noise. Regardless, I could not hear the fans, so this should help in low noise environments, if needed. Each of the two towers have vibration dampers applied and are large enough that there is none of the annoying noise from a fan rattling against the fin array. You get a tube of Noctua's NH-T1 premiere thermal compound, and case screws or vibration compensators to mount the fans to a case, if you choose not to use them on the NH-D14. You even get a screwdriver to put it all together, so no tools are needed on your part. The NH-D14's best attribute is also its biggest detriment. It is huge, no ifs, ands or buts about it. It takes up a lot of real estate in the chassis and will interfere with memory modules that use large heatspreaders for cooling. However, standard-height modules will fit under the fan. On one hand, the size is what allows the incredible performance, but it may not fit in every chassis. In my Thermaltake Armour+, the NH-D14 was right up against the the support rail for the power supply. Not touching, but within a sixteenth of an inch from it. When it comes to pricing, you have to expect that this level of performance and quality will come with a price. Just shy of $90 is steep, but with this kit you really get what you pay for. Steep, but worth it. All things considered, this is the best heatsink I have ever tested. You get a massive dual-tower, six-heatpipe design that incorporates both a 120mm and 140mm fan to deliver outstanding results that show larger gains as the thermal load increases. All that without the noise associated with high end air cooling. Hey, size really does matter, and yes, you can pay for performance!

 

Pros:

 

Cons: